Commit 5fccfd83ec, with the fix for #6806,
switched eval to buffer its output (like other builtins do). But this
prevents using eval with commands that wants to see the tty, especially
fzf. So only buffer the output if the output is piped to the next process.
builtin_eval needs to know whether to set up bufferfills to capture its
output and/or errput; it should do this specifically if the output and
errput is piped (and not, say, directed to a file). In preparation for
this change, add bools to io_streams_t which track whether stdout and
stderr are specifically piped.
Prior to this fix, builtin_eval would direct output to the io_chain of the
job. The problem is with pipes: `builtin_eval` might happily attempt to
write unlimited output to the write end of a pipe, but the corresponding
reading process has not yet been launched. This results in deadlock.
The fix is to buffer all the output from `builtin_eval`. This is not fun
but the best that can be done until we have real concurrent processes.
cherry-pick of a1f1b9c2d9Fixes#6806
Ensure that if eval is invoked as part of a pipeline, any jobs spawned
by eval will have the same pgroup as the parent job.
cherry-pick of 82f2d86718
Partially fixes#6806
Give string expansion an (optional) parent pgroup. This is threaded all
the way into eval(). This ensures that in a mixed pipeline like:
cmd | begin ; something (cmd2) ; end
that cmd2 and cmd have the same pgroup.
Add a test to ensure that command substitutions inherit pgroups
properly.
cherry-pick of 938b683895Fixes#6624
This was a weird one. We split the aliases correctly even with
multiple lines, but then broke it all again when we just printed the
description.
Note that it would be possible to use `string split0` here, but since
anything longer than a line is likely too long for a description
anyway we don't bother.
Fixes#6946.
(cherry picked from commit 1988bd2579)
When this switched over from directly piping commandline to storing
its output and using printf, I inadvertently always added a trailing
newline. That's probably annoying.
Note that this will now always *remove* a trailing newline (because
the command substitution does). That will barely make a
difference (because trailing newlines are quite unusual in the
commandline) and will probably feel better than keeping it - we could
even make a point of removing trailing whitespace in general.
Fixes#6927
(cherry picked from commit 6ebbe5a450)
Fixes#6798
This re-adds some false positives: functions, builtins and abbreviations
are suggested after commands like sudo but I don't think anyone had
complained about that.
(cherry picked from commit 2a89873e6d686fcff1d26d0914a8b9f90b7cc308)
Currently we do not add such command lines to the history, so there
won't be a suggestion from history anyway.
Fixes#6763 which occurs because midnight commander feeds fish commands
like this one (note the leading space)
` cd (printf '%b' '\0057home\0057johannes\0057git\0057fish\0055shell\0057build')`
(cherry picked from commit 390647ae34)
The output of
systemctl list-units
seems to include a marker of '●' or '*' for some units, even if the
output is not going to a terminal and "--no-legend" and "--no-pager"
are given. This appears
to be a recent development, and there does not appear to be a flag to
disable it.
So we simply filter it out in the completions to once again hopefully
offer the actual units.
Fixes#6740
Even if $DISPLAY is unset, xdg-open can be useful, and on systems that
have xdg-open, "open" is most likely some god awful outdated thing
called "openvt" elsewhere.
Fixes#6739
[ci skip]
If given a prompt that includes a non-ascii char and a C locale, fish
currently fails to properly display it.
So you set `function fish_prompt; echo 😃; end` and it shows empty
space.
While the underlying cause is obviously using a C locale and non-C
characters to begin with, this is an unacceptable failure mode.
Apparently I misunderstood wcstombs, so I inadvertently broke this in
2b0b3d3 while trying to fix 5134949's crash.
Just return the offending bit to pre-5134949 levels, so instead of an
infinite recursion we just call a lame function a couple of times.
This teaches the reader fast-path to use self-insert-notfirst, allowing
it to handle spaces. This greatly increases the performance of paste by
reducing redraws.
Fixes#6603. Somewhat improves #6704
This adds basic support for self-insert-notfirst. When we see a
self-insert-nonempty char event, we kick it back to the outer loop,
which only inserts the character if the cursor is not at the beginning.
This adds a new readline command self-insert-notfirst, which is
analogous to self-insert, except that it does nothing if the cursor
is at the beginning. This will serve as a higher-performance implementation
for stripping leading spaces on paste.
The default hg prompt is slow on large repositories (hg status takes
2-3 seconds on mozilla-central) which is unacceptable as a default.
Mimick our git prompt: by default, only show the current branch.
If the new variable $fish_prompt_hg_show_informative_status is set,
then use the old behavior.
[ci skip]
(cherry picked from commit da7b762f4a)
This is apparently quite slow on large svn repos (like 40 seconds
slow), and we don't have a good thing to display other than the full
file information.
So we'll have to disable it for now.
Fixes#6681.
[ci skip]
bbc3fecbe introduced a regression where support for 256 color was not
detected on xterm-like terminals that did not define the TERM_PROGRAM
env variable. Almost no terminal on linux define this variable.
f8ba0ac5bf introduced a bug where INT handlers would themselves be
cancelled, due to the signal. Defer processing handlers until the
parser is ready to execute more fish script.
Fixes the interactive case of #6649.
55e3270 introduced a regression where we would remove all completed
jobs. But jobs that want to print a status message get skipped, so
the status message (and associated event handlers) might not get run.
Fix this by making it explicit which jobs are safe to process, and which
should be skipped.
Fixes#6679.
Which happened when starting the selection at the end of the commandline.
In this case, selections still interact weirdly with autosuggestions (the
first character of the suggestion appears to be part of the selection
when it's not).
Fixes#6680
(cherry-picked from commit 99851c09b3)
Since #6406, read will trim whitespace before the last variable.
In this case there is only one variable, and the line looks like
M CHANGELOG.md
so it does indeed start with whitespace, and the whitespace is quite
significant.
Fixes#6650.
[ci skip]
This used to use doxygen's html blocks, which don't have a *direct*
equivalent in sphinx in code blocks.
Instead of adding this to the pygments highlighter, let's just use
some roles.
It's a teensy bit awkward as we then use block styling, but we want to
add more of our own styling anyway, so we can presumably get this
somehow, and these html tags look awkward and confuse people.
Fix#6640
[ci skip]
Appending to an fd doesn't really make sense, but we allowed the
syntax previously and it was actually used.
It's not too harmful to allow it, so let's just do that again.
For the record: Zsh also allows it, bash doesn't.
Fixes#6614
(cherry picked from commit aba900a71f)
When building fish-shell with the macOS 10.12 SDK, <sys/proc.h> does not
include <sys/time.h> but references `struct itimerval`. This causes a
compilation failure if we don't import <sys/time.h> ourselves.
This was previously masked by an import of <sys/sysctl.h>, which was
removed in fc0c39b6fd.
(cherry picked from commit 47aeaa1535)
This was lost in 35671dd9f0.
Even tho we plan to drop caret redirection, while it's there it should
fully work.
Fixes#6591.
(cherry picked from commit 13b470af07)
Glob ordering is used in a variety of places, including figuring out
conf.d and really needs to be stable.
Other ordering, like completions, is really just cosmetic and can
change if it makes for a nicer experience.
So we uncouple it by copying the wcsfilecmp from 3.0.2, which will
return the ordering to what it was in that release.
Fixes#6593
(cherry picked from commit f053cd27c6)
Perform an ad-hoc code signing with the hardened runtime.
This ensures that these executables can pass notarization.
The code signing ID is controlled by the MAC_CODESIGN_ID CMake
cache variable.
The `function --on-job-exit caller` feature allows a command substitution
to observe when the parent job exits. This has never worked very well - in
particular it is based on job IDs, so a function that observes this will
run multiple times. Implement it properly.
Do this by having a not-recycled "internal job id".
This is only used by psub, but ensure it works properly none-the-less.
faho:
Backport of 6bf9ae9aebFixes#6613
Same issue occurs here, as in #6270 (and fixed in 611a658 for `__fish_describe_command.fish`). Same reason. I've just copied the same workaround and changed the function name to match.
(cherry picked from commit f7edfba5d7)
This makes two changes:
1. Remove the 'brace_text_start' idea. The idea of 'brace_text_start' was
to prevent emitting `BRACE_SPACE` at the beginning or end of an item. But
we later strip these off anyways, so there is no apparent benefit. If we
are not doing brace expansion, this prevented emitting whitespace at the
beginning or end of an item, leading to #6564.
2. When performing brace expansion, only stomp the space character with
`BRACE_SPACE`; do not stomp newlines and tabs. This is because the fix in
came from a newline or tab literal, then we would have effectively
replaced a newline or tab with a space, so this is important for #6564 as
well. Moreover, it is not easy to place a literal newline or tab inside a
brace expansion, and users who do probably do not mean for it to be
stripped, so I believe this is a good change in general.
Fixes#6564
Just another version of the error. We still want to get a bug if it
ever triggers a *wrong* error, so we still list all the options
instead of going for `.*option:.*Z.*`.
Fixes#6554
(cherry picked from commit e8000cfea9)
Solaris/OpenIndiana/Illumos `rm` checks that and errors out.
In these cases we don't actually need it to be a part of $PWD as
it's just for cleanup, so we `cd` out before.
See #5472
See 1ee57e9244Fixes#6555Fixes#6558
(cherry picked from commit 9cbd3d57a0)
OpenBSD uses [unveil(2)](https://man.openbsd.org/unveil) in chromium and
firefox. This means that things outside of directories like ~/Downloads or /tmp are not visible to the
browsers.
Change webconfig so it uses tempfile.NamedTemporaryFile to create our temp file.
We'd use $__fish_data_dir, but that already had the "/fish" component,
and then we'd add it again later, so we would try to find vendor
functions in places like
/usr/share/fish/fish/vendor_functions.d
which obviously don't exist.
Instead let's add the "/fish" component to the xdg dirs early, which
also saves us from having to repeat it later.
Fixes#6428
See #6508
[ci skip]
Sometimes we must spawn a new thread, to avoid the risk of deadlock.
Ensure we always spawn a thread in those cases. In particular this
includes the fillthread.
complete -C'echo $HOM ' would complete $HOM instead of a new token.
Fixes another regression introduced in
6fb7f9b6b - Fix completion for builtins with subcommands
64 is too low (it's actually reachable), and every sensible system should have a limit above
this.
On OpenBSD and FreeBSD it's ULONG_MAX, on my linux system it's 61990.
Plus we currently fail by hanging if our limit is reached, so this
should improve things regardless.
On my linux system _POSIX_THREAD_THREADS_MAX works out to 64 here,
which is just too low, even tho the system can handle more.
Fixes#6503 harder.
This commit recognizes an existing pattern: many operations need some
combination of a set of variables, a way to detect cancellation, and
sometimes a parser. For example, tab completion needs a parser to execute
custom completions, the variable set, should cancel on SIGINT. Background
autosuggestions don't need a parser, but they do need the variables and
should cancel if the user types something new. Etc.
This introduces a new triple operation_context_t that wraps these concepts
up. This simplifies many method signatures and argument passing.
When executing a buffered block or builtin, the usual approach is to
execute, collect output in a string, and then output that string to
stdout or whatever the redirections say. Similarly for stderr.
If we get no output, then we can elide the outputting which means
skipping the background thread. In this case we just mark the process as
finished immediately.
We do this in multiple locations which is confusing. Factor them all
together into a new function run_internal_process_or_short_circuit.
It's now good enough to do so.
We don't allow grid-alignment:
```fish
complete -c foo -s b -l barnanana -a '(something)'
complete -c foo -s z -a '(something)'
```
becomes
```fish
complete -c foo -s b -l barnanana -a '(something)'
complete -c foo -s z -a '(something)'
```
It's just more trouble than it is worth.
The one part I'd change:
We align and/or'd parts of an if-condition with the in-block code:
```fish
if true
and false
dosomething
end
```
becomes
```fish
if true
and false
dosomething
end
```
but it's not used terribly much and if we ever fix it we can just
reindent.
They need to be escaped twice, for the local and the remote shell.
Also don't suggest local files as rsync remote paths (-a -> -xa) and
fix completion for remote paths containing multiple consecutive spaces.
Fixes#1872
[ci skip]
$XDG_DATA_DIRS/vendor_{completions,conf,functions}.d
Additionally, CMake variables extra_{completions,conf,functions}dir are
appended, if they are not already contained in $XDG_DATA_DIRS.
If XDG_DATA_DIRS is not defined, we fall back to
$__fish_datadir/vendor_completions.d:${extra_completionsdir}
for completions. Same for conf and functions.
The logo is actually extracted from the site, but since it's just for
the appimage (I don't even know where it shows it, tbh) it's okay for
now.
Progress towards #6475.
[ci skip]
for-loops that were not inside a function could overwrite global
and universal variables with the loop variable. Avoid this by making
for-loop-variables local variables in their enclosing scope.
This means that if someone does:
set a global
for a in local; end
echo $a
The local $a will shadow the global one (but not be visible in child
scopes). Which is surprising, but less dangerous than the previous
behavior.
The detection whether the loop is running inside a function was failing
inside command substitutions. Remove this special handling of functions
alltogether, it's not needed anymore.
Fixes#6480
'fish_test_helper print_pid_then_sleep' tried to sleep for .5 seconds,
but instead it divided by .5 so it actually slept for 2 seconds.
This exceeds the maximum value on NetBSD so it wasn't sleeping at all
there.
Fixes#6476
This removes a call to `sed` and allows the user to specify shortening
via the variable.
We still default to disabling shortening because this prompt never
did.
[ci skip]
Empty items are used as sentinels to indicate that we've reached the end of
history, so they should not be added as actual items. Enforce this.
Fixes#6032
We just do a cheesy version check and hope it works out.
If this is fixed in 10.15.4, we have to reenable it. If it still isn't
fixed in 10.16, we need to adjust it.
Fixes#6270
This reduces the syscall count for `fish -c exit` from 651 to 566.
We don't attempt to *cache* the pgrp or anything, we just call it once
when we're about to execute the job to see if we are in foreground and
to assign it to the job, instead of once for checking foreground and
once to give it to the job.
Caching it with a simple `static` would get the count down to 480, but
it's possible for fish to have its pgroup changed.
Store the entire function declaration, not just its job list.
This allows us to extract the body of the function complete with any
leading comments and indents.
Fixes#5285
In particular, this allows `true && time true`, or `true; and time true`,
and both `time not true` as well as `not time true` (like bash).
time is valid only as job _prefix_, so `true | time true` could call
`/bin/time` (same in bash)
See discussion in #6442
Use string split instead of cut - which we'd fork for 2*signal
count times in a loop when tab was first pressed. Noticably faster
If giving a signal num, what works everywhere is -NUM, if giving
a signal name, what works everywhere is -s NAME - don't show -sNUM
or -NAME completions; that only works on GNU and it's redundant
anyhow as we show the signal number in the description field for -s
or the signal name for the -NUM case in the pager.
Sort -sNAME completions by the signal number not alphabetical
Shorten descriptions
Extend the commit 8e17d29e04 to block processes, for example:
begin ; stuff ; end
or if/while blocks as well.
Note there's an existing optimization where we do not create a job for a
block if it has no redirections.
job_promote attempts to bring the most recently "touched" job to the front
of the job list. It did this via:
std::rotate(begin, job, end)
However this has the effect of pushing job-1 to the end. That is,
promoting '2' in [1, 2, 3] would result in [2, 3, 1].
Correct this by replacing it with:
std::rotate(begin, job, job+1);
now we get the desired [2, 1, 3].
Also add a test.
This PR is aimed at improving how job ids are assigned. In particular,
previous to this commit, a job id would be consumed by functions (and
thus aliases). Since it's usual to use functions as command wrappers
this results in awkward job id assignments.
For example if the user is like me and just made the jump from vim -> neovim
then the user might create the following alias:
```
alias vim=nvim
```
Previous to this commit if the user ran `vim` after setting up this
alias, backgrounded (^Z) and ran `jobs` then the output might be:
```
Job Group State Command
2 60267 stopped nvim $argv
```
If the user subsequently opened another vim (nvim) session, backgrounded
and ran jobs then they might see what follows:
```
Job Group State Command
4 70542 stopped nvim $argv
2 60267 stopped nvim $argv
```
These job ids feel unnatural, especially when transitioning away from
e.g. bash where job ids are sequentially incremented (and aliases/functions
don't consume a job id).
See #6053 for more details.
As @ridiculousfish pointed out in
https://github.com/fish-shell/fish-shell/issues/6053#issuecomment-559899400,
we want to elide a job's job id if it corresponds to a single function in the
foreground. This translates to the following prerequisites:
- A job must correspond to a single process (i.e. the job continuation
must be empty)
- A job must be in the foreground (i.e. `&` wasn't appended)
- The job's single process must resolve to a function invocation
If all of these conditions are true then we should mark a job as
"internal" and somehow remove it from consideration when any
infrastructure tries to interact with jobs / job ids.
I saw two paths to implement these requirements:
- At the time of job creation calculate whether or not a job is
"internal" and use a separate list of job ids to track their ids.
Additionally introduce a new flag denoting that a job is internal so
that e.g. `jobs` doesn't list internal jobs
- I started implementing this route but quickly realized I was
computing the same information that would be computed later on (e.g.
"is this job a single process" and "is this jobs statement a
function"). Specifically I was computing data that populate_job_process
would end up computing later anyway. Additionally this added some
weird complexities to the job system (after the change there were two
job id lists AND an additional flag that had to be taken into
consideration)
- Once a function is about to be executed we release the current jobs
job id if the prerequisites are satisfied (which at this point have
been fully computed).
- I opted for this solution since it seems cleaner. In this
implementation "releasing a job id" is done by both calling
`release_job_id` and by marking the internal job_id member variable to
-1. The former operation allows subsequent child jobs to reuse that
same job id (so e.g. the situation described in Motivation doesn't
occur), and the latter ensures that no other job / job id
infrastructure will interact with these jobs because valid jobs have
positive job ids. The second operation causes job_id to become
non-const which leads to the list of code changes outside of `exec.c`
(i.e. a codemod from `job_t::job_id` -> `job_t::job_id()` and moving the
old member variable to a non-const private `job_t::job_id_`)
Note: Its very possible I missed something and setting the job id to -1
will break some other infrastructure, please let me know if so!
I tried to run `make/ninja lint`, but a bunch of non-relevant issues
appeared (e.g. `fatal error: 'config.h' file not found`). I did
successfully clang-format (`git clang-format -f`) and run tests, though.
This PR closes#6053.
This is part of our (well, my) quest to spice up the default prompt.
In this case we color the host if $SSH_TTY is set, which is easy to
detect and helps draw attention to the host.
See #6398.
See #6375.
This variable holds an integer that resembles the fish version up to
that initializations were performed. It should be incremented whenever
some new initialization is required after upgrading fish. This should
not change the behavior for existing fish installations, except for a
minor message on installations that upgrade from fish<2.3.0.
[ci skip]
This use of eval is unsafe, not really all that useful and can spew
errors that can't be suppressed. So let's remove it, and in future add
a thing that can do expansions in a safe manner
Fixes#6456.
__fish_complete_suffix accepts a first argument containing a
brace-expansion, like
__fish_complete_suffix '.{c,cpp,py}'
We're gonna be removing the `eval` that does that shortly, so let's
remove all uses in our code.
"To assume" and such.
It doesn't check $SHELL, so it might have some other automagic that
can fail (probably still because of the login shell, but I have no
idea).
Override the special variable that
click-completion (https://github.com/click-contrib/click-completion)
uses to force it instead.
Really fixes#6454.
[ci skip]
This checks $SHELL to determine which completions to print, and $SHELL
is typically set by your login program.
So if the login shell isn't fish, this will print the wrong
completions.
Fixes#6454
[ci skip]
Turns out we never documented the "jump" ones.
That means the still-undocumented bind functions are
- vi-arg-digit
- vi-delete-to
- and
Mostly because I'd have to look up what they actually *do*, and
possibly rename them to be generic.
[ci skip]
It looks like the last status already contains the signal that cancelled
execution.
Also make `fish -c something` always return the last exit status of
"something", instead of hardcoded 127 if exited or signalled.
Fixes#6444
Two blocks of code were trying to do the same thing in different ways;
standardise on one, and only add the compile flags if CMake won't do it
itself (policy CMP0067).
Previously, the block stack was a true stack. However in most cases, you
want to traverse the stack from the topmost frame down. This is awkward
to do with range-based for loops.
Switch it to pushing new blocks to the front of the block list.
This simplifies some traversals.
Work around the issue in CMake where C++ standard doesn't get propagated
to CHECK_CXX_SOURCE_COMPILES. Also correctly check for std::make_unique;
the define was missing from the config.h header.
GCC 4.8 requires the use of `-std=gnu++11` or similar to enable atomic
features. However, older versions of CMake don't pick up the
project-wide target for C++11 when building the configure check targets.
Although CMake policy 0067 could be set to NEW to enable this, it only
exists on CMake 3.8 and newer, while many of our supported platforms are
on an older version.
This was previously required so that, if there was a redirection to a
file, we would fork a process to create the file even if there was no
output. For example `echo -n >/tmp/file.txt` would have to create
file.txt even though it would be empty.
However now we open the file before fork, so we no longer need special
logic around this.
780bac671f did not actually successfully
compile on any platforms, leading to -latomic always being added
(including on platforms it does not exist on).
Work on #5865.
Do this only when splitting on IFS characters which usually contains
whitespace characters --- read --delimiter is unchanged; it still
consumes no more than one delimiter per variable. This seems better,
because it allows arbitrary delimiters in the last field.
Fixes#6406
user_supplied was used to distinguish IO redirections which were
explicit, vs those that came about through "transmogrphication." But
transmogrification is no more. Remove the flag.
"-C" is short for "--case-sensitive", which is entirely okay with "--delete".
The one that isn't okay is "-X", which is short for "--Clear".
Seen on gitter.im
This patch keeps the existing `make` shims via `GNUmakefile` and
`BSDmakefile` but also resolves the issue reported in #6264 with
CMake-generated `Makefile` overwriting the extant `Makefile` causing the
source directory to become dirty once again.
Closes#6264
This prefixes files beginning with `-` with a `./` when generating
completions *in fish code*. Standard completions for directory listings
generated by the C++ directory traversal code are not afected by this
patch.
Most fish completions defer to `__fish_complete_suffix` to generate the
file/directory completions, these *will* be corrected.
As of GCC 7.4 (at least under macOS 10.10), the previous workaround of
casting a must-use result to `(void)` to avoid warnings about unused
code no longer works.
This workaround is uglier but it quiets these warnings.
The C++ spec (as of C++17/n4713) does not specify the sign of `wchar_t`,
saying only (in section 6.7.1: Fundamental Types)
> Type wchar_t shall have the same size, signedness, and alignment
> requirements (6.6.5) as one of the other integral types, called its
> underlying type.
On most *nix platforms on AMD64 architecture, `wchar_t` is a signed type
and can be compared with `int32_t` without incident, but on at least
some platforms (tested: clang under FreeBSD 12.1 on AARCH64), `wchar_t`
appears to be unsigned leading to sign comparison warnings:
```
../src/widecharwidth/widechar_width.h:512:48: warning: comparison of
integers of different signs: 'const wchar_t' and 'int32_t' (aka 'int')
[-Wsign-compare]
return where != std::end(arr) && where->lo <= c;
```
This patch forces the use of wchar_t for the range start/end values in
`widechar_range` and the associated comparison values.
Previously, if the user control-C'd out of a process, we would set a
bogus exit status in the process, but it was difficult to observe this
because we would be cancelling anyways. But set it properly.
If a Control-C is received during expanding a command substitution, we
may execute the job anyways, because we do not check for cancellation
after the expansion. Ensure that does not happen.
This should fix sporadic test failures in the cancellation unit test.
parser_t::eval indicates whether there was a parse error. It can be
easily confused with the status of the execution. Use a real type to
make it more clear.
$GIT_DIR is interpreted by git as an environment variable, pointing at the
.git directory. If git_version_gen.sh is run in an environment with an
exported GIT_DIR, it will re-export GIT_DIR to point at the fish source
directory. This will cause git operations to fail.
This could be reproduced as building fish as part of an interactive rebase
'exec' command. git_version_gen.sh would always fail!
Looking up a variable by a string literal implicitly constructs a wcstring.
By avoiding that, we get a noticeable reduction of temporary allocations.
$ HOME=. heaptrack ./fish -c true
heaptrack stats: # baseline
allocations: 7635
leaked allocations: 3277
temporary allocations: 602
heaptrack stats: # new
allocations: 7565
leaked allocations: 3267
temporary allocations: 530
Closes#6435.
close_fds=True is actually the default in Python 2.7 and 3.2, but not in
ancient (but still in production in Red Hat Enterprise Linux 6) Python
2.6. Enable it there as well.
From the `git-switch` documentation:
If <branch> is not found but there does exist a tracking branch in
exactly one remote (call it <remote>) with a matching name, treat as
equivalent to
$ git switch -c <branch> --track <remote>/<branch>
This adds a test for the obscure case where an fd is redirected to
itself. This is tricky because the dup2 will not clear the CLO_EXEC bit.
So do it manually; also posix_spawn can't be used in this case.
The IO cleanup left file redirections open in the child. For example,
/bin/cmd < file.txt would redirect stdin but also leave the file open.
Ensure these get closed properly.
Prior to this fix, a file redirection was turned into an io_file_t. This is
annoying because every place where we want to apply the redirection, we
might fail due to open() failing. Switch to opening the file at the point
we resolve the redirection spec. This will simplify a lot of code.
Prior to this change, a process after it has been constructed by
parse_execution, but before it is executed, was given a list of
io_data_t redirections. The problem is that redirections have a
sensitive ownership policy because they hold onto fds. This made it
rather hard to reason about fd lifetime.
Change these to redirection_spec_t. This is a textual description
of a redirection after expansion. It does not represent an open file and
so its lifetime is no longer important.
This enables files to be held only on the stack, and are no longer owned
by a process of indeterminate lifetime.
fish has to ensure that the pipes it creates do not conflict with any
explicit fds named in redirections. Switch this code to using
autoclose_fd_t to make the ownership logic more explicit, and also
introduce fd_set_t to reduce the dependence on io_chain_t.
* Make `type -p` and `type -P` behave as documented
* Recognize `-` as an additional sign of no path
Functions created via `source` (like by `alias`) cause `functions --details` to return `-`
rather than `stdin` when invoked upon them.
Set variables for available connections and SSIDs only when the completion is loaded.
This is not perfect but faster than scanning for connections everytime.
Don't complete connection UUID, DBUS-PATH, ACTIVE-PATH because they are unintelligible.
Instead only complete the connection name.
See #6379
[ci skip]
Prior to this fix, a job would hold onto any IO redirections from its
parent. For example:
begin
echo a
end < file.txt
The "echo a" job would hold a reference to the I/O redirection.
The problem is that jobs then extend the life of pipes until the job is
cleaned up. This can prevent pipes from closing, leading to hangs.
Fix this by not storing the block IO; this ensures that jobs do not
prolong the life of pipes.
Fixes#6397
* Add the `--succinct` flag to `type`
* Use `echo` rather than `printf`
* Change `succinct` to `short`; print path if known
* Clean up the printing logic ever so slightly
This was mostly dead, since $fish_color_host is set to normal in
__fish_config_interactive. The assignment was only used if the user
explicitly unsets fish_color_host (which they shouldn't, really).
Anyway it's weird to use cyan, use normal instead.
[ci skip]
The colors are set in __fish_config_interactive before the prompt is
painted for the first time.
Also initialize the $fish_color_status for the (pipe) status, bump the
version for that.
[ci skip]
If a command fails, print the pipestatus in red instead of yellow and
don't print the status of the last process again. See #6375.
Also use $fish_color_status for coloring status consistently.
Also use __fish_pipestatus_with_signal to print SIGPIPE instead
of a numeric code on e.g.: yes | less +q
[ci skip]
This purported to need python > 3.4, but used anypython.
Plus it's not super useful anyway since it can easily be told to
use *all* cpus, so there's no need to set it to the precise number.
See #6400.
[ci skip]
First tell them how to install (though we don't actually do that right
now), then tell them how to start it, and only *then* tell them how to
make it the default or uninstall it.
Just seems sensible to try it first then delete it.
Some more sections here were duplicated or not all that useful, and
it's weird to start with "Commands versus Functions".
Let's explain to people how to start fish, then let's get going.
Currently a job needs to know three things about its "parents:"
1. Any IO redirections for the block or function containing this job
2. The pgid for the parent job
3. Whether the parent job has been fully constructed (to defer self-disown)
These are all tracked in somewhat separate awkward ways. Collapse them
into a single new type job_lineage_t.
In preparation for concurrent execution, invert the control of function and
block execution. Allow a process to return an std::function that performs the
the execution. This can be run on either the main or a background thread
(eventually).
If Python 3.4 or later installed on the system, complement to the
number of physical cores. In addition, even if the number of physical
cores cannot be obtained, it was fixed to run properly.
(command pwd) uses the system's implementation of pwd. At least the GNU
coreutils implementation defaults to -P, which resulted in symlinks being
expanded when switching between directories with nextd/prevd.
- Don't use a guard uvar - we're only setting variables now, and
- that's basically free.
- Allow non-universal color variables
- Simplify the root color setting a bit.
- Some comments
[ci skip]
This was a bit stuffy and verbose, so try to make it a tad more human.
Also don't mention `fish_opt` constantly. It's not actually all that
useful as argparse isn't as difficult to use as we thought.
[ci skip]
This did some weird unescaping to try to extract the first word.
So we're now more likely to be *correct*, and the alias benchmark is
about 20% *faster*.
Call it a win-win.
This splits a string into variables according to the shell's
tokenization rules, considering quoting, escaping etc.
This runs an automatic `unescape` on the string so it's presented like
it would be passed to the command. E.g.
printf '%s\n' a\ b
returns the tokens
printf
%s\n
a b
It might be useful to add another mode "--tokenize-raw" that doesn't
do that, but this seems to be the more useful of the two.
Fixes#3823.
Background fillthreads are used when we want to populate a buffer from an
external command. The most common is command substitution.
Prior to this commit, fish would spin up a fillthread whenever required.
This ended up being quite expensive.
Switch to using the iothread pool instead. This enables reusing the same
thread(s), which prevents needing to spawn new threads. This shows a big
perf win on the alias benchmark (766 -> 378 ms).
This reintroduces commits 22230a1a0d
and 9d7d70c204, now with the bug fixed.
The problem was when there was one thread waiting in the pool. We enqueue
an item onto the pool and attempt to wake up the thread. But before the
thread runs, we enqueue another item - this second enqueue will see the
thread waiting and attempt to wake it up as well. If the two work items
were dependent (reader/writer) then we would have a deadlock.
The fix is to check if the number of waiting threads is at least as large
as the queue. If the number of enqueued items exceeds the number of waiting
threads, then spawn a new thread always.
This added the function offset *again*, but it's already included in
the line for the current file.
And yes, I have explicitly tested a function file with a function
defined at a later line.
Fixes#6350
Since #6287, bare variable assignments do not parse, which broke
the "Unsupported use of '='" error message.
This commit catches parse errors that occur on bare variable assignments.
When a statement node fails to parse, then we check if there is at least one
prefixing variable assignment. If so, we emit the old error message.
See also #6347
This adds initial support for statements with prefixed variable assignments.
Statments like this are supported:
a=1 b=$a echo $b # outputs 1
Just like in other shells, the left-hand side of each assignment must
be a valid variable identifier (no quoting/escaping). Array indexing
(PATH[1]=/bin ls $PATH) is *not* yet supported, but can be added fairly
easily.
The right hand side may be any valid string token, like a command
substitution, or a brace expansion.
Since `a=* foo` is equivalent to `begin set -lx a *; foo; end`,
the assignment, like `set`, uses nullglob behavior, e.g. below command
can safely be used to check if a directory is empty.
x=/nothing/{,.}* test (count $x) -eq 0
Generic file completion is done after the equal sign, so for example
pressing tab after something like `HOME=/` completes files in the
root directory
Subcommand completion works, so something like
`GIT_DIR=repo.git and command git ` correctly calls git completions
(but the git completion does not use the variable as of now).
The variable assignment is highlighted like an argument.
Closes#6048
uClibc-ng does not expose C++11 math
functions to the std namespace, breaking
compilation. This is fine as the argument
type is double.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Improve the iothread behavior by enabling an iothread to stick around for
a while waiting for work. This reduces the amount of iothread churn, which
is useful on platforms where threads are expensive.
Also do other modernization like clean up the locking discipline and use
FLOG.
Fix 'string length: Unknown option': add `--` before $subcommand
Fix count $subcommand always = 1 with `sudo` and `doas`:
give argv as array to __fish_complete_subcommand
[ci skip]
This mostly fixes some wrong indents or replaces some stray tab indents.
I excluded alignment on purpose, because we have a whole bunch of code
that goes like
```fish
complete -c foo -n 'some-condition' -l someoption
complete -c foo -n 'some-longer-condition' -l someotheroption
```
and changing it seems like a larger thing and would include more
thrashing.
See #3622.
This would have prevented #6323.
While we don't want to pepper `command` everywhere, `psub` is kind of
a core thing, so we should try to proof it against common problems.
CMake sets these flags to sane defaults depending on which compiler
you're using, so overriding them isn't very nice.
For example:
with g++, I get
-- Debug: -g
-- RelWithDebInfo: -O2 -g -DNDEBUG
-- MinSizeRel: -O2 -g -DNDEBUG
-- Release: -O3 -DNDEBUG
and with MSVC you get something like
-- Debug: /MDd /Zi /Ob0 /Od /RTC1
-- RelWithDebInfo: /MD /Zi /O2 /Ob1 /DNDEBUG
-- MinSizeRel: /MD /Zi /O2 /Ob1 /DNDEBUG
-- Release: /MD /O2 /Ob2 /DNDEBUG
fish will react to certain variable modifications, such as "TZ." Only do
this if the main stack is modified. This has no effect now because there
is always a single stack, but will become important when concurrent
execution is supported.
This adds string-x.rst for each subcommand x of string. The main page
(string.rst) is not changed, except that examples are shown directly after
each subcommand. The subcommand sections in string.rst are created by
textual inclusion of parts of the string-x.rst files.
Subcommand man pages can be viewed with either of:
```
man string collect
man string-collect
string collect <press F1 or Alt-h>
string collect -h
```
While `string -h ...` still prints the full help.
Closes#5968
Changes identity `is` for equality `==` check. To remove python warnings when updating auto complete
```
/usr/local/Cellar/fish/3.0.2/share/fish/tools/deroff.py:770: SyntaxWarning: "is" with a literal. Did you mean "=="?
if len(comps) is 2:
/usr/local/Cellar/fish/3.0.2/share/fish/tools/deroff.py:954: SyntaxWarning: "is" with a literal. Did you mean "=="?
if len(comps) is 2:
Parsing man pages and writing completions to /Users/james/.local/share/fish/generated_completions/
6155 / 6155 : zic.8
```
This reverts commit f620ddf03b.
Setting the paste handler isn't performance-sensitive.
On the other hand setting it this way makes things less transparent,
less flexible (if e.g. a paste handler is installed while the shell is running),
and causes #6286.
Fixes#6286.
[ci skip]
This removes the explicit html coloring that was used in the tutorial.
Where necessary we just add pseudo-html like `<red>...</red>` to
explain it to the users.
I don't know how to reintroduce coloring here, but it's not super
important as the user can always just check for themselves.
See #5696
[ci skip]
Presently the completion engine ignores builtins that are part of the
fish syntax. This can be a problem when completing a string that was
based on the output of `commandline -p`. This changes completions to
treat these builtins like any other command.
This also disables generic (filename) completion inside comments and
after strings that do not tokenize.
Additionally, comments are stripped off the output of `commandline -p`.
Fixes#5415Fixes#2705
The history search logic had a not very useful "fast path" which was also
buggy because it neglected to dedup. Switch the "fast path" to just a
history search type which always matches.
Fixes#6278
PATH and CDPATH have special behavior around empty elements. Express this
directly in env_stack_t::set rather than via variable dispatch; this is
cleaner.
This adds support for `fish_trace`, a new variable intended to serve the
same purpose as `set -x` as in bash. Setting this variable to anything
non-empty causes execution to be traced. In the future we may give more
specific meaning to the value of the variable.
The user's prompt is not traced unless you run it explicitly. Events are
also not traced because it is noisy; however autoloading is.
Fixes#3427
When considering an autosuggestion from history, we attempt to validate the
command to ensure that we don't suggest invalid (e.g. path-dependent)
commands. Prior to this fix, we would validate the last command in the
command line (e.g. in `cd /bin && ./stuff` we would validate "./stuff".
This doesn't really make sense; we should be validating the first command
because it has the potential to change the PWD. Switch to validating the
first command.
Also remove some helper functions that became dead through this change.
This just makes more sense, as people don't want to enter exact
matches if they delete interactively.
It also brings it in line with "search".
Fixes#6142
Rejects #6070
Until now, something like
`math '7 = 2'`
would complain about a "missing" operator.
Now we print an error about logical operators not being supported and
point the user towards `test`.
Fixes#6096
This reorders many CHANGELOG entries.
The main idea is to keep the "NOTABLE fixes and improvements" for the headline items,
so a bunch of entries (like "Empty uvars can now be exported") are moved to more specific sections.
Other than that, there's some rewording, and the new feature flag is mentioned in Deprecations,
because that's effectively what it is.
[ci skip]
They were just wrappers around `cmake` and caused cmake with the Makefile generator
to mark in-tree builds as dirty, since it would overwrite them with its own.
Fixes#6264
MacOS Catalina apparently ships a stripped down svn that doesn't have
`svnversion`, which we use to print the revision.
For now skip the entire step to remove error spam.
Fixes#6267.
[ci skip]
Every builtin or function shipped with fish supports flag -h or --help to
print a slightly condensed version of its manpage.
Some of those help messages are longer than a typical screen;
this commit pipes the help to a pager to make it easier to read.
As in other places in fish we assume that either $PAGER or "less" is a
valid pager and use that.
In three places (error messages for bg, break and continue) the help is
printed to stderr instead of stdout. To make sure the error message is
visible in the pager, we pass it to builtin_print_help, every call of which
needs to be updated.
Fixes#6227
Some distros (Arch) use python command for Python 3, so we need to update the scripts to work with it. We cannot just switch to python3 command because MacOS does not ship it.
* functions/__fish_print_hostnames: Fix ssh_configs no values return
`string replace` not working with mutlilines variable.
So split per line first.
* functions/__fish_print_hostnames: remove quotes at `split '\n'`
"\n with quotes" will cause `string split` weird issues.
* functions/__fish_print_hostnames: using `read -alz -d \n`
Fix `$contents` issues together
Since the url is inside a AngularJS markup {{url}}, it's better to use **ng-href**.
From [AngularJS Documentation](https://docs.angularjs.org/api/ng/directive/ngHref):
<br>
"Using AngularJS markup like {{hash}} in an href attribute will make the link go to the wrong URL if the user clicks it before AngularJS has a chance to replace the {{hash}} markup with its value. Until AngularJS replaces the markup the link will be broken and will most likely return a 404 error. The ngHref directive solves this problem."
The `--entire` would enable output even though the `--quiet` should
have silenced it. These two don't make any sense together so print an
error, because the user could have just left off the `-q`.
We used to just check for the presence of "--" on the command line to
make judgements about which completions to suggest. Now, even if "--" is
present, we can still make different suggestions by taking the cursor's
position into account.
If "--" is present in the command line, it's usually safe to assume that
the user is going to want to complete a file tracked by git so let's
only suggest branches if "--" isn't present.
When there is already a "src:", we assume that it is a valid ref and
just complete "dst". This allows completion of dest if src is e.g. a
commit SHA (completing all possible refs would probably impact
performance).
See issue #3035.
This spewed errors because the `math` invocation got no second
operand:
Testing file checks/sigint.fish ... math: Error: Too few arguments
'1571487730 -'
but only if the `date` didn't do milliseconds, which is the case on
FreeBSD and NetBSD.
(also force the variable to be global - we don't want to have a
universal causing trouble here)
sys/sysctl.h is deprecated on glibc, so it leads to warnings.
According to fa4ec55c96, it was included for KERN_PROCARGS2 for
process expansion, but process expansion is gone, so it's unused now.
(there is another use of it in common.cpp, but that's only on FreeBSD)
Also 1f06e5f0b9 only included
tokenizer.h (present since the initial commit) if KERN_PROCARGS2
wasn't available, so it can't have been important.
This builds and passes the tests on:
- Archlinux, with glibc 2.30
- Alpine, with musl
- FreeBSD
- NetBSD
Universal exported variables (created by `set -xU`) used to show up
both as universal and global variable in child instances of fish.
As a result, when changing an exported universal variable, the
new value would only be visible after a new login (or deleting the
variable from global scope in each fish instance).
Additionally, something like `set -xU EDITOR vim -g` would be imported
into the global scope as a single word resulting in failures to
execute $EDITOR in fish.
We cannot simply give precedence to universal variables, because
another process might have exported the same variable. Instead, we
only skip importing a variable when it is equivalent to an exported
universal variable with the same name. We compare their values after
joining with spaces, hence skipping those imports does not change the
environment fish passes to its children. Only the representation in
fish is changed from `"vim -g"` to `vim -g`.
Closes#5258.
This eliminates the issue #5348 for universal variables.
Consider a group of short options, like -xzPARAM, where x and z are options and z takes an argument.
This commit enables completion of the argument to the last option (z), both within the same
token (-xzP) or in the next one (-xz P).
complete -C'-xz' will complete only parameters to z.
complete -C'-xz ' will complete only parameters to z if z requires a parameter
otherwise, it will also complete non-option parameters
To do so this implements a heuristic to differentiate such strings from single long options. To
detect whether our token contains some short options, we only require the first character after the
dash (here x) to be an option. Previously, all characters had to be short options. The last option
in our example is z. Everything after the last option is assumed to be a parameter to the last
option.
Assume there is also a single long option -x-foo, then complete -C'-x' will suggest both -x-foo and
-xy. However, when the single option x requires an argument, this will not suggest -x-foo.
However, I assume this will almost never happen in practise since completions very rarely mix
short and single long options.
Fixes#332
In e167714899 we allowed recursive calls
to complete. However, some completions use infinite recursion in their
completions and rely on `complete` to silently stop as soon as it is
called recursively twice without parameter (thus completing the
current commandline). For example:
complete -c su -s -xa "(complete -C(commandline -ct))"
su -c <TAB>
Infinite recursion happens because (commandline -ct) is an empty list,
which would print an error message. This commmit explicitly detects
such recursive calls where `complete` has no parameter and silently
terminates. This enables above completion (like before raising the
recursion limit) while still allowing legitimate cases with limited
recursion.
Closes#6171
This stops reading argument names after another option appears. It does not break any previous uses and in fact fixes uses like
```fish
function foo --argument-names bar --description baz
```
* `function` command handles options after argument names (Fixes#6186)
* Removed unneccesary test
Corrects #6110
BSD `seq` produces a down-counting sequence when the second argument is
smaller than the first, e.g.:
$ seq 2 1
2
1
$
While GNU `seq` produces no output at all:
$ seq 2 1
$
To accommodate for this behavior, only run `seq` when we are sure that
the second argument is greater than or equal to the first (in this case,
the second argument `line_count` should be greater than 1).
* Add completions/sfdx.fish
* completions/sfdx.fish: add completion for options
* completions/sfdx.fish: add a completion for --manifest(-x) option which need package.json
* completions/sfdx.fish: replace redundant function with already existing one
Revert "gut gpg.fish/gpg1.fish/gpg2.fish; migrate functionality to __fish_complete_gpg.fish"
This reverts commit d558218d03.
Revert "break version-specific completions out into independent function;"
This reverts commit 9160e77b01.
Revert "split gpg2- and gpg1-specific completions to conditional block"
This reverts commit a069b95f63.
* Fix default Alt+W keybinding
The old keybinding would chop off the last line of the `whatis` output
when using a multi-line prompt. This fix corrects that.
* Make variable local and remove unneeded if statement
* Test that token is non-empty
1. Added missing commands and arguments.
2. Removed alternative spelling of some commands (e.g. clear-cache|clearcache) since a choice of spelling is not really useful for completion.
3. Fixed a typo: np-ansi → no-ansi.
4. Removed redundant backslash in front of $COMPOSER_HOME.
The updated completion was initially generated using the bamarni/symfony-console-autocomplete package and then incorporated into the existing code.
Reproducer: type `: \<RET><M-p>`. This used to print an error due to builtin test receiving
too many arguments.
It looks like (commandline -j) can return multiple items, because a job can be broken up in multiple
lines terminated by backslashes.
With the new support for self-insert inserting a bound sequence,
the default binding for space as expanding abbreviations can be simplified
to just `self-insert expand-abbr`. This also fixes the bug where space
would cancel pager search.
Prior to this fix, self-insert would always wait for a new character.
Track in char_event what sequence generated the readline event, and then
if the sequence is not empty, insert that sequence.
This will support implementing the space binding via pure readline
functions.
Previously, tab-completion would move the cursor to the end of the current token, even
if no completion is inserted. This commit defers moving the cursor until we insert a completion.
Fixes#4124
I got tired of seeing ' ... ok (0 sec)' so now with GNU date/gdate
installed there is millisecond output shown. One can get rough
nanoseconds from gdate.
If interactive, `complete` commands are highlighted like they would
be if typed. Adds a little fun contrast and it's easier to read.
Moved a function out of fish_indent to highlight.h
we now print --long options for ones I arbitrarily decided
are less likely to be remembered.
Also fixed the `--wraps` items at the end not being escaped
Most of our completion scripts are written using the short options
anyhow, and this makes it less likely the output will span several
lines per command
This commit makes git completions aware of files that are both staged as renamed, and have unstaged
modifications/are deleted.
__fish_git_files now potentially prints these files twice:
$ __fish_git_files renamed modified
foo Renamed file
foo Modified file
Fixes#6031
Adds synopses for those commands missing them.
Moves all synopsis sections to code blocks. This improves the appearance, although highlighting as
fish code may not be the ideal appearance.
This sequence can be generatd by control-spacebar. Allow it to be bound
properly.
To do this we must be sure that we never round-trip the key sequence
through a C string.
By not manipulating each line or even each file at a time, we can go
back to `string` and piece together a pipeline that will execute
significantly faster than shelling out to `awk` will. This also removes
one of the few dependencies on `awk` in the codebase.
With this change, `__fish_print_hostnames` now finishes ~80% faster than
it used to a few commits back.
Reordering the `getent hosts` and read from `/etc/hosts` combined with
minimizing shelling and job invocations for parsing the output results
in a profiled and benchmarked ~42% decrease in the time it takes to run,
and that's on a machine with a very small hosts list in the first place.
This update also fixes the hadling of IPv6 addresses in the hosts
output, which were previously ignored, and ignores 127.* loopback
addresses in addition to the 0.0.0.0 address (plus adds support for
shorter IPv4 notations).
Fish completes parts of words split by the separators, so things like
`dd if=/dev/sd<TAB>` work.
This commit improves interactive completion if completion strings legitimately
contain '=' or ':'. Consider this example where completion will suggest
a:a:1 and other files in the cwd in addition to a:1
touch a:1; complete -C'ls a:'
This behavior remains unchanged, but this commit allows to quote or escape
separators, so that e.g. `ls "a:<TAB>` and `ls a\:<TAB>` successfully complete
the filename.
This also makes the completion insert those escapes automatically unless
already quoted.
So `ls a<TAB>` will give `ls a\:1`.
Both changes match bash's behavior.
Instead of warning (debug level 1), we now emit an error (debug level 0) if a known bad version of
WSL is detected. However, `FISH_NO_WSL_CHECK` can now be defined to skip both the check and the
startup message.
"space-delimited" sounds like you'd set it like `set
__fish_git_prompt_showupstream "auto verbose"`. This will not work.
It's a real actual proper list, which aren't space-delimited.
[ci skip]
fish is designed to append to the history file in most cases. However
save_internal_via_appending was never returning success, so we were
always doing the slow rewrite path. Correctly return success.
Fixes#6042
It appears Gcc 4.8 doesn't get this particular expression, so we just
revert to the old `type foo = bar` style from the new `type foo{bar}`.
Fixes#6027.
Arrow keys are often not conveniently located on keyboards, so the use of arrow keys for common keyboard shortcuts can be a turn-off for some.
I found that fish supports alternate keybindings for these cases but I didn't seem them documented in these places where the arrow keys versions are highlighted.
Meaning empty variables, command substitutions that don't print
anything.
A switch without an argument
```fish
switch
case ...
end
```
is still a syntax error, and more than one argument is still a runtime
error.
The none-argument matches either an empty-string `case ''` or a
catch-all `case '*'`.
Fixes#5677.
Fixes#4943.
This test uses universal variables, and so it can fail when run
multiple times.
It might be a good idea to do this in general, but for now let's just
try it here.
See for example: https://git-scm.com/docs/git-cherry
git cherry is quite helpful when trying to findout if merges between
branches are complete, when there were cherry-picks in addition to
merges.
Previously when propagating explicitly separated output, we would early-out
if the buffer was empty, where empty meant contains no characters. However
it may contain one or more empty strings, in which case we should propagate
those strings.
Remove this footgun "empty" function and handle this properly.
Fixes#5987
Prior to this fix, fish would attempt to react if a local fish_complete_path
or fish_function_path were set. However this has never been very well tested
and will become impossible with concurrent execution. Always use the global
values.
Soon we will have more complicated logic around whether to call tcsetpgrp.
Prepare to centralize the logic by passing in the new term owner pgrp,
instead of having child_setup_process perform the decision.
I tested this manually (`littlecheck.py -s fish=fish tests/checks/eval.fish`) from the base directory, which means I got
"tests/checks/eval", while the real test gets "checks/eval".
I then reran `make test_fishscript`, but that didn't pull in the
updated test - we should really handle that better.
This exitted if the cursor was at the end of the line as well (i.e. if
delete-char failed). That's a bit too eager.
Also documentation, which should have already been included.
I'm gonna add more tests to this and I don't want to touch the old stuff.
Notice that this needs to have the output of the complete_directories
test adjusted because this one now runs later.
That's something we should take into account in future.
Previously, elements already existing in the path variable would keep their position when the path was being constructed from the config files. This caused issues given that $PATH typically already contains "/usr/bin:/bin" when fish initializes within a macOS terminal app. In this case, these would keep their position at the front of the $PATH, even though the system path_helper configs explicitly place them _after_ other paths, like "/usr/local/bin". This would render binaries in "/usr/local/bin" as effectively "invisible" if they also happen to live in "/usr/bin" as well. This is not the intended
This change makes the __fish_macos_set_env config function emulate the macOS standard path_helper behavior more faithfully, with:
1. The path list being constructed *from scratch* based on the paths specified in the config files
2. Any distinct entries in the exist path environment variable being appended to this list
3. And then this list being used to *replace* the existing path environment variable
The result, for a vanilla fish shell on macOS, is that the $PATH is now set to:
/usr/local/bin /usr/local/sbin /usr/bin /bin /usr/sbin /sbin
Where previously it was set to:
/usr/bin /bin /usr/local/bin /usr/sbin /sbin
This new $PATH exactly matches the order of paths specified in `/etc/paths`.
We used to have a global notion of "is the shell interactive" but soon we
will want to have multiple independent execution threads, only some of
which may be interactive. Start tracking this data per-parser.
history now often writes to the history file asynchronously, but the history
test expects to find the text in the file immediately after running the
command. Hack a bit in history to make this test more reliable.
Prior to this diff, fish had different signal handling functions for
different signals. However it was hard to coordinate when a signal needed
to be the default handler, and when it was custom. In #5962 we overwrote
fish's custom WINCH handler with the default_handler when fish script asked
for WINCH to be handled.
Just have a single big signal handler function. That way it can never be
set to the wrong thing.
Fixes#5969
When executing a job, if the first process is fish internal, then have
fish claim the job's pgroup.
The idea here is that the terminal must be owned by a pgroup containing
the process reading from the terminal. If the first process is fish
internal (a function or builtin) then the pgroup must contain the fish
process.
This is a bit of a workaround of the behavior where the first process that
executes in a job becomes the process group leader. If there's a deferred
process, then we will execute processes out of order so the pgroup can be
wrong. Fix this by setting the process group leader explicitly as fish
when necessary.
Fixes#5855
This required a bit of thinking.
What we do is we have one test that fakes $HOME, and then we do the
various config tests there.
The fake config we have is reused and we exercise all of the same codepaths.
This prints a green "ok" with the duration, just like the rest of the
tests.
Note that this clashes a bit with
https://github.com/ridiculousfish/littlecheck/pull/3.
(also don't check for python again and again and again)
This is a bit weird sometimes, e.g. to test the return status (that
fish actually *returns $status*), we use a #RUN line with %fish
invoking %fish, so we can use the substitution.
Still much nicer.
The missing scripts are those that rely on config.
Especially as, in this case, the documentation is quite massive.
Caught by porting string's test to littlecheck.
See #3404 - this was already supposed to be included.
This is a nice test (ha!) for how this works and what littlecheck can
do for us.
1. Input is now the actual file, not "Standard Input" anymore. So
any errors mentioning that now include the filename.
2. Regex are really nice for filenames, but especially for line
numbers
3. It's much nicer to have the output where it's created, instead of
needing to follow three files at the same time.
$__fish_git_prompt_use_informative_chars will use the informative
chars without requiring informative mode (which is really frickin'
slow!).
See #5726.
[ci skip]
This previously effectively checked `string split ' '`s return status,
which was false if it didn't split anything. And while that should be
true if getent fails (because it should produce no output), it's also
true if it doesn't print a line with multiple aliases. Which should be
fairly typical.
Instead we use our new-found $pipestatus to check what getent returns,
in the assumption that it'll fail if it doesn't support hosts.
Follow up to 8f7a47547e.
[ci skip]
`getent hosts` is expensive-ish - ~50ms, so we don't want to run it
twice just to figure out it works.
Apparently this works everywhere but CYGWIN and possibly older
OpenBSD, but we don't want to explicitly blacklist those.
[ci skip]
25afc9b377 made this unnecessary by
having child processes wait for a signal after fork(), but this change
was later reverted. If we artificially slow down fish (e.g. with a sleep)
after the fork call, we see commands getting backgrounded by mistake.
Put back the tcsetgrp() call.
This was clearly intended for index, but because it was called "fish"
it was overwritten by the "fish" command man page.
I also added the tutorial and faq. Both of those might not be *ideal*
as man pages (the tutorial makes references to colors that won't show
up), but it's better to provide them than not.
Hat-tip to @wwared
See #5521.
[ci skip]
This makes test_low_level, test_interactive, test_invocation, and
test_fishscript independent. This allows running a smaller subset of tests.
To prevent all tests running in parallel, we also have new targets
serial_test_low_level, serial_test_interactive, etc. which have the
dependency chain that enforces serial execution.
* Prevent not-yet-loaded functions from loaded when erased
Today, `functions --erase $function` does nothing if the function
hasn't been autoloaded yet.
E.g. run, in an interactive session
> functions --erase ls
> type ls
and be amazed that it still shows our default `ls --color=auto`
wrapper function.
This seems counter-intuitive - removing a function ought to remove it,
whether it had been executed before or not.
* doc/changelog
Instead of requiring a flag to enable newline trimming, invert it so the
flag (now `--no-trim-newlines`) disables newline trimming. This way our
default behavior matches that of sh's `"$(cmd)"`.
Also change newline trimming to trim all newlines instead of just one,
again to match sh's behavior.
The `string collect` subcommand behaves quite similarly in practice to
`string split0 -m 0` in that it doesn't split its output, but it also
takes an optional `--trim-newline` flag to trim a single trailing
newline off of the output.
See issue #159.
This adds support for .check files inside the tests directory. .check
files are tests designed to be run with littlecheck.
Port printf test to littlecheck and remove the printf.in test.
Our existing .rst files have lines with trailing whitespace in them,
which I can only assume is deliberate, so update the editorconfig to
stop trimming trailing whitespace for these files.
It's always a bit annoying that `*` requires quoting.
So we allow "x" as an alternative, only it needs to be followed by
whitespace to distinguish it from "0x" hexadecimal notation.
To support distinct parsers having different working directories, we need
to keep the working directory alive, and also retain a non-path reference
to it.
Because an exported universal variable must be exported in all variable
stacks, explicit invalidation is infeasible. Switch the universal variables
to a generation count.
Prior to this fix, fish would invalidate the exported variable list
whenever an exported variable changes. However we soon will not have a
single "exported variable list." If a global variable changes, it is
infeasible to find all exported variable lists and invalidate them.
Switch to a new model where we store a list of generation counts. Every
time an exported variable changes, the node gets a new generation. If the
current generation list does not match the cached one, then we know that
our exported variable list is stale.
We also might want to remove the mailing list and/or IRC channel as I
think most devs aren't on them anymore, but let's just add the channel
that I prefer first.
[ci skip]
This was undocumented, not all that useful and potentially unwanted.
In particular it means that things like
mysql -p(read)
will still keep the password in history.
Also it allows us to simply implement asking for the history deletion
term.
See #5791.
This makes the following changes:
1. Events in background threads are executed in those threads, instead of
being silently dropped
2. Blocked events are now per-parser instead of global
3. Events are posted in builtin_set instead of within the environment stack
The last one means that we no longer support event handlers for implicit
sets like (example) argv. Instead only the `set` builtin (and also `cd`)
post variable-change events.
Events from universal variable changes are still not fully rationalized.
This cuts down on the wcs2string here by ~25%.
The better solution would be to cache narrow versions of $PATH, since
we compute that over and over and over and over again, while it rarely changes.
Or we could add a full path-cache (where which command is), but that's
much harder to invalidate.
See #5905.
This sets the explicit path to the default one, which should be okay,
since the default path never changes (not even if $XDG_CONFIG_HOME
does).
Then it saves a narrow version of that, which saves most of the time
needed to `sync` in most cases.
Fixes#5905.
Note that this isn't technically *w*util, but the differences between
the functions are basically just whether they do the wcs2string
themselves or not.
This fixes a race condition in the topic monitor. A thread may decide to
enter the wait queue, but before it does the generation list changes, and
so our thread will wait forever, resulting in a hang.
It also simplifies the implementation of the topic monitor considerably;
on reflection the whole "metagen" thing isn't providing any value and we
should just compare generations directly.
In the new design, we have a lock-protected list of current generations,
along with a boolean as to whether someone is reading from the pipe. The
reader (only one at a time) is responsible for broadcasting notifications
via a condition variable.
3.x has changed enough that knowing whether it's working in 2.7.x
doesn't help all that much anymore.
There've been 3222 commits since 2.7.1, which is about a third of the
commits to fish *total*.
[ci skip]
Not -Werror, just the warnings themselves.
Let's just see what happens. It's warning-free on my system.
Worst case we disable it again, better case we refine the set.
The enum starts at 0 (defined to be!), so we can eliminate this one.
That allows us to remove a reliance on the position of
beginning_of_line, and it would trigger a "type-limits" warning.
Also leave a comment because I actually hit that.
Someone has hit the 10MiB limit (and of course it's the number of
javascript packages), and we don't handle it fantastically currently.
And even though you can't pass a variable of that size in one go, it's
plausible that someone might do it in multiple passes.
See #5267.
If you changed $__fish_git_prompt_show_informative_status, it
triggered a variable handler, which erased the chars, but neglected to
unset $___fish_git_prompt_init, so we just kept chugging along with
empty characters.
What's the hardest thing in CS again? Cache something something?
[ci skip]
This widens the remaining ones that don't take a char
anywhere.
The rest either use a char _variable_ or __FUNCTION__, which from my
reading is narrow and needs to be widened manually. I've been unable
to test it, though.
See #5900.
This solves the main part of (careful linebreak)
issue #5900.
I'm betting all the errors that do use narrow IO are broken, including
a bunch of asserts.
This is roughly in order of importance.
Buffering is really quite nice, $pipestatus is a new thing, $PATH was
a big hubhub, people seem to like eval...
[ci skip]
3e055f does not exist. I'm assuming it's
b2a1da602f.
Unfortunately, I'm not entirely sure what kind of parser errors
weren't propagated, so this note isn't all that useful.
[ci skip]
This adds a new mechanism for logging, intended to replace debug().
The entry points are FLOG and FLOGF. FLOG can be used to log a sequence of
arguments, FLOGF is for printf-style formatted strings.
Each call to FLOG and FLOGF requires a category. If logging for a category
is not enabled, there is no effect (and arguments are not evaluated).
Categories may be enabled on the command line via the -d option.
This way we use our core file completion code, which is much more
flexible than we can easily achieve directly in script (which would
require e.g. an `expand` builtin, and case-insensitive globs).
Fixes#5896.
Now that our interactive signal handlers are a strict superset of
non-interactive ones, there is no reason to "reset" signals or take action
when becoming non-interactive. Clean up how signal handlers get installed.
The signal handlers for interactive and non-interactive SIGINT were distinct
and talked to the reader. This wasn't really justified and will complicate
having multiple threads. Unify these into a single signal handler.
is_interactive_read is a suspicious flag which prevents a call to
parser_t::skip_all_blocks from a ^C signal handler. However we end
up skipping the blocks later when we exit the read loop.
This flag seems unnecessary. Bravely remove it.
When setting a variable without a specified scope, we should give priority
to an existing local or global above an existing universal variable with
the same name.
In 16fd780484 there was a regression that
made universal variables have priority.
Fixes#5883
This displays a colored prompt, which we emulate by adding explicit
roles that are translated to css classes.
For other things, like "eror" this might not be enough because those
often need the rest of the line to still be styled, and I've not found
a way to add some explicit styling to a code block.
See #5696.
[ci skip]
Otherwise we'd undo the history search when you press e.g. execute,
which means you'd execute the search term.
Only `cancel` should walk it back, like it previously did hardcoded to
escape.
Fixes#5891.
These were inadvertently disabled by a bug which was introduced in
cd7e8f4103 . Fix the bug so the tests run
again.
They don't all pass yet; they regressed during the period they were
disabled.
This mainly is conceptually a bit simpler. The comment about making it
cheaper is entirely misplaced since this is quite far away from being
important.
Even expanding 1000 abbrs, it doesn't show up in the profile.
This was, under some circumstances, apparently off by one.
If a suggestion was really long, like
```fish
infocmp | string split , | string trim | string match -re . | while read -d = -l key val; test -z "$val"; and continue; string match -q '*%*' -- $val; and continue; test (string replace -ra '\e([\[\]]|\(B).*[\comJKsu]' '' -- a(tput $key)b) = ab; or echo $key $val; end > xterm
```
(I'm assuming longer than $COLUMNS), it would staircase like with a wrong wcwidth.
This reverts commit 15a5c0ed5f.
I noticed my debug output for 24bit color mode was garbled due to
this being wrong. I spent a little time trying to get the compiler
to tell us about these, but -Wformat doesn't do anything for wchar
printf functions, and __attribute__((format(printf, n, m))) will
cause an error with wchar_t's, so I gave up and decided to manually
check out every '%s' in the entire project. I found (only) one
more.
debug(0, "%s", wchars) will report warnings for incorrect
specifiers but debug(0, L"%s", wchars) is unable. Thus there may
be reason to prefer not using L"..." as an argument if all else
is equal and it's not necessary.
Allows `fish_indent -w **.fish` to restyle all fish files under the
current directory.
(This also has the sideeffect of reducing style.fish time by ~10s, as
we only need to invoke `fish_indent` once, instead of once per-file)
Blocks will soon need to be shared across parsers. Migrate the loop status
(like break or continue) from the block into the libdata. It turns out we
only ever need one, we don't need to track this per-block.
Make it an enum class.
Brace expansion with single words in it is quite useless - `HEAD@{0}`
expanding to `HEAD@0` breaks git.
So we complicate the rule slightly - if there is no variable expansion
or "," inside of braces, they are just treated as literal braces.
Note that this is technically backwards-incompatible, because
echo foo{0}
will now print `foo{0}` instead of `foo0`. However that's a
technicality because the braces were literally useless in that case.
Our tests needed to be adjusted, but that's because they are meant to
exercise this in weird ways.
I don't believe this will break any code in practice.
Fixes#5869.
Prior to this fix, a function_block stored a process_t, which was only used
when printing backtraces. Switch this to an array of arguments, and make
various other cleanups around null terminated argument arrays.
We previously checked if fish_mode_prompt existed as a function, but
that's a bad change for those who already set it to an empty function
to have a mode display elsewhere.
Updated widechar_width takes care of it.
Technically, this does ~3 comparisons more per-character (because it
checks variation selectors and such), but that shouldn't really matter.
get_current_winsize() is intended to be lazy. It does the following:
1. Gets the termsize from the kernel
2. Compares it against the current value
3. If changed, sets COLUMNS and LINES variables
Upon setting these variables, we notice that the termsize has changed
and invalidate the termsize. Thus we were doing this work multiple times
on every screen repaint.
Put back an old hack that just marked the termsize as valid at the end
of get_current_winsize().
This just sets some special characters that we use in the reader, so
it only needs to be done before the reader is set up.
Which, as it stands, is in env_init().
This stops trying to see if the previous line is wider if it is a
prefix of the current one.
Which turns out to be true often enough that it's a net benefit.
This passes character width as an argument for a few functions.
In particular, it hardcodes a width of "1" for a space literal.
There's no reason to compute wcwidth for the length of the prompt.
This measured *all* the characters on the commandline, and saved all
of them in another wcstring_list_t, just to then do... nothing with
that info.
Also, it did wcslen for something that we already have as wcstring,
reserved a vector and did a bunch of work for autosuggestions that
isn't necessary if we have more than one line.
Instead, we do what we need, which is to figure out if we are
multiline and how wide the first line is.
Fixes#5866.
line_shared_prefix explains in its comment that
> If the prefix ends on a combining character, do not include the
previous character in the prefix.
But that's not what it does.
Instead, what it appears to do is to return idx for *every* combining
mark. This seems wrong to begin with, and it also requires checking
wcwidth for *every* character.
So instead we don't do that. If we find the mismatch, we check if it's
a combining mark, and then go back to the previous character (i.e. the
one before the one that the combining mark is for).
My tests found no issues with this, other than a 20% reduction in
pasting time.
The old commit #3f820f0 "Disable ONLCR mapping of NL output to CR-NL"
incorrectly used c_iflag instead of c_oflag, and I copied that error
in my patch. Fixed that. However, there seems to be other problems
trying to use "\x1B[A", which I have not tried to debug, so comment that out.
(However, #3f820f0 seems to mostly work if we fix it to use c_oflag.)
This read something like `o=!_validate_int`, and the flag modifier
reading kept the pointer after the `!`, so it created a long flag
called `_validate_int`, which meant it would not only error out form
```fish
argparse 'i=!_validate_int' 'o=!_validate_int' -- $argv
```
with "Long flag '_validate_int' already defined", but also set
$_flag_validate_int.
Fixes#5864.
We were flip-flopping between the two terms, so we now use one. We
still mention "array" in the chapter, and it's still `read --array`,
though.
Fixes#5846.
[ci skip]
This was quite famously rather complicated.
We drop a bunch of cases - we can't handle tmux-starting-terminals
100% accurately, so we just don't try. It should be quite rare that
somebody starts a different terminal from tmux.
We drop the `tput` since it is useless (like terminfo in general for
feature-detection, because everyone claims to be xterm).
So we just check if we are in konsole, iTerm, vte or genuine-xterm.
Fixes#3696.
See #3481.
As mentioned in #2900, something like
```fish
test -n "$var"; and set -l foo $var
```
is sufficiently idiomatic that it should be allowable.
Also fixes some additional weirdness with semicolons.
This runs build_tools/style.fish, which runs clang-format on C++, fish_indent on fish and (new) black on python.
If anything is wrong with the formatting, we should fix the tools, but automated formatting is worth it.
This removes semicolons at the end of the line and collapses
consecutive ones, while replacing meaningful semicolons with newlines.
I.e.
```fish
echo;
```
becomes
```fish
echo
```
but
```fish
echo; echo
```
becomes
```fish
echo
echo
```
Fixes#5859.
This was a sort of side channel that was only used to propagate redraws
after universal variable changes. We can eliminate it and handle these
more directly.
tsan does funny things to signals, preventing signals from being delivered
in a blocking read. Switch the topic monitor to non-blocking reads under
tsan.
This completed the commandline with options removed, which looked like
env VAR=VAL command option
Which didn't really actually work.
Fixes#5856.
[ci skip]
fish_indent_lexer formats lines not starting with a prompt indicator
as output, as long as there is a prompt indicator elsewhere.
So these tags are useless and wrong.
See #5696.
[ci skip]
This keeps all unknown options in $argv, so
```fish
argparse -i a/alpha -- -a banana -o val -w
```
results in $_flag_a set to banana, and $argv set to `-o val -w`.
This allows users to use multiple argparse passes, or to simply avoid
specifying all options e.g. in completions - `systemctl` has 46 of
them, most not having any effect on the completions.
Fixes#5367.
This cleans up how functions are stored and autoloaded. It eliminates the
recursive lock. Instead there is a single normal owning_lock that protects
the entirety of the function data. Autoloading is re-implemented via the
new autoloader_t.
autoloader_t will be the reimplementation of autoloading. Crucically it no
longer manages any locking or loading itself; instead all locking and loading
is performed by clients. This makes it easier to test and helps limit its
responsibilities.
autoloading has a "feature" where functions are removed in an LRU-fashion.
But there's hardly any benefit in removing autoloaded functions. Just stop
doing it.
* Add speedtest-cli/speedtest completion
Added a completion file for speedtest-cli utility (https://github.com/sivel/speedtest-cli) as shipped from various package repositories.
* added no-files parameter
* Remove inheritance to speedtest
* Create speedtest.fish
This is a long-standing issue with how `complete --do-complete` does
its argument parsing: It takes an optional argument, so it has to be
attached to the token like `complete --do-complete=foo` or (worse)
`complete -Cfoo`.
But since `complete` doesn't take any bare arguments otherwise (it
would error with "too many arguments" if you did `complete -C foo`) we
can just take one free argument as the argument to `--do-complete`.
It's more of a command than an option anyway, since it entirely
changes what the `complete` call _does_.
This command can be used to "`cat`" the contents of `$path` as of `$rev`.
These are "silent" completions, e.g. while this adds a completion for
`git show master:foo`, the completions for `git show <TAB>` are not
affected; these "advanced" completions kick in only after at least
`git show master:<TAB>` to prevent completion pollution or slowing down
tab completions in the typical case (as this would cause each valid and
possibly unique $rev completion result to complete to `n*$rev`
completions for *n* files.
[ci skip]
Dealing with macOS output in a fast manner using `string` is surprisingly hard, given that it features lines like
gls(1), ls(1) - list directory contents
Printing the "gls" with the description and the "ls" with the description requires a `while read` loop, and that's too slow.
This reverts commit 7784a5f23c.
[ci skip]
* Some comment fixes and renaming of is_iterm2_escape_seq.
The comment for is_iterm2_escape_seq incorrectly says "CSI followed by ]".
This is wrong, because CSI is ESC followed by [ (or the seldom-used 0x9b).
The procedure actually matches Operating System Command (OSC) escape codes.
Since there is nothing iterm2-specific about OSC, is_osc_escape_seq
would be a better name.
Also s_desired_append_char documents a non-existent parameter.
* Update broken iterm2 url in comment.
This was added in 04a96f6 but not strictly required to fix#5803
(verified), with the intention of hiding invisible background jobs
(created by invoking a function within a pipeline) from the user, but
that also broke intentionally created jobs from displaying as well.
I'm thinking it can't be done without keeping track of caller context vs
job context.
Closes#5824.
env_scoped_t lives between environment_t and env_stack_t.
It represents the read-only logic of env_stack_t and will be used to back
the new environment snapshot implementation.
These tests used raw, unescaped parentheses to perform `test` logical
grouping, but the test failures weren't caught because the parser
evaluation errors were not being propagated (fixed in bdbd173e).
It was unconditionally returning `parse_execution_success`. This was
causing certain parser errors to incorrectly return after evaluation
with `$status` equal to `0`, as reported after `eval`, `source`, or
sub-`fish` execution.
Prior to this change, fish used a global flag to decide if we should check
for changes to universal variables. This flag was then checked at arbitrary
locations, potentially triggering variable updates and event handlers for
those updates; this was very hard to reason about.
Switch to triggering a universal variable update at a fixed location,
after running an external command. The common case is that the variable
file has not changed, which we can identify with just a stat() call, so
this is pretty cheap.
This reverts commit cdce8511a1.
This change was unsafe. The prior version (now restored) took the lock and
then copied the data. By returning a reference, the caller holds a
reference to data outside of the lock.
This function isn't worth optimizing. Hardly any functions use this
facility, and for those that do, they typically just capture one or two
variables.
* Convert `function_get_inherit_vars()` to return a reference to the
(possibly) existing map, rather than a copy;
* Preallocate and reuse a static (read-only) map for the (very) common
case of no inherited vars;
* Pass references to the inherit vars map around thereafter, never
triggering the map copy (or even move) constructor.
NB: If it turns out the reference is unsafe, we can switch the inherit vars
to be a shared_ptr and return that instead.
Universal newlines behaves differently between Python 2.7 and 3.x,
leading to problems when running Sphinx with Python 2.7.
fish_indent always uses \n, so there's no need to use universal newline
detection.
This also allows full UTF-8 in documentation sources.
Closes#5808.
I did not realize builtins could safely call into the parser and inject
jobs during execution. This is much cleaner than hacking around the
required shape of a plain_statement.
* Honour `dirprev` scope
Honour the scope of the `dirprev` variable if it is universal
and avoid to shadow it with a global. This enables to share
the `cd` history between sessions.
* Honor dirnext and __fish_cd_direction scope
If these variables exist in the universal scope, do not shadow them
- fix the carat position expanding e.g. `command $,`
- improve the error reporting for not-allowed command subtitutions
by figuring out where the expansion failed instead of using
SOURCE_LOCATION_UNKNOWN
- allow nullptr for parse_util_licate_brackets_range() out_string
argument if we don't need it to do any work.
Fixes#5812
`eval` has always been implemented as a function, which was always a bit
of a hack that caused some issues such as triggering the creation of a
new scope. This turns `eval` into a decorator.
The scoping issues with eval prevented it from being usable to actually
implement other shell components in fish script, such as the problems
described in #4442, which should now no longer be the case.
Closes#4443.
While `eval` is still a function, this paves the way for changing that
in the future, and lets the proc/exec functions detect when an eval is
used to allow/disallow certain behaviors and optimizations.
This adds an option --print-rusage-self to the fish executable. When set,
this option prints some getrusage stats to the console in a human-readable
way. This will be used by upcoming benchmarking support.
Followup to 394623b.
Doing it in the parser meant only top-level jobs would be reaped after
being `disown`ed, as subjobs aren't directly handled by the parser.
This is also much cleaner, as now job removal is centralized in
`process_clean_after_marking()`.
Closes#5803.
This prevents the `disown` builtin from directly removing jobs out of
the jobs list to prevent sanity issues, as `disown` may be called within
the context of a subjob (e.g. in a function or block) in which case the
parent job might not yet be done with the reference to the child job.
Instead, a flag is set and the parser removes the job from the list only
after the entire execution chain has completed.
Closes#5720.
This lets non-developers simply `cd` into the fish source directory and
execute `make` to build the project. The Makefile searches for CMake and
hands over the build to it if it is available, otherwise an error
message is emitted. All dependency checking is left to CMake.
Non-fish-devs shouldn't have to concern themselves with what build
system fish developers have chosen, and building a random C++ project
should not be a chore in familiarizing one's self with all the various
build platforms out there.
CMake is instructed to use `ninja` if it is available, otherwise the
standard Unix Makefiles generator option is used.
(This has already been the behavior on BSDs since CMake was adopted.)
There's an explicit label for "cartesian-product", but the title is
"Cartesian Product*s*". So linking via with `thing <#link>`_ links the
title, so without the "s" it doesn't work.
From what I know, linking via :ref:`thing <label>` is preferred and
works better with other exports and across files?
I think I should take a doc holiday.
[ci skip]
No longer uses global vars to cache set_color output, this was
from before set_color was a builtin, it is pointless now.
This is also a prompt from before we had bright named colors,
and it appears it was relying on -o red to get bright red.
so use brred, etc.
This merges a bunch of changes that migrate logic from env.cpp to a new file
env_dispatch.cpp. env_dispatch is concerned with dispatching changes to
variables, while env.cpp is the "core."
When popping a scope from the environment stack, we currently do a lot of
nonsense like looking for changed curses variables. We want to centralize
this in env_stack_t so that it can be migrated to the env_dispatch logic.
Move this logic up one level in preparation for doing that.
This new file is supposed to encapsulate all of the logic around
reacting to variable changes, as opposed to the environment core.
This is to help break up the env.cpp monolith.
Prior to this fix, a job would only inherit a pgrp from its parent if the
first command were external. There seems to be no reason for this
restriction and this causes tcsetgrp() churn, potentially cuasing SIGTTIN.
Switch to unconditionally inheriting a pgrp from parents.
This should fix most of #5765, the only remaining question is
tcsetpgrp from builtins.
Prior to this fix, in every call to job_continue, fish would reclaim the
foreground pgrp. This would cause other jobs in the pipeline (which may
have another pgrp) to receive SIGTTIN / SIGTTOU.
Only reclaim the foreground pgrp if it was held at the point of job_continue.
This partially addresses #5765
In tests we would like to arrange for an executable to invoke certain
system calls, e.g. to claim or relinquish control of the terminal. This is
annoying to do portably via e.g. perl. fish_test_helper is a little
program where we can add custom commands to make it act in certain ways.
This adds a reference to one specific FAQ, so it adds a label for that
one question. It does not add the rest, because they currently aren't
linked. If you add a reference to an FAQ, you should add the label as
well.
[ci skip]
We don't refer to "readline functions" anywhere else, and "injecting"
them "into the reader" is an overly jargony way of expressing it that
only makes sense to someone familiar with the internals. And even then
the term "readline" is already taken by the "readline" library, used
by bash et al, but not by us.
So we pick the term "input functions", like we did in bind.
See https://stackoverflow.com/questions/55542839/what-does-commandline-f-repaint-in-fish-shell/55543411#55543411.
[ci skip]
This set the term modes to the shell-modes, including disabling
ICRNL (translating \cm to \cj) and echo.
The rationale given was that `reader_interactive_init()` would only be
called >= 250ms later, which I _highly_ doubt considering fish's total
startup time is 8ms for me.
The main idea was that this would stop programs like tmuxinator that
send shortcuts early from failing _iff_ the shortcut was \cj, which
also seems quite unusual.
This works both with `rm -i` and `read` in config.fish, because `read`
explicitly calls `reader_push`, which then initializes the shell modes.
The real fix would involve reordering our init so we set up the
modesetting first, but that's quite involved and the remaining issue
should barely happen, while it's fairly common to have issues with a
prompt in config.fish, and the workaround for the former is simpler, so let's leave it for now.
Partially reverts #2578.
Fixes#2980.
Pursuant to 0be7903859, there still
remained one issue with the test when run from within a symlinked
directory after fish gained support for cding into symlinks.
This change should make the test function OK both when the tests are run
out of a PWD containing a symlink in its hierarchy and when run
otherwise.
Putting larger members before smaller ones will reduce structure
sizes. bools are 1 byte. on 64bit systems I think they reduced:
wgetopt.h:46: 64 to 56 bytes
builtin_history.cpp:30: 48 to 32 bytes
builtin_status.cpp:91: 32 to 24 bytes
tinyexpr.cpp:69: 40 to 32 bytes
The data stored in these containers is small enough that it is worth
creating distinct sets for each lookup.
In a microbenchmark of these changes, the single-lookup version of the
function with lookups gated on the length of input (bypassed entirely if
the input is longer than the longest key in the container) provided a
1.5x-3.5x speedup over the previous implementation.
Additionally, as the collections are static and their contents are never
modified after startup, it makes no sense to continously calculate the
location of and allocate an iterator for the `!= foo.end()` comparison;
the end iterator is now statically cached.
I'm not expecting massive speed gains out of this change, but the parser
does perform enough of these to make it worth optimizing in this way.
This reverts commit 7a74198aa3.
Believe it or not this commit actually increased copying. When accepting
a value you know you're going to take ownership of, just accept it by
value; then temporaries can invoke the move ctor and blah blah blah.
We really need a lightweight refcounted pass-by-value string to make this
less error prone.
If we switch the bind mode, we add a "force-repaint" there just to
redraw the mode indicator.
That's quite wasteful and annoying, considering that sometimes the prompt can take
half a second.
So we add a "repaint-mode" function that just reexecutes the
mode-prompt and uses the cached values for the others.
Fixes#5783.
As it turns out it didn't work much better, and it fell behind in
support when it comes to things that wcwidth traditionally can't
express like variation selectors and hangul combining characters, but
also simply $fish_*_width.
I've had to tell a few people now to rebuild with widecharwidth after
sending them on a fool's errand to set X variable.
So keeping this option is doing our users a disservice.
* Add "expand-abbr" bind function
This can be used to explictly allow expanding abbreviations.
* Make expanding abbr explicit
NOTE: This accepts them for space only, we currently also do it for \n
and \r.
* Remove now dead code
We no longer trigger an abbr implicitly, so we can remove the code
that does it.
* Fix comment
[ci skip]
This isn't officially supported (yet?), and it's currently broken.
It doesn't include ssl certificates,
and I can't see a way to add them or disable verification before it attempts to clone the git repo.
Ironically:
[ci skip]
Apparently an anonymous hyperlink looks like `__something__`.
I had to find this by deleting parts of the document and building to
narrow it down until I had the line, because sphinx wouldn't give a
line number.
See #5696.
[ci skip]
This was:
- Some `` mismatches - it's "``something``", not "``something`".
- Some "explicit targets", which IMHO are quite a misfeature - `word
<link>`_ has to be unique, which I don't see a usecase for. Instead
use `word <link>`__, with a double-underscore at the end.
- One case of `||` which I just removed
See #5696.
[ci skip]
This both formats it as a code-block, and adds the synopsis of each
subcommand to the corresponding section again so you don't need to
scroll back-and-forth so much.
[ci skip]
Apparently there must indeed be a toctree somewhere in the document to
get the links to the other docs to show up.
Even a ":hidden:" toctree doesn't help - that just leads to an empty
toc in the sidebar (no idea yet where that's defined!).
I've added it to the end so it's not that weird "Commands" section in
the middle.
[ci skip]
The final test in `realpath.in` was based on the no-longer-valid
assumption that $PWD cannot be a symlink. Since the recent changes in
fish 3.0 to allow `cd`ing into "virtual" directories preserving symlinks
as-is, when `make test` was run from a path that contained a symlink
component, this test would fail the `pwd-resolved-to-itself` check.
As the test is not designed to initialize then cd into an absolute path
guaranteed to not be symbolic, so this final check is just wrong.
Directly access the job list without the intermediate job_iterator_t,
and remove functions that are ripe for abuse by modifying a local
enumeration of the same list instead of operating on the iterators
directly (e.g. proc.cpp iterates jobs, and mid-iteration calls
parser::job_remove(j) with the job (and not the iterator to the job),
causing an invisible invalidation of the pre-existing local iterators.
This has been driving nuts for years. The output of the diff emitted
when a test fails was always reversed, because the diff tool is called
with `${difftool} ${new} ${old}` so all the `-` and `+` contexts are
reversed, and the highlights are all screwed up.
The output of a `make test` run should show what has changed from the
baseline/expected, not how the expected differs from the actual. When
considered from both the perspective of intentional changes to the test
outputs and failed test outputs, it is desirable to see how the test
output has changed from the previously expected, and not the other way
around.
(If you were used to the previous behavior, I apologize. But it was
wrong.)
It's not _perfect_, but should hopefully ease the introduction a
teensy bit.
We use `timedatectl` because it's a reasonably simple command that
still uses subcommands and some generated candidates.
[ci skip]
This printed weird things like
```fish
$ functions -x
functions: Unknown option '-x'
(Type 'help functions' for related documentation)
```
Instead, let's make it
```fish
$ functions -x
functions: Unknown option '-x'
(Type 'help functions' for related documentation)
```
This was printed basically everywhere.
The user knows what they executed on standard input.
A good example:
```fish
set c (subme 513)
```
used to print
```
fish: Too much data emitted by command substitution so it was discarded
set -l x (string repeat -n $argv x)
^
in function 'subme'
called on standard input
with parameter list '513'
in command substitution
called on standard input
```
and now it is
```
fish: Too much data emitted by command substitution so it was discarded
set -l x (string repeat -n $argv x)
^
in function 'subme' with arguments '513'
in command substitution
```
See #5434.
Now:
```
cd: Unknown option '-r'
~/dev/fish-shell/share/functions/cd.fish (line 40):
builtin cd $argv
^
in function 'cd' with arguments '-r'
in function 'f'
in function 'd'
in function 'b' with arguments '-1q --wurst'
in function 'a'
called on standard input
```
See #5434.
This printed things like
```
in function 'f'
called on standard input
in function 'd'
called on standard input
in function 'b'
called on standard input
in function 'a'
called on standard input
```
As a first step, it removes the empty lines so it's now
```
in function 'f'
called on standard input
in function 'd'
called on standard input
in function 'b'
called on standard input
in function 'a'
called on standard input
```
See #5434.
This switches env_var_t to be an immutable value type, and stores its
contents via a shared_ptr. This eliminates string copying when fetching
env_var_t values.
If a function process is deferred, allow it to be unbuffered.
This permits certain simple cases where functions are piped to external
commands to execute without buffering.
This is a somewhat-hacky stopgap measure that can't really be extended
to more general concurrent processes. However it is overall an improvement
in user experience that might help flush out some bugs too.
In a job, a deferred process is the last fish internal process which pipes
to an external command. Execute the deferred process last; this will allow
for streaming its output.
This searched for package.json in any parent, so just like finding
.git and .hg directories it _needs_ to use the physical pwd because
that's what git/hg/yarn use.
In general, if you do _any_ logic on $PWD, it should be the physical
path. Logical $PWD is basically only good for display and cd-ing
around with symlinks.
[ci skip]
This merges a bunch of changes that clean up how the reader loop and input
works.
Prior to this fix, we abused wchar_t by cramming readline functions into
"private" regions. Readline functions then were further abused with
meta-readline functions like R_NULL or R_TIMEOUT.
This fix introduces a new type char_event_type_t which wraps up the "meta"
character types. A char event may be a null (try again), timeout, readline,
or real input character. These are all distinct values.
The reader loop is then refactored to handle these cases separately.
I believe this was selected to be artificially low for the sake
of it displaying well in prompts. But people should expect to get
the same output as can be gotten from `hostname`.
Fixes#5758
This called `eval $fish_browser --version` to figure out if it is
lynx.
That's really not worth it just to support edge-cases using a rather
unusual browser, to work around a bug in it.
Instead we just see if the browser name starts with "lynx", which
should work in 99.9% of cases.
This merges in a number of improvements specifically aimed at console
sessions (i.e. using fish at the tty, not over SSH or in an X-based
terminal emulator). When a console session is detected, the system
wcwidth is used to line up width info between fish and the system tty,
and only simple characters are used as symbols.
Tested under Linux, macOS, FreeBSD, Solaris, Cygwin, WSL, and others.
Closes#5552. Ref #789, #3672.
The code already allowed for variable width (multicell) *display* of the
newline omitted character, but there was no way to define it as being
more than one `wchar_t`.
This lets us use a string on console sessions (^J aka newline feed)
instead of an ambiguous character like `@` (used in some versions of
vim for ^M) or `~` (what we were using).
The system version of `wcwidth()` reflects the capabilities of the
system's own virtual terminal's view of the width of the character in
question, while fish's enhanced version (`widechar_wcwidth`) is much too
smart for most login terminals, which generally barely support anything
beyond ASCII text.
If, at startup, it is detected that we are running under a physical
console rather than within a terminal emulator running in a desktop
environment, take that as a hint to use the system-provided `wcwidth`.
The commit began passing the length of the wide string rather than the
length of the narrowed string after conversion via `wcstombs`. We *do*
have the actual length, but it's not (necessarily) the same as the
original value. We need to pass the result of `wcstombs` instead.
This created another local version of the variable just for the if-block.
Can't say I love the space prefix, but then I think we have too many
of these modes anyway.
If you use these to figure out if there _are_ staged files, or dirty
or whatever, you currently need to check the output, which relies on
the configured character.
Instead, we let them also return a useful status.
Notably, this is *not* simply the status of the git call.
__fish_git_prompt_X returns 0 if the repo is X.
This works for untracked, but the "diff" things return 1 if there is a
diff, so we invert the status for them.
See #5748.
[ci skip]
POSIX dictates here that incomplete conversions, like in
printf %d\n 15.2
or
printf %d 14g
are still printed along with any error.
This seems alright, as it allows users to silence stderr to accept incomplete conversions.
This commit implements it, but what's a bit weird is the ordering between stdout and stderr,
causing the error to be printed _after_, like
15
14
15.1: value not completely converted
14,2: value not completely converted
but that seems like a general issue with how we buffer the streams.
(I know that nonfatal_error is a copy of most of fatal_error - I tried
differently, and va_* is weird)
Fixes#5532.
Before this change, - was sorted with other punctuation before
A-Z. Now, it sorts above the rest of the characters.
This has a practical effect on completions, where when there are
both -s and --long with the same description, the short option
is now before the long option in the pager, which is what is now
selected when navigating `foo -<TAB>`. The long options can be
picked out with `foo --<TAB>`. Before, short options which
duplicated a long option literally could not be selected by
any means from the pager.
Fixes#5634
This tweaks wcsfilecmp such that certain punctuation characters will
come after A-Z.
A big win with `set <TAB>` - the __prefixed fish junk now comes
after the stuff users should care about.
Classic case of not seeing `and` as a new command:
`__fish_git_using_command config and anotherthing`
causes `and anotherthing` to be passed as arguments to
`__fish_git_using_command` instead of being executed.
[ci skip]
A function file for a function used only by one completion (and
unlikely to be used anywhere else).
If another user shows up, we can move it out again.
Part of #5279
[ci skip]
This disables an extra round of escaping in the `string replace -r`
replacement string.
Currently, to add a backslash to an a or b (to "escape" it):
string replace -ra '([ab])' '\\\\\\\$1' a
7 backslashes!
This removes one of the layers, so now 3 or 4 works (each one escaped
for the single-quotes, so pcre receives two, which it reads as one literal):
string replace -ra '([ab])' '\\\\$1' a
This is backwards-incompatible as replacement strings will change
meaning, so we put it behind a feature flag.
The name is kinda crappy, though.
Fixes#5474.
As a simple replacement for `wc -l`.
This counts both lines on stdin _and_ arguments.
So if "file" has three lines, then `count a b c < file` will print 6.
And since it counts newlines, like wc, `echo -n foo | count` prints 0.
Mostly related to usage _(L"foo"), keeping in mind the _
macro does a wcstring().c_str() already.
And a smattering of other trivial micro-optimizations certain
to not help tangibly.
This should be the last call to `grep` outside of a script
specifically related to `grep`.
(With the exception of `zpool`, which I've already written, but which
will probably be merged later)
Similar to the last commit, only for the in-terminal-paste stuff.
Also cleans up the comments on bracketed paste a bit - nobody has
stepped forward to report problems with old emacsen or windows, so
there's no need for a TODO comment.
See #4327.
If we're at the beginning of the commandline, we trim leading whitespace so we don't trigger histignore.
Since that's the main issue of problems with histignore:
Closes#4327.
C++11 provides std::min/std::max which we're using all over,
obviating the need for our own templates for this.
util.h now only provides two things: get_time and wcsfilecmp.
This commit removes everything that includes it which doesn't
use either; most because they no longer need mini or maxi from
it but some others were #including it unnecessarily.
Hangul uses three codepoints to combine to one glyph. The first has a
width of 2 (like the final glyph), but the second and third were
assigned a width of 1, which seems to match EastAsianWidth.txt:
> 1160..11FF;N # Lo [160] HANGUL JUNGSEONG FILLER..HANGUL JONGSEONG SSANGNIEUN
Instead, we override that and treat the middle and end codepoint as combiners,
always, because there's no way to figure out what the terminal will
think and that's the way it's supposed to work.
If they stand by themselves or in another combination, they'll indeed
show up with a width of 1 so we'll get it wrong, but that's less
likely and not expressible with wcwidth().
Fixes#5729.
This only did prefix matching, which is generally less useful.
All existing users _should_ be okay with this since they want to
provide completions.
Fixes#5467.
Fixes#2318.
Reverts 71329a250b.
That tried to fix problems with pkgconfig by not recreating it.
Instead, use the function we already have for not trying too hard to
create a directory.
Fixes#5735.
This addresses a few places where -Wswitch-enum showed one or two missing
case's for enum values.
It did uncover and fix one apparent oversight:
$ function asd -p 100
echo foo
end
$ functions --handlers-type exit
Event exit
asd
It looks like this should be showing a PID before 'asd' just like
job_exit handlers show the job id. It was falling
through to default: which just printed the function name.
$ functions --handlers-type exit
Event exit
100 asd
Also prevents file completions where they are not approprite, and
additionally shortened the descriptions to fit in two pager columns
in an 80-wide terminal for some platforms.
Apparently if you install gnu coreutils on OpenBSD, the tools are
g-prefixed. So we definitely want to just alias that rather than
provide our lousy shell script implementation.
Apparently that's actually faster than jq, and it's more likely to be
installed.
Also it should convince the arch packager to remove the jq dependency.
The indentation is weird, though.
[ci skip]
+ tweaks for Linux: shorter descriptions, suppress file completions
+ Add correct completions for macOS, NetBSD, FreeBSD, OpenBSD, DragonFly
+ Solaris dmesg has no options, so complete nothing there
25d83ed0d7 (included in 3.0.0) added a `string` check that
did not use `--`, so negative numbers were interpreted as options.
Apparently nobody is using this.
(Again, this is for the `seq` fallback used on OpenBSD)
This tried to skip conversion if the locale had MB_CUR_MAX == 1, but
in doing so it just entered an infinite recursion (because
writestr(wchar_t*) called writestr(wchar_t*)).
Instead, just let wcstombs handle it.
Fixes#5724.
Since Unicode 9, the width of some characters changed to 2.
Depending on the system, it might have support for it, or it might
not.
Instead of hardcoding specific glibc etc versions, we check what the
system wcwidth says to "😃", U+1F603 "Grinning Face With Big Eyes".
The intention is to, in most cases, make setting $fish_emoji_width
unnecessary, but since it sets the "guessed_emoji_width", that variable still takes precedence if it is set.
Unfortunately this approach has some caveats:
- It relies on the locale being set to a unicode-supporting one.
(C.UTF-8 is unfortunately not standard, so we can't use it)
- It relies on the terminal's wcwidth having unicode9 support IFF the
system wcwidth does.
This is like #5722, but at runtime.
The additional caveat is that we don't try to achieve a unicode
locale, but since we re-run the heuristic when the locale changes (and
we try to get a unicode locale), we should still often get the correct
value.
Plus if you use a C locale and your terminal still displays emoji,
you've misconfigured your system.
Fixes#5722.
These aren't perfect, but the tool is pretty much hostile to proper
completions - it includes a "--machine-readable" option, but `vagrant
global-status --machine-readable` prints great output like
```
1551816037,,ui,info,id
1551816037,,ui,info,name
1551816037,,ui,info,provider
1551816037,,ui,info,state
1551816037,,ui,info,directory
1551816037,,ui,info,
1551816037,,ui,info,-------------------------------------------------------------------------
1551816037,,ui,info,d3ea265
1551816037,,ui,info,default
1551816037,,ui,info,virtualbox
1551816037,,ui,info,poweroff
1551816037,,ui,info,/home/alfa/dev/oi-userland
1551816037,,ui,info,
1551816037,,ui,info,fdf42c4
1551816037,,ui,info,default
1551816037,,ui,info,virtualbox
1551816037,,ui,info,poweroff
1551816037,,ui,info,/home/alfa/dev/vagrant/NetBSD
1551816037,,ui,info,
1551816037,,ui,info,f8f6eff
1551816037,,ui,info,default
1551816037,,ui,info,virtualbox
1551816037,,ui,info,poweroff
1551816037,,ui,info,/home/alfa/dev/vagrant/fedora
1551816037,,ui,info,
1551816037,,ui,info, \nThe above shows information about all known Vagrant environments\non this machine. This data is cached and may not be completely\nup-to-date (use "vagrant global-status --prune" to prune invalid\nentries). To interact with any of the machines%!(VAGRANT_COMMA) you can go to that\ndirectory and run Vagrant%!(VAGRANT_COMMA) or you can use the ID directly with\nVagrant commands from any directory. For example:\n"vagrant destroy 1a2b3c4d"
```
and still takes 500ms to do so. The actual information is in a json
file, which we can't expect to read, and it doesn't have linebreaks or
such which we could use to hack-parse it.
So this is the best we can do for the most important bits (the
machineids), so let's just add this as-is.
[ci skip]
`ipset list --name` is a privileged operation, and it prints an
"Operation not permitted" error when done as a normal user.
What's worse, this did it on loading (the command substitution wasn't
quoted), so we'd print the error as soon as you did `ipset `.
Only do the operation when necessary, and don't print the error.
This'll effectively only make it work for root shells (not e.g. `sudo
ipset`), but I don't want to sprinkle `sudo` in the completion.
(Also why does listing stuff require root? That's not how it works
e.g. for ips. But I don't actually know what ipset is for, so maybe
there is a good reason.)
[ci skip]
Prior to this fix, the wait command used waitpid() directly. Switch it to
calling process_mark_finished_children() along with the rest of the job
machinery. This centralizes the waitpid call to a single location.
It turns out that `string split0` didn't actually ever do any
splitting. The arg_iterator_t already split stdin on NUL, and split0 just
performed an additional search that could never succeed (since
arguments from argv already can't contain NUL).
Let the arg_iterator_t not perform any splitting if asked, and then
let split0 split in 0.
One slight wart is that split0 ignores a trailing NUL, which normal
split doesn't.
Fixes#5701.
That seems suspect.
It removes files starting with "# Autogenerated", but those files
usually do not show up in ~/.config - they're in ~/.local/share.
So let's be careful and not mess with the user's config.
(I'm pretty sure that the previous commit re-enabled cleanup as the
`~` was quoted before then)
[ci skip]
This did `argparse`, but only handled "--help". Any other options
would be ignored.
Instead, we just pass all the options through to python, and that'll
display help if needed.
This allows passing e.g. `--verbose 1` to help with debugging.
[ci skip]
This is another case where we used pid when we meant pgroup.
Since 55b3c45f95, the assumption that
both are the same no longer holds in all cases, so this check was wrong.
Might fix#5663.
In fish we play fast and loose with status codes as set directly (e.g. on
failed redirections), vs status codes returned from waitpid(), versus the
value $status. Introduce a new value type proc_status_t to encapsulate
this logic.
This used "argparse" to parse the args, which broke since CXXFLAGS
contained options.
Instead we pass "--all" before any other arguments, and then stop
argparse at nonoptions.
A key frustration with the prior version of mkvextract completions was
that even in a position where a filename would be expected, no
completions for a filename were offered. This update introduces more
rigorous argument handling, most importantly restricting
track/attachment completion to when both a mode and a file are
specified.
For some reason Ubuntu's version of screen includes timestamps in the
output of `screen -list`. The timestamps aren't present on other
distributions (tested on Fedora and Arch Linux), nor when building from
source. This commit fixes the regex so that with or without the
timestamp, fish will correctly show suggestions for screen sessions.
prettify_node_recursive is replaced with prettify_node_nrecursive
explicite stack is used instead.
Signed-off-by: Janczar Kurek <janczar.kurek@student.uj.edu.pl>
Now Sphinx is my best friend
This switches the docs to build with Sphinx instead of Doxygen.
There's a lot remaining to do: see #5696. However it is painful to mirror
docs changes from Doxygen to Sphinx, so it is better to switch over now
even in this incomplete state.
The last Doxygen build is tagged as 'last_doxygen'
Closes#5640
sphinx expects that the description for a command (as appearing in its man
page) be provided in conf.py, not in the rst file itself. LLVM handles this
with some custom Python code that parses it out of the file. Do the same
thing in fish.
fish's signal handlers are now sufficiently innocuous that there should
be no reason to block signals (outside of temporarily, when creating a
thread and we need to manipulate the signal mask).
Prior to this fix, an "event" was used as both a predicate on which events
to match, and also as the event itself. Re-express these concepts
distinctly: an event is something that happened, an event_handler is the
predicate and name of the function to execute.
Prior to this fix, fish had a signal_list_t that accumulated signals.
Signals were added to an array of integers, with an overflow flag.
The event machinery would attempt to atomically "swap in" the other list.
After this fix, there is a single list of pending signal events, as an array
of atomic booleans. The signal handler sets the boolean corresponding to its
signal.
In a galaxy far, far away, event_blockage_t was intended to block only cetain
events. But it always just blocked everything. Eliminate the event block
mask.
Arch changed the version string to include the package rel, so it
looks like
systemd 241 (241.7-2-arch)
which would break our simple `string replace` and `test`.
Fixes#5689.
[ci skip]
This was treated as a glob where it was still enabled, most likely removing the "-E" option from argparse,
which caused `sudo -E` to not be parsed correctly, breaking completion.
(There was no error because the glob was used with `set`)
Fixes#5675.
[ci skip]
As it turns out, NetBSD's rand(3) is awful - it's possible that in any
given run it'll only return odd numbers, which means
while (rand() % 10)
will never stop.
Since random(3) is also standardized and works, let's use that!
This commit merges support for a new event publishing mechanism
"topic_monitor" that allow for waiting on multiple event types. It then
replaces waitpid() logic inside `process_mark_finished_children` with new
and simpler logic built around this mechanims. Lastly it migrates the
builtin and function output from processes to background threads.
Now that we use an internal process to perform builtin output, simplify the
logic around how it is performed. In particular we no longer have to be
careful about async-safe functions since we do not fork.
Also fix a bunch of comments that no longer apply.
This uses the new internal process mechanism to write output for builtins.
After this the only reason fish ever forks is to execute external processes.
This introduces "internal processes" which are backed by a pthread instead
of a normal process. Internal processes are reaped using the topic
machinery, plugging in neatly alongside the sigchld topic; this means that
process_mark_finished_children() can wait for internal and external
processes simultaneously.
Initially internal processes replace the forked process that fish uses to
write out the output of blocks and functions.
This adds an "in-process" interpretation of dup2s, allowing for fish to
output directly to the correct file descriptor without having to perform
an in-kernel dup2 sequence.
The sigchld generation expresses the idea that, if we receive a sigchld
signal, the generation will be different than when we last recorded it. A
process cannot exit before it has launched, so check the generation count
before process launch. This is an optimization that reduces failing
waitpid calls.
This is a big change to how process reaping works, reimplenting it using
topics. The idea is to simplify the logic in
process_mark_finished_children around blocking, and also prepare for
"internal processes" which do not correspond to real processes.
Before this change, fish would use waitpid() to wait for a process group,
OR would individually poll processes if the process group leader was
unreapable.
After this change, fish no longer ever calls blocking waitpid(). Instead
fish uses the topic mechanism. For each reapable process, fish checks if
it has received a SIGCHLD since last poll; if not it waits until the next
SIGCHLD, and then polls them all.
topic_monitor allows for querying changes posted to one or more topics,
initially sigchld. This will eventually replace the waitpid logic in
process_mark_finished_children().
Comment from the new header:
Topic monitoring support. Topics are conceptually "a thing that can
happen." For example, delivery of a SIGINT, a child process exits, etc. It
is possible to post to a topic, which means that that thing happened.
Associated with each topic is a current generation, which is a 64 bit
value. When you query a topic, you get back a generation. If on the next
query the generation has increased, then it indicates someone posted to
the topic.
For example, if you are monitoring a child process, you can query the
sigchld topic. If it has increased since your last query, it is possible
that your child process has exited.
Topic postings may be coalesced. That is there may be two posts to a given
topic, yet the generation only increases by 1. The only guarantee is that
after a topic post, the current generation value is larger than any value
previously queried.
Tying this all together is the topic_monitor_t. This provides the current
topic generations, and also provides the ability to perform a blocking
wait for any topic to change in a particular topic set. This is the real
power of topics: you can wait for a sigchld signal OR a thread exit.
Changes according to the feedback have been made:
- What is a shell section has been moved before Installation and Start section
- Content changes have been made as suggested in both of the above sections.
Just a bunch of rewriting descriptions and some arguments.
Most arguments here are uncompleteable, and most of these options will
never be used.
[ci skip]
This resolves the issue where running pre-compiled Linux packages from
binary package manager repositories lead fish to think that we are not
running under WSL.
- Closes#5619.
- Ping neovim/neovim#7330
NetBSD's man is unusual in that it doesn't understand an empty
$MANPATH component as "the system man path", and doesn't have a
`manpath` or `man --path`.
It has a `-m` option that would be useful, but other mans also have a
`-m` option that isn't, so detecting it is tough.
It does have a `-p` option that almost does what one would want here,
so we hack around it to make things work.
Fixes#5657.
[ci skip]
This happens on OpenIndiana/Solaris/Illumos/SunOS.
Elsewhere we use read_blocked, which already returned in this
case (and which we might want to use here as well!).
On some systems, this sometimes uses unicode quotation marks.
Not on mine, but on Travis it does.
The only other workaround I can think of is setting locale to C, but
that implies not being able to test anything unicode-related in the
entire invocation tests.
So for now disable this test.
It turns out the default gettext on the sunny operating system with
the many names interprets at least `\n` itself, so we'd end up
swallowing it.
This allows us to move past the interactive tests and onto the expect
ones.
See #5472.
Illumos/OpenIndiana/SunOS/Solaris has an rm/rmdir that tries to
protect the user by not allowing them to delete $PWD.
Normally, this would be a good thing as deleting $PWD is a stupid
thing to do. Except in this case, we absolutely need to do that.
So instead we weasel around it by invoking an sh to cd out of the
directory to then invoke an `rmdir` to delete it. That should throw
off any attempts at protection (we could also have tried $PWD/. or
similar, but that's possibly still protected against).
This is the last failing test on
Illumos/OpenIndiana/SunOS/Solaris/afunnyquip, so:
Fixes#5472.
This tested #1728, where redirecting a directory (`begin; something;
end < .`) would cause `status` to misbehave.
Unfortunately, on Illumos/OpenIndiana/SunOS, this returns a different
error (EINVAL instead of EISDIR), so we can't check that with our test harness, because
we can't redirect it.
Since it's not important that this gives the same error across
systems (and indeed we provide no way of intercepting the error!),
use an invocation test instead, because that allows different output per-uname.
See #5472.
Matches upstream path_helper which is invoked in /etc/profile and only
applies to login shells. Enables running interactive, non-login shells
with altered PATH values.
Reverts change in c0f832a7, which reverts change in adbaddf.
`fish_title` as invoked by fish itself is not running in an interactive
context, and attempts to read from the input fd (e.g. via `read`) cause
fish to segfault, go into an infinite loop, or hang at the read prompt
depending on the exact command line and fish version.
This patch addresses that by explicitly closing the input fd when
invoking `fish_title`.
Reported by @floam in #5629. May close that issue, but situation is
unclear.
These are files with staged modifications, and additional unstaged
ones.
In practice what happened was that you ran
git add somefile
then editted it some more and tried to
git add <TAB>
which didn't offer it anymore.
Now, we offer it if either modified or modified-staged is set.
Currently modified-staged isn't ever set alone, but through
all-staged, so we still need to keep offering the file then.
(This shows that the current switch/case might have some holes)
Fixes#5648.
[ci skip]
Taking advantage of the maybe_t's, the logic and nesting here
can be a bit less intense.
Small adjustments to debug output, and found a more accurate
version number for Lion Terminal.app.
Longer term we should have a terminal_t class or something
encapsulating all the kinds of terminal detection we have
with methods that return the color support, and also stuff
like whether the terminal has the newline glitch, the
ambiguous width character behavior, etc.
fish forks child processes when (for example) writing out builtin output.
After fork it resets signal handlers, but if a signal is delivered before
the signal handlers are reset, it will inherit fish's default handlers,
which do things like swallow SIGINT. Teach fish's default signal handlers
to detect this case and re-raise signals with default handlers.
This improves the reliability of control-C in the face of builtins.
This exposes it more, since it's quite an important function.
We should do the same with the other vcs functions.
We leave a compatibility shim in place for now.
I hope this is now complete.
Also, shorten enough descriptions to make `string match --<TAB>`
show a two column pager with 80 cols.
We really should have shown more retraint in the design of `string`,
not all of the flags required both a long and short option created.
I hope this is now complete.
Also, shorten enough descriptions to make `string match --<TAB>`
show a two column pager with 80 cols.
We really should have shown more retraint in the design of `string`,
not all of the flags required both a long and short option created.
300ms was waaay too long, and even 100ms wasn't necessary.
Emacs' evil mode uses 10ms (0.01s), so let's stay a tad higher in case
some terminals are slow.
If anyone really wants to be able to type alt+h with escape, let them
raise the timeout.
Fixes#3904.
`/tmp` isn't present / writeable on every system. Instead of always
using `/tmp`, try to use standard environment variables and
configuration to find a temporary directory.
Adapted from #3974, with updates based on those comments.
Closes#3845.
This is a large change to how io_buffers are filled. The essential problem
comes about with code like (example):
echo ( /bin/pwd )
The output of /bin/pwd must go to fish, not the tty. To arrange for this,
fish does the following:
1. Invoke pipe() to create a pipe.
2. Add an io_bufferfill_t redirection that owns the write end of the pipe.
3. After fork (or equiv), call dup2() to replace pwd's stdout with this pipe.
Now when /bin/pwd writes, it will send output to the read end of the pipe.
But who reads it?
Prior to this fix, fish would do the following in a loop:
1. select() on the pipe with a 10 msec timeout
2. waitpid(WNOHANG) on the pwd proc
This polling is ugly and confusing and is what is replaced here.
With this new change, fish now reads from the pipe via a background thread:
1. Spawn a background pthread, which select()s on the pipe's read end with
a long (100 msec) timeout.
2. In the foreground, waitpid() (allowing hanging) on the pwd proc.
The big win here is a major simplification of job_t::continue_job() since
it no longer has to worry about filling buffers. This will make things
easier for concurrent execution.
It may not be obvious why the background thread still needs a poll (100 msec).
The answer is for cases where the write end of the fd escapes, in particular
background processes invoked inside command substitutions. psub is perhaps
the only important case of this (other shells typically just hang here).
This makes some significant architectual improvements to io_pipe_t and
io_buffer_t.
Prior to this fix, io_buffer_t subclassed io_pipe_t. io_buffer_t is now
replaced with a class io_bufferfill_t, which does not subclass pipe.
io_pipe_t no longer remembers both fds. Instead it has an autoclose_fd_t,
so that the file descriptor ownership is clear.
This switches IO redirections after fork() to use the dup2_list_t,
instead of io_chain_t. This results in simpler code with much simpler
error handling.
Prior to this fix, we would write to a fifo via cat >$filename & .
However in some cases (and soon in all cases) we open the file before
the fork, not after. This results in a deadlock because the file open
cannot succeed until a write begins.
Switch to using tee to write to the file. Because tee opens the file itself,
fish is no longer responsible and the deadlock is resolved.
This represents a "resolved" io_chain_t, where all of the different io_data_t
types have been reduced to a sequence of dup2() and close(). This will
eliminate a lot of the logic duplication around posix_spawn vs fork, and pave
the way for in-process redirections.
This is a large change to how io_buffers are filled. The essential problem
comes about with code like (example):
echo ( /bin/pwd )
The output of /bin/pwd must go to fish, not the tty. To arrange for this,
fish does the following:
1. Invoke pipe() to create a pipe.
2. Add an io_bufferfill_t redirection that owns the write end of the pipe.
3. After fork (or equiv), call dup2() to replace pwd's stdout with this pipe.
Now when /bin/pwd writes, it will send output to the read end of the pipe.
But who reads it?
Prior to this fix, fish would do the following in a loop:
1. select() on the pipe with a 10 msec timeout
2. waitpid(WNOHANG) on the pwd proc
This polling is ugly and confusing and is what is replaced here.
With this new change, fish now reads from the pipe via a background thread:
1. Spawn a background pthread, which select()s on the pipe's read end with
a long (100 msec) timeout.
2. In the foreground, waitpid() (allowing hanging) on the pwd proc.
The big win here is a major simplification of job_t::continue_job() since
it no longer has to worry about filling buffers. This will make things
easier for concurrent execution.
It may not be obvious why the background thread still needs a poll (100 msec).
The answer is for cases where the write end of the fd escapes, in particular
background processes invoked inside command substitutions. psub is perhaps
the only important case of this (other shells typically just hang here).
This makes some significant architectual improvements to io_pipe_t and
io_buffer_t.
Prior to this fix, io_buffer_t subclassed io_pipe_t. io_buffer_t is now
replaced with a class io_bufferfill_t, which does not subclass pipe.
io_pipe_t no longer remembers both fds. Instead it has an autoclose_fd_t,
so that the file descriptor ownership is clear.
This switches IO redirections after fork() to use the dup2_list_t,
instead of io_chain_t. This results in simpler code with much simpler
error handling.
Prior to this fix, we would write to a fifo via cat >$filename & .
However in some cases (and soon in all cases) we open the file before
the fork, not after. This results in a deadlock because the file open
cannot succeed until a write begins.
Switch to using tee to write to the file. Because tee opens the file itself,
fish is no longer responsible and the deadlock is resolved.
This represents a "resolved" io_chain_t, where all of the different io_data_t
types have been reduced to a sequence of dup2() and close(). This will
eliminate a lot of the logic duplication around posix_spawn vs fork, and pave
the way for in-process redirections.
By exclusively waiting by pgrp, we can fail to reap processes that
change their own pgrp then either crash or close their fds. If we wind
up in a situation where `waitpid(2)` returns 0 or ECHLD even though we
did not specify `WNOHANG` but we still have unreaped child processes,
wait on them by pid.
Closes#5596.
By exclusively waiting by pgrp, we can fail to reap processes that
change their own pgrp then either crash or close their fds. If we wind
up in a situation where `waitpid(2)` returns 0 or ECHLD even though we
did not specify `WNOHANG` but we still have unreaped child processes,
wait on them by pid.
Closes#5596.
* brew.fish: Add `update-reset` subcommand
This command resets all tap's remotes to the latest available upstream. Ideal for debugging before reporting bugs or just housekeeping.
Add missing newlines.
* Add `brew.fish` changes to CHANGELOG.md
We were checking for the $TMUX variable to determine if we were
running under tmux. However when running the tests, the terminal becomes
expect, even though the TMUX variable is still set, so we spew tmux-isms
at expect. Check the value of $TERM for 'screen'.
Since fish began resolving symlinks it broke the running-from-build-dir
detection in fish.cpp if the build directory were a symlink (which is
common on some platforms where the default user HOME directory is a
symlink in the first place, e.g. FreeBSD).
If we read an R_EOF, we'd try to match mappings to it.
In emacs mode, that's not an issue because the generic binding was
always available, but in vi-normal mode there is no generic binding,
so we'd endlessly loop, waiting for another character.
Fixes#5528.
If we read an R_EOF, we'd try to match mappings to it.
In emacs mode, that's not an issue because the generic binding was
always available, but in vi-normal mode there is no generic binding,
so we'd endlessly loop, waiting for another character.
Fixes#5528.
Originally I sought out to configure the foreground color of the
selected text in the pager. After reading a thread on a github issue I
was inpired to do more: now you can conifgure any part of the pager when
selected, and when a row is secondary. More specifically this commit adds the
ability to specify a pager row's:
- Prefix
- Completion text
- Description
- Background
when said row is selected or secondary.
This allows disabling _just_ the informative status.
We still also use the dirty and untracked variables, but only if
informative status hasn't explicitly been enabled.
If either of the two git config variables:
- bash.showDirtyState
- bash.showUntrackedFiles
is explicitly set to false, we will disable informative status, and
fall back on the non-informative version (most likely still with
either dirty or untracked files, since we already use the variables
for that).
These vars are read by the official git prompt, so we use them instead
of inventing our own "fish.showInformativeStatus".
(Note: This also uses $__fish_git_prompt_showdirtystate and friends,
but only when there's nothing set in the repo, and there's really no
reason to set those to false if using the informative status)
Fixes#5551.
[ci skip]
This will print out along with the stuff we've guessed about color
support. We get a lot of bug reports about these messing up rendering,
this is useful diagnostic output.
Ask the system where utilities are available with confstr (POSIX).
This is the same string printed by `getconf PATH`, which likely
includes more directories.
Expands the utility of `type -p foo` by allowing it to print the path to
the script that defines `foo` when `foo` is a valid function that was
sourced from a path on disk (rather than interactively defined).
This does not change the behavior of `type -P`/`type --force-path`,
which should have already been used if the desire was to resolve the
path to an executable file (otherwise the output would have been blank
if a function was shadowing an executable file of the same namea), so no
backwards compatibility issues are expected.
Using printf like
printf "The message"
is unsafe, because if the message contains any formatting characters,
they'll be interpreted.
In this case it's not all that important because the message contains
only filenames of our tests and static strings, but still.
I was surprised to see:
> set_color normal | string escape
\e\[30m\e\(B\e\[m
I only expected to see a sgr0 here.
Cleanup a nearby `else { if (...) {` and comment with a bogus example.
A person stuck installing it just for fish on their server
doesn't want to waste time installing the wrong one, so assuage that.
Also tweak to look nicer with 80 columns
As discussed in #5492, it would be good if running fish_config without
Python actually told the user to install Python.
Further, let's give the person some hints on how to configure these
things by hand, since they may have to.
I believe this should take care of the reported problem with the
corrected definition for `wcstod_l`. For future reference, any changes
to `config.h.in` should also be reflected in `osx/config.h`
`xlocale.h` is not available on Linux, so we can't just universally
include it.
`HAVE_XLOCALE_H` was already being tested/set in the CMake script as a
possible requirement for `wcstod_l` support, this just adds it to
`config_cmake_h.in` and uses it in `wutil.h` to gate the include.
This was broken in a8eb02f9f5 when the
detection was corrected for FreeBSD. This patch makes the detection work
for both Linux and FreeBSD instead of one or the other (tested).
Using `setlocale` is both not thread-safe and not correct, as
a) The global locale is usually stored in static storage, so
simultaneous calls to `setlocale` can result in corruption, and
b) `setlocale` changes the locale for the entire application, not
just the calling thread. This means that even if we wrapped the
`wcstod_l` in a mutex to prevent the previous point, the results
would still be incorrect because this would incorrectly influence the
results of locale-aware functions executed in other threads while
this thread is executing.
The previous comment mentioned that `uselocale` hadn't worked. I'm not
sure what the failing implementation looked like, but `uselocale` can be
tricky. The committed implementation passes the tests for me under Linux
and FreeBSD.
Don't do it when the relative path is simple (purely descending),
unless the token starts with ":/".
Also stop offering directories - if they need to be disambiguated, the
normal completion logic will take care of that.
Fixes#5574.
[ci skip]
Don't do it when the relative path is simple (purely descending),
unless the token starts with ":/".
Also stop offering directories - if they need to be disambiguated, the
normal completion logic will take care of that.
Fixes#5574.
[ci skip]
If a job is disowned that, for some reason, has a pgid that is special
to waitpid, like 0 (process with pgid of the calling process), -1 (any
process), or our actual pgid, that would lead to us waiting for too
many processes when we later try to reap the disowned processes (to
stop zombies from appearing).
And that means we'd snag away the processes we actually do want to
wait for, which would end with us in a waiting loop.
This is tough to reproduce, the easiest I've found was
fish -ic 'sleep 5 &; disown; set -g __fish_git_prompt_showupstream auto; __fish_git_prompt'
in a git repo.
What we do is to not allow special pgids in the disowned_pids list.
That means we might leave a zombie around (though we probably wait on
0 somewhere), but that's preferable to infinitely looping.
See #5426.
Previously, using special regex characters or slashes would result in an
error message, when pressing tab in a command-line such as
"man /usr/bin/time ".
Previously, using special regex characters or slashes would result in an
error message, when pressing tab in a command-line such as
"man /usr/bin/time ".
This was the actual issue leading to memory corruption under FreeBSD in
issue #5453, worked around by correcting the detection of `wcstod_l` so
that our version of the function is not called at all.
If we are 100% certain that `wcstod_l` does not exist, then then the
existing code is fine. But given that our checks have failed seperately
on two different platforms already (FreeBSD and Cygwin/newlib), it's a
good precaution to take.
This broke when --preset was introduced.
We allow a "--preset" or "--user" to appear right after the "bind",
and save the value, but don't use it yet.
Fixes#5534.
[ci skip]
PWD is not set in fish vars because it is read only.
Use getenv() to fetch it, allowing fish to inherit a virtual PWD.
This cherry pick includes both:
24f251e04 Correctly remove the test directory again in cd test
91a9c9897 Correctly inherit a virtual PWD
Fixes#5525
We can't complete these, and now the user can do
```
set -g __fish_git_alias_$alias $command
```
e.g.
```
set -g __fish_git_alias_co checkout
```
if the arguments in the alias end up going to `git alias`.
Fixes#5412.
[ci skip]
This was an oversight from the previous commit. Not that it matters
much, because we already removed $files.
Still, this would fail if someone defined a global $files, so let's fix it.
[ci skip]
shorter descriptions that can fit in a terminal window, and option arguments added.
hide one option that is only functional on Cygwin unless we are on Cygwin
Make it so that the generated completion has the form \t\n
when the optional description has been ommitted - otherwise
the original option's description gets inherited and is seen hundreds
of times repeating in the pager.
Our weird %-expanding function wrappers around kill et all defined
"--wraps" for the same name.
As it turns out, fish follows that one, and executes the completion
multiple times.
I didn't notice because these tend to be rather quick on linux, but on
macOS that's apparently a real issue.
Fixes#5541.
[ci skip]
For some reason, we have two places where a variable can be read-only:
- By key in env.cpp:is_read_only(), which is checked via set*
- By flag on the actual env_var_t, which is checked e.g. in
parse_execution
The latter didn't happen for non-electric variables like hostname,
because they used the default constructor, because they were
constructed via operator[] (or some such C++-iness).
This caused for-loops to crash on an assert if they used a
non-electric read-only var like $hostname or $SHLVL.
Instead, we explicitly set the flag.
We might want to remove one of the two read-only checks, or something?
Fixes#5548.
GNU ls's --indicator-style=classify is the same as POSIX -F.
Refactor and change command testing logic so that we define the
function in the same place for all platforms, and use -F on all
the platforms when stdout is a TTY.
This enables fuzzy-matching outside of the current directory again.
As it turns out, the performance impact here isn't as large as I
thought - it's massively dependent on caching.
Fixes#5476.
(cherry picked from commit 73bae383e0)
There was a bogus check for is_interactive_session. But if we are in
reader_readline we are necessarily interactive (even if we are not in
an interactive session, i.e. a fish script invoked some interactive
functionality).
Remove this check.
Fixes#5519
There was a bogus check for is_interactive_session. But if we are in
reader_readline we are necessarily interactive (even if we are not in
an interactive session, i.e. a fish script invoked some interactive
functionality).
Remove this check.
Fixes#5519
A while loop now evaluates to the last executed command in the body, or
zero if the loop body is empty. This matches POSIX semantics.
Add a bunch of tricky tests.
See #4982
A while loop now evaluates to the last executed command in the body, or
zero if the loop body is empty. This matches POSIX semantics.
Add a bunch of tricky tests.
See #4982
This is effectively a pick of 2ebdcf82ee
and the subsequent fixup. However we also avoid setting WNOHANG unless
waitpid() indicates a process was reaped.
Fixes#5438
This is effectively a pick of 2ebdcf82ee
and the subsequent fixup. However we also avoid setting WNOHANG unless
waitpid() indicates a process was reaped.
Fixes#5438
If it's a foreground job, it is related to the currently running exec.
This fixes exec in functions, i.e.
function reload
exec fish
end
would previously always ask about the "function reload" job.
Fixes#5449.
Fixesoh-my-fish/oh-my-fish#664.
I believe this should take care of the reported problem with the
corrected definition for `wcstod_l`. For future reference, any changes
to `config.h.in` should also be reflected in `osx/config.h`
For some reason, we have two places where a variable can be read-only:
- By key in env.cpp:is_read_only(), which is checked via set*
- By flag on the actual env_var_t, which is checked e.g. in
parse_execution
The latter didn't happen for non-electric variables like hostname,
because they used the default constructor, because they were
constructed via operator[] (or some such C++-iness).
This caused for-loops to crash on an assert if they used a
non-electric read-only var like $hostname or $SHLVL.
Instead, we explicitly set the flag.
We might want to remove one of the two read-only checks, or something?
Fixes#5548.
Our weird %-expanding function wrappers around kill et all defined
"--wraps" for the same name.
As it turns out, fish follows that one, and executes the completion
multiple times.
I didn't notice because these tend to be rather quick on linux, but on
macOS that's apparently a real issue.
Fixes#5541.
[ci skip]
On `set fish_color_cwd <TAB>`, a bunch of named colors are
shown in the pager. Each and every one has a description of "Color".
These are all very obviously colors, and none are not colors,
the description does not tell us anything specific about the item.
Descriptions in situations like this are actually a hinderance
because of the way they cause less to fit into the pager. Remove it
There's just waaayy too many things that could go wrong with it, so it
annoys more than it helps, especially since we don't get any
indication what failed.
E.g. on FreeBSD, the test failed without a usable message just because
`tput` couldn't find an attribute (so colors were unset).
Some $TERMs like tmux and linux use an sgr0 ("reset") value that ends
in \co instead of "m". We need to adjust our regex here to catch that,
or we'd miscount lines with it.
This broke when --preset was introduced.
We allow a "--preset" or "--user" to appear right after the "bind",
and save the value, but don't use it yet.
Fixes#5534.
[ci skip]
Our is_hex_digit() was redundant, we can just use iswxdigit; the libc
implementation is a more efficient table lookup anyhow.
Do is_octal_digit() in terms of iswdigit instead of using wcschr.
A person stuck installing it just for fish on their server
doesn't want to waste time installing the wrong one, so assuage that.
Also tweak to look nicer with 80 columns
As discussed in #5492, it would be good if running fish_config without
Python actually told the user to install Python.
Further, let's give the person some hints on how to configure these
things by hand, since they may have to.
This was an oversight from the previous commit. Not that it matters
much, because we already removed $files.
Still, this would fail if someone defined a global $files, so let's fix it.
[ci skip]
We can't complete these, and now the user can do
```
set -g __fish_git_alias_$alias $command
```
e.g.
```
set -g __fish_git_alias_co checkout
```
if the arguments in the alias end up going to `git alias`.
Fixes#5412.
[ci skip]
This enables fuzzy-matching outside of the current directory again.
As it turns out, the performance impact here isn't as large as I
thought - it's massively dependent on caching.
Fixes#5476.
Which is 4, apparently.. (builtin_set.cpp returns ENV_NOT_FOUND)
here. This was previously hardcoded to our 121, which used to be
what builtins used for invalid arguments.
4 is pretty arbitrary but at least this is more consistent.
I had previously introduced a lot of updates and fixes to npm registry
based completions for `yarn` but hadn't ported them to `npm` as well
(although they can be dropped in as-is). This patch shares the code
between the two, which resides in an explicitly sourced multi-function
fish script.
The informational message is only shown the first time an attempt at
completing `yarn add` is made per session. This should vastly improve
the discoverability of this feature as regular yarn/npm users would
never have `all-the-package-names` installed normally.
This merges a bunch of changes that eliminate fish variables as a global
concept. Instead fish variables are tied to an instance of environment_t
(read-only) or env_stack_t (read/write), which is explicitly threaded
through every site. This is nice cleanup and also preparation for
concurrent execution, where multiple independent threads may need to see
different variables.
This requires threading environment_t through many places, such as completions
and history. We introduce null_environment_t for when the environment isn't
important.
The compiler flag `-Werror=unguarded_availability` was hard-coded for
macOS, but is not supported by GCC on macOS 10.10 (Yosemite). Test for
support with CHECK_CXX_COMPILER_FLAG before forcing it.
`xlocale.h` is not available on Linux, so we can't just universally
include it.
`HAVE_XLOCALE_H` was already being tested/set in the CMake script as a
possible requirement for `wcstod_l` support, this just adds it to
`config_cmake_h.in` and uses it in `wutil.h` to gate the include.
Removes the dependency on the current user's home directory, instead
overriding it to be within the current hierarchy.
Fixes the tests on Debian buildd, where the home directory is
deliberately unwriteable to pick up errors in builds.
The one thing I was missing:
`echo -n` isn't POSIX. In practice, it appears the only shell to encounter this
is macOS' crusty old bash in sh-mode. Just replace it with `touch`.
This reverts commit fc5e8f9fec.
This makes the script worse, but it's good enough.
The required changes are:
- `shopt -s nullglob`, which we simply don't use (we have one glob, but that's
guaranteed to match because we ship the files)
- One array, which we replace with a direct use of the glob (plus it
used `echo` again?)
- The `function` word, which I'm still annoyed is even a thing!
- Variable indirection (`color=${!color_var}` - instead we pass the
value directly - which makes the script uglier!)
- One array, which we replace with a function
- A use of `type -t`, replaced with `command -v`
- A use of `${var:begin:end}` substring expansion, replaced with trickery.
- `set -o pipefail` is replaced with a function
Note that checkbashisms still complains about `command -v`, because
we're not using it with "-p". But we _want_ to check the current
$PATH, and `command -v` is POSIX.
This still uses `local`, which technically isn't in POSIX.
The tests now appear to pass in:
- bash
- dash
- zsh
- mksh
- busybox
Starting with Fedora 30 and RHEL 8, ambiguous python shebangs will now
throw errors during the RPM build process instead of just warnings,
since these systems have moved to Python 3 by default, and Python 2 may
not be available in the future.
See [this
page](https://fedoraproject.org/wiki/Changes/Make_ambiguous_python_shebangs_error)
for more details.
Drop these shebangs as the scripts are only ever called from fish
wrappers.
Rather than killing the process with close, read EOF after sending the
"exit" command and wait for OS cleanup (per the expect examples).
Not cleaning up with wait caused expect to crash on all 32-bit platforms
including i586 and armv7l with "alloc: invalid block: 0xbf993ccb: 3d 3b".
64-bit platforms were not affected, for reasons that are not clear.
CentOS 7 does not have rhel_version as one of its macros, so trying to
build results in CMake errors, since we get `cmake` instead of
`cmake3`. These additional conditions allow the spec to build
successfully on CentOS 7.
Using %rhel should allow one set of conditionals to work across CentOS 7
and RHEL 7.
This has been tested on both.
This was broken in a8eb02f9f5 when the
detection was corrected for FreeBSD. This patch makes the detection work
for both Linux and FreeBSD instead of one or the other (tested).
Using `setlocale` is both not thread-safe and not correct, as
a) The global locale is usually stored in static storage, so
simultaneous calls to `setlocale` can result in corruption, and
b) `setlocale` changes the locale for the entire application, not
just the calling thread. This means that even if we wrapped the
`wcstod_l` in a mutex to prevent the previous point, the results
would still be incorrect because this would incorrectly influence the
results of locale-aware functions executed in other threads while
this thread is executing.
The previous comment mentioned that `uselocale` hadn't worked. I'm not
sure what the failing implementation looked like, but `uselocale` can be
tricky. The committed implementation passes the tests for me under Linux
and FreeBSD.
This was the actual issue leading to memory corruption under FreeBSD in
issue #5453, worked around by correcting the detection of `wcstod_l` so
that our version of the function is not called at all.
If we are 100% certain that `wcstod_l` does not exist, then then the
existing code is fine. But given that our checks have failed seperately
on two different platforms already (FreeBSD and Cygwin/newlib), it's a
good precaution to take.
CMake seems to have trouble finding libraries from multiarch packages
that do not have the compatibility symlink installed to the
arch-independent library directory. Users must either manually supply
the path to the library in question via command-line parameters or we
can fall back to CMake's alternate method of finding packages based off
of pkg-config rather than using the hard-coded `FindCurses` CMake module
specific to the CMake version/distribution installed.
e4b6007f33 introduced the following warning:
configure.ac:327: warning: AC_LANG_CONFTEST: no AC_LANG_SOURCE call
detected in body
Fix by using the right autoconf macros for the job.
Turns out busybox diff (used on alpine) defaults to unified output,
which we can't use because that prints filenames, and those are
tempfiles made by psub.
Instead, we use builtins to print the first line and compare the others.
This isn't all that important, and it breaks on musl just because the message is different.
Just skip it for now, until we figure out how to better test this.
This `set TERM`. Which, if $TERM is inherited, is already exported,
but not if it isn't.
This is the case on sr.ht's arch images, so we failed without a TERM variable.
This checks if uname exists (we already call it elsewhere without
check, nobody has complained, uname is in POSIX), then calls to see if
it's "Linux", and only then offers any completions.
Since we don't have any other version to offer, the check is worse
than useless.
This helps on netbsd, because enter_standout_mode et al are const
there.
These methods don't alter their argument, so they should have been
const to begin with.
This is non-const on macOS, but some of the args we pass are always
const on netbsd.
I have no idea why you'd ever want this to modify its argument, but whatever.
This is the more correct fix for #5447, as regardless of which process
in the job (be it the first or the last) finished first, once we have
waited on a process without ~WNOHANG we don't do that for any subsequent
processes in the job.
It is also a waste to call into the kernel to wait for a process we
already know is completed!
@ridiculousfish had introduced this in 3a45cad12e
to work around an issue with Coverity Scan where it couldn't tell the
mutex was correctly locked, but even with the `fish_mutex_t` hack, it
still emits the same warnings, so there's no pointing in keeping it.
This is necessary for the history race condition test to succeed.
(That test is permanently disabled under WSL (as it always fails) so I
didn't catch this on my end.)
Use `pthread_atfork()` to mark child processes as dirty when `fork()` is
invoked rather than needing to call into the kernel each time
`ASSERT_IS_NOT_FORKED_CHILD()` is called.
This makes simple test cases that hit `ASSERT_IS_NOT_FORKED_CHILD()` 1.8x faster.
------------------------
With a7998c4829 reverted but before this optimization:
```
mqudsi@ZBOOK ~/r/fish-shell> hyperfine -S build/fish 'for i in (seq 100000); test 1 = 1; end'
Benchmark #1: for i in (seq 100000); test 1 = 1; end
Time (mean ± σ): 717.8 ms ± 14.9 ms [User: 503.4 ms, System: 216.2 ms]
Range (min … max): 692.3 ms … 740.2 ms
```
With a7998c4829 reverted and with this optimization:
```
mqudsi@ZBOOK ~/r/fish-shell> hyperfine -S build/fish 'for i in (seq 100000); test 1 = 1; end'
Benchmark #1: for i in (seq 100000); test 1 = 1; end
Time (mean ± σ): 397.2 ms ± 22.3 ms [User: 322.1 ms, System: 79.3 ms]
Range (min … max): 376.0 ms … 444.0 ms
```
Without a7998c4829 reverted and with this optimization:
mqudsi@ZBOOK ~/r/fish-shell> hyperfine -S build/fish 'for i in (seq 100000); test 1 = 1; end'
Benchmark #1: for i in (seq 100000); test 1 = 1; end
Time (mean ± σ): 423.4 ms ± 51.6 ms [User: 363.2 ms, System: 61.3 ms]
Range (min … max): 378.4 ms … 541.1 ms
```
By using a user-land thread-local integer and lock-free (at least under
x86/x64) atomics, we can implement a safe `assert_is_main_thread()`
without calling into the kernel. Thread-local variables are part of
C++11.
This is called a lot in some performance-sensitive areas, so it is worth
optimizing.
This fixes#5438 by having fish block while waiting on a foreground job
via its individual processes by enumerating the procs in reverse order,
such that we hang waiting for the last job in the IO chain to terminate,
rather than the first.
If it's a foreground job, it is related to the currently running exec.
This fixes exec in functions, i.e.
function reload
exec fish
end
would previously always ask about the "function reload" job.
Fixes#5449.
Fixesoh-my-fish/oh-my-fish#664.
Mainly this removes the "TYPE_MASK" macro that just masks off the
higher bits, which I don't think were ever actually used.
Much of this seems like anticipation of future direction, but we're
going somewhere else.
This removes the need to run c-compilation on one file, and allows us
to in future c++-ify this a bit.
There's a lot of bit-fiddling here that is quite unnecessary, better
error-handling would be nice...
So far this removes a few more unused things (because I would have had
to port them), including:
- Functions with ARITY > 3 (even 3 isn't used, but just so we don't
get complacent)
- Variables
- Most functions moved out of the header, because only te_interp is used.
- The te_print function
The function `add_disowned_pgid` adds process *group* ids and not
process ids. It multiplies the value by negative 1 to indicate a wait
on a process group, so the original value must be positive.
If a job is disowned that, for some reason, has a pgid that is special
to waitpid, like 0 (process with pgid of the calling process), -1 (any
process), or our actual pgid, that would lead to us waiting for too
many processes when we later try to reap the disowned processes (to
stop zombies from appearing).
And that means we'd snag away the processes we actually do want to
wait for, which would end with us in a waiting loop.
This is tough to reproduce, the easiest I've found was
fish -ic 'sleep 5 &; disown; set -g __fish_git_prompt_showupstream auto; __fish_git_prompt'
in a git repo.
What we do is to not allow special pgids in the disowned_pids list.
That means we might leave a zombie around (though we probably wait on
0 somewhere), but that's preferable to infinitely looping.
See #5426.
* Severely extended the sorin theme
This theme should now mostly match the original.
* Removed superfluous whitespace
* Inlined external links as ASCII art
* Made myself the author of the sorin theme
* Removed superfluous read delemiter
* Renamed __fish_git_action to fish_print_git_action
* Adde a minor comment
Return STATUS_INVALID_ARGS when failing due to evaluation errors,
so we can tell the difference between an error and falseness.
Add a test for the ERANGE error
The rest of the high-numbered exit codes are not values used by scripts
or builtins, they are internal to fish and come out of
the parser for example.
Prior to adding STATUS_INVALID_ARGS, builtins were usually exiting 2
if they had a special exit status for the situation of bad arguments.
Set it to 2.
We were not parsing an in-range number when we claimed we were,
and were thus failing to error with invalid numbers and returned
a wrong test result. Fixed#5414
Also, provide the detail we can for the other error cases.
Return STATUS_INVALID_ARGS when failing due to evaluation errors,
so we can tell the difference between an error and falseness.
Add a test for the ERANGE error
The rest of the high-numbered exit codes are not values used by scripts
or builtins, they are internal to fish and come out of
the parser for example.
Prior to adding STATUS_INVALID_ARGS, builtins were usually exiting 2
if they had a special exit status for the situation of bad arguments.
Set it to 2.
Cleaned up the code to no longer replicate in fishscript what fish
already does (and caches to boot) in C++ in setting up the paths to the
user configuration directory.
Also introduced a `$__fish_user_data_dir` instead of the sporadic
definitions of `$userdatadir` that may or may not go through
`XDG_DATA_HOME`.
We were not parsing an in-range number when we claimed we were,
and were thus failing to error with invalid numbers and returned
a wrong test result. Fixed#5414
Also, provide the detail we can for the other error cases.
I spent some time figuring out $TERM_PROGRAM_VERSION and Terminal.app's
capabilities over time. [1]
Only use OSC 7 if running on the version of Terminal.app that added it
or newer. In the past this would have been harder because `test` couldn't
do float comparisons.
cleanup:
Don't bother setting a local $TERM_PROGRAM if it's unset: quoting
is enough to keep test happy. For the version numbers, 0"$var" is safe
against unset variables for numerical comparisons.
[1]: https://github.com/fish-shell/fish-shell/wiki/Terminal.app-characteristics
Instead of maybe adding "-s" and "-M" if "-s" hasn't already been
given, just add "-s" to _every_ bind invocation, and "-M" to those who
need it.
Fixes#5028.
Largely reverts 007d794b6e.
fish_indent is extremely resource-intensive on large inputs and can crash; it also does not handle
invalid characters gracefully.
Work on #5402.
Just sets locale to "C" (because that's the only one we need), does
wcstod and resets the locale.
No idea why uselocale(loc) failed for me, but it did.
Fixes#5407.
This previously used /dev/tty to make sure we have `source` connected
to a terminal. Only as it turns out, FreeBSD doesn't have that (https://builds.sr.ht/~faho/job/15308).
So instead, let's just use the expect tests since stdin there is by
definition a terminal.
This happens in firejail, and it means that we can't use it as an
argument to most pgid-taking functions.
E.g. `wait(0)` means to wait for the _current_ process group,
`tcsetpgrp(0)` doesn't work etc.
So we just stop doing this stuff and hope it works.
Fixes#5295.
The solarized themes now define pager colors, while other schemes
don't.
So if a user picks one of them, and then another, they'd keep the
pager colors.
Instead, since the default theme is now complete, any theme that does
not define its own pager colors will always get the default ones.
[ci skip]
This was missing a bunch of variables from __fish_config_interactive.
Ideally we wouldn't have to duplicate this info, but I don't have a
great solution either.
This removes ~140ms from every single prompt.
When not in a git repo, this prompt now takes ~9ms, as opposed to
~150ms before.
Fixes#5266 harder.
[ci skip]
This is quite ugly, but in lieu of putting in a proper ansi
parser (i.e. the output part of a terminal), since this is the only
such sequence we have seen until now, let's just match it.
Fixes#5312.
[ci skip]
- Remove use of `eval`
- Use `git rev-parse` instead of `git status` as its faster,
- especially in large repos. (in qt5: 600ms vs 1ms)
- Use return status instead of test -n
This should change nothing about the output.
This uses some more string, but the main improvement is using "git
rev-list" instead of parsing "git branch" output that happens to be localized.
[ci skip]
I can't see the value in this, given that we have a bunch of minimalist ones.
The "escaping" here is gnarly enough that I don't want to attempt to clean it up.
man.fish can be clarified a bit, by removing a superfluous early return. Additionally, performance can be
(ever so slightly) improved, by using the empty string to suffix an extra colon when `$MANPATH` is empty, as
described in `manpath(1)`. As `man` will internally call `manpath` as it starts, this eliminates a redundancy.
This adds the color variables from the docs to both the python script
and the js controller.
Among others, this includes "search_match", i.e.
"fish_color_search_match".
It still does not include the pager colors because the variable names
wouldn't match.
This reverts commit 1cb8b2a87b.
argv[0] has the full path in it for a user when he executes it
out of $PATH. This is really annoying in the title which uses $_.
Also check if that is actually defined, not the cur_term proxy.
In #5371, we figured out that there are terminfo entries without this
capability, so this would do a NULL-dereference.
OCLINT was ignoring this, but we can just not do the bad thing.
Declare argc and argv const. These are in the stack, they can
be modified, but we won't.
Fix a typo
... rather than hard code it to "fish". This affects
what is found in $_ and improves the errors:
For example, if fish was ran with ./fish, instead of
something like:
fish: Expected 3 surprises, only got 2 surprises
we'll see:
./fish: Expected 3 surprises, only got 2 surprises
like most other shell utilities. It's just a tiny bit
of detail that can avoid confusion.
This broke fishtape, which did
somestuff | fish -c "source"
Because `source` didn't have a redirection, it refused to read from
stdin.
So, to keep the common issue of `source (command that does not print)`
from seeminly stopping fish, we instead actually check if stdin is a terminal.
This was causing problems if "fish" wasn't in exec_path, like
if the binary had been renamed.
I also noticed that even with 'fish' not renamed, only paths.data
was made relative to my source tree. paths.sysconf, paths.doc, and
paths.bin were all relative to /usr/local.
This had a bunch of "do_{backward,forward}" movements that differed
only in one argument.
Just keep them together, so it's less code, and less needs to be
changed.
`ls` was suggesting options that are are not valid for my system,
omitting options that are on my system. Different BSD OSes have
different option extensions, and some of them do conflict with eachother.
I carefully checked the manuals of netbsd, macos, freebsd, and openbsd
`ls` and made the completions show the right completions in full for them.
Some verbiage tweaks as well.
- No longer uses sed, sort, uniq, uname
- Stop doing too-clever filtering (e.g. the kernel thread stuff never
- really worked)
- Don't truncate for all OSen, instead just use the (correctly
- truncated) comm field.
The colors happening for the interactive tests didn't match the
expected output. For `history search` commands we test, have them
pipe through `cat` so the fishscript does not use a pager or try
to colorize.
- Colorize history search output when interactive, using
fish_indent. This is the same way we colorize `type` output.
- Ask less to act like `cat` if the output will fit in the
terminal window, so it's less jaring with short output.
- history is viewed in a pager when interactive, but pagers
typically strip escape codes. We accomplish the above by
doing exactly what `git` does[1] when it has colored output
for a pager:
if $LESS is unset, set it to enable -R, -F, and -X options.
if $LV is unset, set it to -c.
[1]: 398dd4bd03/pager.c (L87)
If the user is in a directory which has been unlinked, it is possible
for the path .. to not exist, relative to the working directory.
Always pass in the working directory (potentially virtual) to
path_get_cdpath; this ensures we check absolute paths and are immune
from issues if the working directory has been unlinked.
Also introduce a new function path_normalize_for_cd which normalizes the
"join point" of a path and a working directory. This allows us to 'cd' out of
a non-existent directory, but not cd into such a directory.
Fixes#5341
realpath() will return NULL and sets errno if it fails.
We asserted that realpath(".") does not fail. We also didn't really
check that it was successful. Made sure we'll get a perror telling
us about what went wrong if something like this happens again.
Updated tests and added test case
Fixes#5351
* Replace "env" with "expr" in the test manpage
I'm pretty sure `env` isn't capable of comparing numbers and the author meant `expr`.
* Update the docs regarding floats support in test
Fixes some potentially unsafe uses of direct substitution into regex
expressions and also switches some completions to regex-based now that
there is a safe way of using it.
If fish detects that it was started with a pgrp of 0 (which appears to
oddly be the case when run under firejail), create new process group for
fish and give it control of the terminal.
This selectively reverts 55b3c45 in cases where an invalid pgrp is
detected. Note that this is known to cause problems in other cases, such
as #3805 and Microsoft/WSL#1653, although the former may have been
ameliorated or even addressed by the recent job control overhaul, so
that's why we are careful to only assign fish to its own pgroup if an
invalid pgroup was detected and not as the normal case.
This reverts commit 54050bd4c5.
Type job_list_t was changed from a list to a deque in
commit 54050bd4c5.
In process_clean_after_marking(), we remove jobs while iterating.
dequeues do not support that. Make it a list again.
This fixes the `~floam/` case, where the out_tail_idx pointer needs to
point to the "/", not the last letter.
The `~/` and `~floam` cases still work.
Unfortunately, I'm unsure of how to test this.
Fixes#5325.
In writing the completion script for openocd I found the need to
complete paths at the command-line as if they were relative to a
path other than the current $PWD. Given that `$PWD` is currently
global in fish (i.e. no side-effect free `cd` within a subshell)
this is probably good to have for other completions too.
This also fixes a bug in support for explicitly supplying the
description for completions via a `$argv` parameter, which prefixed
the description with `\t` (which is correct) except it did so in
the local scope within an `if` statement, meaning the changes never
had any effect and in the output the description was directly
concatenated to the completions, instead of separated by a tab.
Incorrectly assumed that pandoc uses XDG_CONFIG_HOME, it turns out the
path is hard-coded as $HOME/.pandoc unless explicitly otherwise
specified in the command-line.
Limit the fish_wcstod fast path to ASCII digits only, to fix the problem
observed in the discussion for a700acadfa
where LANG=de_DE.UTF-8 would cause `test` to interpret commas instead of
periods inside floating point values.
This merges a stack that removes the WAIT_BY_PROCESS and NESTED job flags.
Instead jobs are taught about their parents, and parents are interrogated to
determine whether they are fully constructed, and therefore whether it is
safe to call waitpid().
Now jobs are aware of their parent jobs, and can interrogate those jobs,
to determine if every job in the chain is fully constructed.
Remove flags and the static stacks that manipulated them.
The parent of a job is the parent pipeline that executed the function or
block corresponding to this job. This will help simplify
process_mark_finished_children().
Prior to this fix, cding into a symlink and then completing .. would complete
from the physical directory instead of the logical directory, which could not
actually be cd'd to. Teach cd completiond to use the logical directory.
Don't attempt to complete against package names if the user is trying to
enter a switch to speed things up.
Also work around #5267 by not wrapping unfiltered `all-the-package-name`
calls in a function.
select_try() returned IO_ERROR to indicate that there's no file descriptors
from which to read. Name this return value properly.
Also migrate this type into proc.cpp since it's not used outside of the
header.
This is an opposite case from the usual "pipe into grep-the-function"
where my `pbpaste` emitted a lot of content exceeding the OS pipe
buffer. The `block_on_fg` condition was just `send_sigcont` in the
original job control rewrite, and it was incorrect to sub it for
WAIT_BY_PROCESS on its own.
However, this requires always blocking when select_try returns an
interrupted/incomplete read or else fish doesn't block and stays running
in a tight loop in the background (and incorrectly writing to a terminal
it doesn't own under higher debug levels), which I *think* is OK.
Instantiate the std:locale instance used within the character comparison
callback outside the lambda and take a reference to it instead of
creating the locale object for each character in the sequence.
This is part of a very tight loop with lots of inputs during the
evaluation of fuzzy string matches for completions/autosuggestions and
is worth optimizing.
This was introduced in 1b1bc28c0a but did
not cause any problems until the job control refactor, which caused it
to attempt to signal the calling `exec` builtin's own (invalid) pgrp
with SIGHUP.
Also improved debugging for `j->signal()` failures by printing the
signal we tried sending in case of error, rename the function to
`hup_background_jobs`, and move it from `reader.h`/`reader.cpp` to
`proc.h`/`proc.cpp`.
When a function is encountered by exec_job, a new context is created for
its execution from the ground up, with a new job and all, ultimately
resulting in a recursive call to exec_job from the same (main) thread.
Since each time exec_job encounters a new job with external commands
that needs terminal control it creates a new pgrp and gives it control
of the terminal (tcsetpgrp & co), this effectively takes control away
from the previously spawned external commands which may be (and likely
are) expecting to still have terminal access.
This commit attempts to detect when such a situation arises by handling
recursive calls to exec_job (which can only happen if the pipeline
included a function) by borrowing the pgrp from the (necessarily still
active) parent job and spawning new external commands into it.
When a parent job spawns new jobs due to the evaluation of a new
function (which shouldn't be the case in the first place), we end up
with two distinct jobs sharing one pgrp (to fix#3952). This can lead to
early termination of a pgrp if finished parent job children are reaped
before future processes in either the parent or future child jobs can
join it.
While the parent job is under construction, require that waitpid(2)
calls for the child job be done by process id and not job pgrp.
Closes#3952.
Use SIGCHLD to determine whether or not waitpid(2) calls can be elided,
but only with extreme caution. If we receive SIGCHLD but are not able to
reap all jobs, we need to iterate through them again.
For this to work, we need to make sure that we reap all children that we
can reap after a SIGCHLD, i.e. it's not OK to just reap the first and
return or else we can never clear the dirty state flag.
In all cases, as expensive as a call to waitpid() may be, if a child
process is available for reaping it is always cheaper to wait on it then
reap it than to call select_try() and end up timing out.
The old code was rather haphazard with regards to error control, and
would make mutable changes before operations that could fail without any
viable error handling options.
Convert `select_try()` to return a well-defined enum describing its
state, and handle each of the three possible cases with clear reasons
why we are blocking or not blocking in each subsequent call to
`process_mark_finished_children()`.
* Use the newly-introduced signal_block_t RAII wrapper
* Remove EINTR loops as all signals are blocked
* Clean up control flow thanks to RAII wrappers
* Rename parameter to clarify what it does and update docs accordingly
* Update outdated comments referencing SIGSTOP code that was removed a
long time ago.
* Remove no-op CHECK_BLOCK() call
* Convert JOB_* enums to scoped enums
* Convert standalone job_is_* functions to member functions
* Convert standalone job_{promote, signal, continue} to member functions
* Convert standolen job_get{,_from_pid} to `job_t` static functions
* Reduce usage of JOB_* enums outside of proc.cpp by using new
`job_t::is_foo()` const helper methods instead.
This patch is only a refactor and should not change any functionality or
behavior (both observed and unobserved).
* Debug level 3: describe all commands being executed (this is, after all,
a shell and one can argue that this is the most important debug
information avaliable)
* Debug level 4: details of execution, mainly fork vs no-fork and io
handling
Also introduced j->preview() to print a short descriptor of the job
based on the head of the first process so we don't overwhelm with
needless repitition, but also so that we don't have to rely on
distinguishing between repeated, non-unique/non-monotonic job ids that
are often recycled within a single "execution cycle" (pressing enter
once).
Per @ridiculousfish's suggestions in #5219,
`process_mark_finished_children()` has been updated to work in an easier-
to-follow manner. Its behavior is now straight forward, it always checks
for finished processes but only blocks if `block_on_fg` is true.
We're not using the SIGCHLD count in s_sigchld_generation_cnt for
anything any more, as it's not actually a reliable metric since we can
experience one SIGCHLD as a result of two processes exiting (see #1768),
but only reap one of them if the other is in a not-fully-constructed job
(see #5219), a state we cannot possibly detect without calling
`waitpid()` on all child processes, which we are explicitly avoiding.
We never insert elements into the middle of a job list, only move
elements to the top. While that can be done "efficiently" with a list, it
can be done faster with a deque, which also won't thrash the cache when
enumerating over jobs.
This speeds up enumeration in the critical path in
`process_mark_finished_children()`.
* Instead of reaping all child processes when we receive a SIGCHLD, try
reaping only processes belonging to process groups from fully-
constructed jobs, which should eliminate the need for the keepalive
process entirely (WSL's lack of zombies not withstanding) as now
completed processes are not reaped until the job has been fully
constructed (i.e. all processes launched), which means their process
group should still be around for new processes to join.
* When `tcgetpgrp()` calls return 0, attempt to `tcsetpgrp()` before
invoking failure handling code.
* When forking a builtin and not running interactively, do not bail if
unable to set/restore terminal attributes.
Fixes#4178. Fixes#3805. Fixes#5210.
This is to avoid development versions of fish 3.0 freaking out when the
file format is changed. We now have better support for for future universal
variable formats so it's unlikely we'll have to change the file name again.
We do a bunch of escaping before to make `eval` work, and that needs to be removed as well or fragment-urls don't work.
This reverts commit e9568069a7.
Use clang/clang++'s own autocompletion support to complete arguments. It
is rather convoluted as clang generates autocompletions for a portion of
the current token rather than the entire token, e.g. while `--st` will
autocomplete to `--std=` (which is fine by fish), `--std=g` will
autocomplete to `gnu...` without the leading `--std=` which breaks fish'
support for the completion.
Additionally, on systems where clang/clang++ is the system compiler
(such as FreeBSD), it is very often for users to invoke a newer version
of clang/clang++ installed as clang[++]-NN instead of clang. Using a
monkey-patched version of `complete -p` to support that without breaking
(future) completions for commands like `clang-format`.
Closes#4174.
In private mode, access to previous history is blocked and new history
does not persist and is only available for the duration of the current
session.
This mode can be used when it is not desirable for commandline history
to leak into a session, e.g. via autocomplete or when it is desirable to
test the behavior of fish in the absence of history items without
permanently clearing the history.
I'm sure there are a lot more features that can be incorporated into
private mode, such as restricting access to certain user-specific
configuration files, etc.
This addresses a lot of the concerns raised in #1363 (which was later
changed to track mosh-specific problems). See also #102.
When we discard output because there's been too much, we print a
warning, but subsequent uses of the same buffer still discard.
Now we explicitly reset the flag, so we warn once and everything works
normal after.
Fixes#5267.
For things like
source $undefined
or
source (nooutput)
it was quite annoying that it read from tty.
Instead we now require a "-" as the filename to read from the tty.
This does not apply to reading from stdin if it's redirected, so
something | source
still works.
Fixes#2633.
This adds flags --path and --unpath to builtin set, analogous to
--export and --unexport. These flags change whether a variable is
marked as a path variable.
Universal variables cannot yet be path variables.
This switches quoted expansion like "$foo" to use foo's delimiter instead of
space. The delimiter is space for normal variables and colonf or path variables.
Expansions like "$PATH" will now expand using ':'.
This commit begins to bake in a notion of path-style variables.
Prior to this fix, fish would export arrays as ASCII record separator
delimited, except for a whitelist (PATH, CDPATH, MANPATH). This is
surprising and awkward for other programs to deal with, and there's no way
to get similar behavior for other variables like GOPATH or LD_LIBRARY_PATH.
This commit does the following:
1. Exports all arrays as colon delimited strings, instead of RS.
2. Introduces a notion of "path variable." A path variable will be
"colon-delimited" which means it gets colon-separated in quoted expansion,
and automatically splits on colons. In this commit we only do the exporting
part.
Colons are not escaped in exporting; this is deliberate to support uses
like
`set -x PYTHONPATH "/foo:/bar"`
which ought to work (and already do, we don't want to make a compat break
here).
This reverts commit 3f820f0edf.
While the premise described by @nbuwe is sound in #4505, we are now
apparently relying on this behavior is some places (although
inadvertently as there doesn't seem to be a deliberate acknowledgement
of that anywhere).
Turning off ONLCR causes things like indented multiline commands to not
appear correct at the tty (subsequent lines appear both at column 0 and
again indented).
Per @nbuwe's excellent explanation in #4505, we can save on output
to the tty by maintaining column location after NL by disabling the
ONLCR terminal mode.
Closes#4505.
Adds a new match mode for `string_fuzzy_match_t` that matches against a
case-insensitive subsequence within a string, e.g. `LL` now (partially)
matches against `hello`. This is implemented as a separate mode, given a
lower priority of match than a same-case match (when present).
Note that `fuzzy_match_subsequence_insertions_only` has purposely not
been extended with a case-insensitive version as that would be a)
unlikely to match often, and b) adding a second inefficient fuzzy search
to something that's queried a lot. Perhaps `subsequence_insertions_only`
can simply be changed to be a case-insensitive comparison in the future?
Closes#1196. Affects #3978.
If you're using the old binding that only clears the commandline and
doesn't preserve its contents and start a new line, you can use
```fish
bind \cc "commandline -f cancel; commandline ''"
```
instead.
Closes#4298.
For some weird reason we only used $editor if it wasn't empty, but
then failed to fail if it was.
This will now print an error and use fish, just like if the $EDITOR
value is invalid in any other way.
Fixes#5257.
Came in handy for tracking down the performance regression in #5219. This will
take the output of two (necessarily identical) `fish --profile ...` runs and
produce a third profile log in which all times are the difference between the
first and the second profile provided.
(I'm not sure if build_tools is the right place for it, but I think it's OK?)
This is a wrapper that calls kitty to dynamically provide completions,
as generated by kitty itself, via `kitty + complete setup fish`.
ref: https://sw.kovidgoyal.net/kitty/#fish
Load fish docs and configuration out of the source and/or build
directories rather from the installed paths when running directly out
of the cmake build directory.
Closes#5255.
Prior to this fix, fish would swallow SIGINT in non-interactive mode. This
meant that scripts could only be Ctrl-C'd if fish was executing an external
command.
Unblock SIGINT in non-interactive mode.
Fixes#5253
Fixes broken macOS build. I'm not sure how the code used to compile
without including `dyld.h` previously, perhaps a different header used
to pull it in?
Retrieves the fully resolved path to the currently executing fish binary
(regardless of PATH). Can be used to ensure that the same fish is
launched again from a script.
`get_executable_path()` moved from fish binary to libfish, also cleaned
up some duplicated (but differing!) definitions of PATH_MAX (which was
used by that function) in the process.
Remove dependency on the Linux compatibility layer's procfs being
installed and mounted when running under FreeBSD by directly querying
the MIB for the path to the running fish executable
(KERN_PROC_PATHNAME). Tested under FreeBSD 11.2-RELEASE.
The hg prompt walks up the directory hierarchy to decide if we are in a
repo subdirectory. Because hg is an external command, it resolves symlinks.
Switch to using pwd -P so hg and fish will have the same view of the hg repo.
Based on comment:
https://github.com/fish-shell/fish-shell/pull/5190#issuecomment-421912360
This switches fish to a "virtual" PWD, where it no longer uses getcwd to
discover its PWD but instead synthesizes it based on normalizing cd against
the $PWD variable.
Both pwd and $PWD contain the virtual path. pwd is taught about -P to
return the physical path, and -L the logical path (which is the default).
Fixes#3350
This new function performs normalization of paths including dropping
/./ segments, and resolving /../ segments, in preparation for switching
fish to a "virtual" PWD.
There are a few opportunities to improve the formatting as well as a
handful of typos in this document. I was looking into contributing and
noticed that it might be worthwhile to address them.
I just submitted a PR to fix a few issues in CONTRIBUTING.MD, so I took
a few minutes too look over README.md as well. This is the only room
for improvement I noticed.
Mostly resolves#4862, though there remains the lingering question of
whether or not to emit a warning to /dev/tty or stderr when a
non-literal-zero index evaluates to zero.
An update to `CMakeLists.txt` set the default build type to
`RelWithDebInfo`, so there's no need to tell users to consider appending
`-DCMAKE_BUILD_TYPE=Release` at all.
[skip ci]
Coalesces commands with leading (if even possible) and trailing
whitespace into the same item, improving the experience when iterating
over history entries.
Closes#4908.
This allows for marking certain bindings as part of a preset, which allows us to
- only erase those when switching presets
- go back to the preset binding when erasing a user binding
- only show user customization if requested
- make bare bind statements in config.fish work (!!!11elf!!!)
Fixes#5191.
Fixes#3699.
This reverts commit 8c14f0f30f.
This list is not reliable - there are many ways for fish to quit that does not
invoke these functions. It's also not necessary since the history is correctly
saved on exec.
Prior to this change, env_get_pwd_slash() would try to infer the PWD from
getcwd() if $PWD were missing. But this results env_get_pwd_slash() doing
something radically different than $PWD, and also is a lot of code for a
scenario that cannot be reliably reproduced. Just return "/" in this case.
If the replacement in `string replace` is invalid, prior to this fix we would
enter into an infinite loop trying to parse it. Instead report errors correctly.
Fixes#3381
Rather than having tokenizer_error as pointers to objects, switch it back
to just an error code value. This makes reasoning about it easier since
it's immutable values instead of mutable objects, and it avoids allocation
during startup.
At some point the completion code was refactored and in the event where
no explicit function description was passed into `resolve_description()`
it would attempt to use the `desc_func` parameter but pass in the
_remaining_ part of the completion rather than the full text, which
would obviously fail.
e.g. if completing `foo<TAB>`, for function `foobar` it would attempt to
find the description for a function named `bar` instead of `foobar`.
Closes#5206.
This reverts commit 9c63ad3209 until I can
figure out what is causing the assertion and test failures.
It *seems* to be that passing in the correct function name to the
description lookup is causing a previously present error to be realized,
but I can't yet be certain.
At some point the completion code was refactored and in the event where
no explicit function description was passed into `resolve_description()`
it would attempt to use the `desc_func` parameter but pass in the
_remaining_ part of the completion rather than the full text, which
would obviously fail.
e.g. if completing `foo<TAB>`, for function `foobar` it would attempt to
find the description for a function named `bar` instead of `foodbar`.
Closes#5206.
There's been no reproducible case entered for #5080, but the stack trace
indicates the problem is with env_get_pwd_slash() returning an empty
string, which isn't a string that terminates in `/`.
In addition to making the failure case to return the path `./` (which
has the benefit of having the same meaning as $PWD), trying a little bit
harder to retrieve the real PWD by using getcwd(3). While
get_current_dir(3) is documented as relying on PWD, getcwd(3) does not
mention any such caveats, so it's possible that it will work even if
something is breaking PWD.
Just a thought, but it's possible if due to some recursion PWD surpassed
some predetermined value (maybe PATH_MAX) that PWD (on certain platforms
or under certain enivronments) won't be set (hence the code that deals
with ERANGE errors from the getcwd(3) call).
Closes#5080.
As reported in fish-shell/fish-shell#5180, when the USER environment
variable is not set and fish is started, `get_runtime_path()` returns a
blank string. At some point in the past, this was called after
`setup_user()` in env.cpp, but this is no longer the case.
This commit removes the reliance on the $USER environment variable
entirely, and instead uses `getpwuid(geteuid()).pw_name` to retrieve the
current username.
Closes#5180.
For some reason I started getting literal \n appearing in Doxygen-generated
help files. These are coming from newlines in aliases defined in
Doxyfile.user. These should be safe to remove because they are HTML-specific
and there is still whitespace before them. Remove these newlines.
This didn't reproduce on Linux; Doxygen is full of mysteries.
`exec` now exhibits the same behavior as `exit` and prompts the user to
confirm their intention to end the current process if there are
background jobs running. Running `exec` again immediately thereafter
will force the exec to go through.
Additionally, background jobs are reaped upon exec to prevent process
leaking (same as `exit`).
Fix#5133 changed builtins to acquire the terminal, but this regressed
caused fish to be stopped when running in background via `sudo fish`.
Fix this by only acquiring the terminal if the terminal was owned by the
builtin's pgroup.
Fixes#5147
- Add support for:
- Jumping to the character before a target.
- Repeating the previous jump (same direction, same precision).
- Repeating the previous jump in the reverse order.
- Enhance vi bindings.
When running a builtin, if we are an interactive shell and stdin is a tty,
then acquire ownership of the terminal via tcgetpgrp() before running the
builtin, and set it back after.
Fixes#4540
Factor the history search fields into a new class.
As a side effect, this shares the deduplication logic, so that token search
no longer returns duplicates.
Fixes#4795
This changes the behavior of builtin math to floating point by default.
If the result of a computation is an integer, then it will be printed as an
integer; otherwise it will be printed as a floating point decimal with up to
'scale' digits past the decimal point (default is 6, matching printf).
Trailing zeros are trimmed. Values are rounded following printf semantics.
Fixes#4478
The Informative VCS sample prompt currently sets the `__fish_git_prompt_char_conflictedstate` variable which is unused.
It should instead set the `__fish_git_prompt_char_invalidstate` variable.
Ordering of directories above files was introduced in a recent change to
the same script. By default it does not matter as completions are sorted
by fish internally, but this allows the use of `-k` to sort files before
directories (or piped to `sort -r` for vice-versa).
Use `apt-cache show` instead of `apt-cache packagenames` to efficiently
print package names and a brief description instead of the placeholder
(localized) "Package" text that was previously printed. This applies to
both available and installed packages (for inistall and remove operations,
respectively).
TODO: update `__fish_print_packages` for non-debian platforms to do the
same.
When listing packages already installed (e.g. for use with `apt remove
...`), do not consider packages return by `dpkg --get-selections` with
state 'deinstall'.
Previously the `string replace` pattern was matching both 'install' and
'deinstall' packages.
This reverts commit e2a3dae58b.
This idea failed because ./share was not complete when bliding via cmake;
it misses critical files such as config.fish.
fish tries to be relocatable by looking for directories relative to its
executable. These directories are not found when running fish from
within a cmake build because the etc directory is not present. Stop requiring
this directory to be present since it's not critical for running fish.
Fixes#4825
Don't mmap history files on remote file systems
This merges some changes to history that may help to mitigate the crashes seen in #5088 . These SIGBUS crashes occur when reading a memory mapping whose underlying file was truncated. It's not clear why this should occur more often on NFS (or ever). However memory mapping over NFS is sketchy anyways so this is desirable regardless.
Migrate the mmap() logic into a new class history_file_contents_t which
will serve to encapsulate conditional logic if we choose to use read()
instead of mmap().
Utilized the `--install` flag added in commit #8c09d6e.
Limit `eopkg remove/autoremove/check ...` completions to installed packages.
Limit `eopkg install/upgrade/info ...` completions to available packages.
Prior to this fix, __fish_describe_command would error if the
input contained any special characters, because it would be interpolated
into a regex. Hack in a guard to do nothing if the input contains
anything other than [a-zA-Z0-9_ ]
* update nim.fish sample prompt
- Use an helper function to wrap informations
- Add VIRTUAL_ENV infos, if any
- Add __fish_git_prompt, wrapped for the theme
- Add comments
- Remove ASCII failback symbols for tty
(no more useful for me, but if someone really needs it, just ask)
* fish.nim: test -n __fish_git_prompt
Added a new flag `--installed` via `argparse` to `__fish_print_packages`
which indicates that only installed packages should be listed.
TODO: Other non-debian/apt platforms should take advantage of this flag/
behavior as well.
This adds a new string command split0, which splits on zero bytes.
split0 has superpowers because its output is not further split on
newlines when used in command substitutions.
separated_buffer_t encapsulates the logic around discarding (which
was previously duplicated between output_stream_t and io_buffer_t),
and will also encapsulate the logic around explicitly separated
output.
It was only introduced in 2.16, which was released in January 2018.
Instead, we just use a bare "--ignored", which is equivalent to "--ignored=traditional".
The difference to "--ignored=matching" mode shouldn't matter to us here.
Fixes#5074.
Executes `whatis` safely, returns at most one line, and strips the name
of the command from the start of line, returning a value fit for use as
the description parameter for a completion argument value.
Fixes
- Use the actual path when skipping unusable paths to fix all Include
directives being skipped when there is no ~/.ssh directory
- Prevent "No matches for wildcard" message
Improvements
- Skip paths that are directories since we only want files
- Remove `cd` as it is not needed
__fish_complete_suffix assumed that the only literal . in a path
would be the . before an extension, and stripped accordingly. This
behavior has been there for a long time, but broke many things
including completion of relative paths and completion of paths with
a literal . in a directory name.
__fish_complete_suffix does not just complete extensions (or at the
very least, it no longer does just that) but rather any suffix, so
isolating the path name without the extension was unnecessary in all
cases.
If just one of the range ends is negative, this now forces direction away from it.
I.e. if the beginning is negative, we go in reverse.
If the end is negative, we go forwards.
This fixes cases like
$var[2..-1]
if $var only has one element.
I'm not sure what was wrong with the old syntax, but I needed to switch
the outer quotes to ' and the inner quotes to " in order for the
completions to work when they weren't explicitly sourced.
Additionally, realized that the overload for __fish_complete_suffix can
be used to get the filtered list of kernel modules from /boot/kernel in
the initial run.
Allows the most painful of curl's arguments to be completed by fish by
restoring file-based completions for paths prefixed with `@` (which are
typically used after parameters like --data).
With a blank $suff (i.e. complete all files), __fish_complete_suffix
returned directories twice, once with the trailing `/` and once without.
This fixes that, and additionally speeds up the code by no longer
shelling out to `sort -u` as we no longer rely on brace expansion to
enumerate directories and files simultaneously.
In general, this behavior would occur when a directory exists that
matches the suffix search pattern (so a dir named 'foo.bar' with a
search pattern '.bar' would return 'foo.bar' twice).
Runtime has dropped from ~22ms to ~8ms on my machine, while also
returning more correct results.
Default to RelWithDebInfo (-O2 -g) if no custom CMAKE_BUILD_TYPE is
defined. Also add flags for use with CMAKE_BUILD_TYPEs Debug, Release,
and RelWithDebInfo.
While supported by gcc and clang, \e is a gcc-specific extension and not
formally defined in the C or C++ standards.
See [0] for a list of valid escapes.
[0]: https://stackoverflow.com/a/10220539/17027
We've tried numerous approaches to mitigate the race condition between
`posix_spawn` and the `setpgid` call, but unfortunately due to the flags
we pass to `posix_spawn`, it (rarely? never?) results in `vfork()` being
used, which means it is never executed atomically. Since it is executed
out-of-band, we must manually call `setpgid` in case `posix_spawn`
hasn't gotten around to doing that yet, but in the event that it has, an
EACCES error can be returned.
Closes#4884. Closes#4715. See also #4778.
On systems where the terminfo for TERM does not contain a string for
attributes such as enter_underline_mode, etc. fish was crashing with a
fatal error message and a note to email the developers.
These are non-essential text attribute changes and should not trigger
such a failure.
This allows snippets to use everything that is defined in config.fish,
which is our _base_ initialization.
Among other things, it enables snippets to use $PATH as it will appear
in the user's config.fish, or even to change $PATH.
Also, this is how it was in 2.7.1 and before (with the small change
that abbrs were upgraded after).
As defined in the `go help packages`:
Many commands apply to a set of packages:
go action [packages]
Usually, [packages] is a list of import paths.
This patch introduces automatic lookup of said packages from GOPATH
using `go list`, and provides them as options go subcmds.
I'm not sure what was up with the old completions,
`$__fish_service_commands` is not set anywhere and completions for the
command (not the service) were not being generated on my machine.
macOS and (AFAICT) most Linux distributions ship with the Info-ZIP
version of unzip, which has the `unzip -h` flag; but other
implementations of unzip do not necessarily have it (i.e. FreeBSD).
`unzip` under FreeBSD does not support `unzip -h`. Under both Linux and
FreeBSD, `unzip -v` presents the list of options, though. Using this
instead of `unzip -h` to detect the Debian-patched version of the
Info-ZIP unzip program.
[9/13] Building CXX object CMakeFiles/fishlib.dir/src/builtin_string.cpp.o
../src/builtin_string.cpp:1221:12: warning: mangled name of 'string_transform' will change in C++17 due to non-throwing exception specification in function signature [-Wc++17-compat-mangling]
static int string_transform(parser_t &parser, io_streams_t &streams, int argc, wchar_t **argv, decltype(std::towlower) func) {
^
1 warning generated.
This prints an escape sequence, so it can break scp or similar when
someone has an unqualified
fish_vi_key_bindings
in config.fish and happens to run a terminal that can set the cursor.
Our completion machinery calls our `__fish_describe_command` function
to describe commands via apropos. Only it trusts the output a bit too
much, so it crashes when any line from that is shorter than the
original string.
Fix this by skipping any string that is shorter than the original,
since it can't be a match anyway.
Also stop doing wcslen so often - std::strings are nice!
Fixes#5014.
These completions are apparently based on an auto-generated version,
so there's a whole bunch of rewording to be done here.
Also for some reason some of the options are mentioned more than once?
I can't seem to find a reason why the shell interpreter needs to be bash
and not just sh here. Needed to replace `BASH_SOURCE[0]` with the legacy
`$0` supported by sh, but otherwise it seems to still work.
Many non-Linux platforms do not ship with bash out-of-the-box (and as a
shell, I don't think we need to encourage the further proliferation of
bash ;-), this lets fish build on a clean install of FreeBSD, which does
not have bash.
There really is no need to
- Timeout just because the _first_ character was a control character
- Timeout because of any control character other than escape
The reason to timeout because the '\e' sequence can appear by itself (signifying
pressing the escape key) and still make
sense - e.g. vi-mode has it bound to a rather important function!
But a \c can't appear by itself, so we can just block.
This allows binding sequences like \cx\ce and inputting them at a
leisurely pace rather than the frantic escape_timeout one.
It should also improve sequences that _include_ escape somewhere else.
E.g. something like a\eb ("a, then alt+b") should now time out for the "\eb" part,
allowing users to bind a\e ("a, then escape") to something else. Why you'd want to do
that, I have no idea. But it's more consistent, and that's nice!
For regex-mode, this should be enough to read NUL-delimited strings to act on, but not
quite patterns and replacements.
Glob-mode requires more work - it uses wcscmp internally, which is unsuitable.
Also the various styles have one function each with barely any
difference - mostly passing the corresponding STYLE argument.
Pack them into one function for escape and one for unescape to save
about 100 lines.
We're now actually handling wchar_t here, so comparing the 0x80 bit
would break for UTF-16, causing ASCII false-positives.
Also simplifies a bit, since we no longer need a second variable.
printf 'a\0b' | string length
used to print "1". Now it prints "3".
Note that this switches to using C++'s std::string::length, which
might give differing results.
Under FreeBSD, as annoying as it is, switches must directly follow the
command or subcommand in question, and cannot come after actual payload
argument. Calling `zpool get all -H` instead of `zpool get -H all`
caused error messages to be spewed to the console under FreeBSD when
simply completing `zfs <TAB>`, this should fix that. The change should
also be compatible with other operating systems (namely Linux) that
don't have this requirement, as they (generally) allow arguments to come
before _or_ after the primary non-switch argument (though I do not have
access to a zfs-enabled Linux machine to test this).
Previously, trying to complete a token with any of these
expansion-related characters would cause the completion to return no
results, as it would emit expanded values which weren't matched by the
autocompleter.
Only the first non-switch parameter to python must be a .py file, but
everything thereafter is "just another argument". This enables file
completions for 2nd+ arguments.
Akin to __fish_complete_suffix, __fish_complete_directories now attempts
to complete the current commandline token if no token is explicitly
passed in as an argument.
The prompt is a fallback that is overridden via a function file
anyway.
Do that with the title as well, so we can use just builtins.
This removes error messages when $fish_function_path is borked.
Introduced by #4849 (add wait for processes by name)
../src/builtin_wait.cpp:23:14: warning: using the result of an assignment as a condition without parentheses [-Wparentheses]
while (j = jobs.next()) {
~~^~~~~~~~~~~~~
../src/builtin_wait.cpp:23:14: note: place parentheses around the assignment to silence this warning
while (j = jobs.next()) {
^
( )
../src/builtin_wait.cpp:23:14: note: use '==' to turn this assignment into an equality comparison
while (j = jobs.next()) {
^
==
1 warning generated.
Turns out that `make -pn` actually takes a while - about 300ms on
fish's makefile.
That's quite a bit of time just to throw away the output and use the
exit code.
So we just check for "GNU" in the version string.
It would be nice to just _do_ the completion and fall back on the
BSD-style if it doesn't work, but that is tricky to do with the pipe
to `awk` - the awk expression actually does not fail if `make` does
not print output.
And I don't know enough about awk to change that.
While this is a bit faster (mostly because it needs less processing on fish's side),
it lacks the neat description bit and the ":/" stuff doesn't work.
The boost is also not large in absolute terms (a few milliseconds).
This reverts commit 1f8e4dad9f.
This uses the same logic that git uses to determine the satus of files
and doesn't require any parsing on our end. Brings in support for
relative paths (such as `git add ../f<TAB>`). Should be faster and more
reliable than manually parsing porcelain status.
This doesn't support as many cases as the old `__git_ls_files` function
did (e.g. `renamed` is not supported, nor is `added`), both of which
_can_ be implemented on top of the new logic - but neither of which were
actually being used, anyway.
Usefulness is decreased by #4970, speed still bottlenecked by #4969.
cc @faho
This is based on what the official git completions do, and it's quite
fast.
Also only complete files after a "--" separator for `checkout`.
Fixes#4858.
This is much quicker - on the order of 100ms vs 50ms.
We shorten to 10 characters, which is statistically suitable - 3 out
of 600k commits in the linux kernel need 11 characters.
Turns out the segfaults we've been getting in our tests are because we set $TERM to "dumb".
So we only clear the line if the terminal isn't dumb.
This reverts commit 745a88f2f6.
Fixes#2320.
One key use of process expansion, used in currently-shipped code, is for running a function on
current shell exit.
Restore the use of %self as a valid argument (and add `self`) and document this change.
(faho: Remove bare "self")
For usage in completion scripts.
Unlike `__fish_is_first_token` (which is probably not correctly named),
`__fish_is_first_arg` returns true regardless of whether existing tokens start with `-`
or not, to be used when an arg cannot be used with any other argument.
`__fish_prev_arg_in` is similar to `__fish_seen_...` but it explicitly
tests the preceding token only, for arguments that take only a single
parameter.
This enables users to opt in (or out) of specific features by setting
the fish_features environment variable.
For example `set -U fish_features stderr-nocaret` to opt into removing the
caret redirection.
This partially reverts 5b489ca30f, with
carets acting as redirections unless the stderr-nocaret flag is set.
This flag is off by default but may be enabled on the command line:
fish --features stderr-nocaret
This introduces a new command line option --features which can be used for
enabling or disabling features for a particular fish session.
Examples:
fish --features stderr-nocaret
fish --features 3.0,no-stderr-nocaret
fish --features all
Note that the feature set cannot be changed in an existing session.
This teaches the status command to work with features.
'status features' will show a table listing all known features and whether
they are currently on or off.
`status test-feature` will test an individual feature, setting the exit status to
0 if the feature is on, 1 if off, 2 if unknown.
This introduces a new type features_t that exposes feature flags. The intent
is to allow a deprecation/incremental adoption path. This is not a general
purpose configuration mechanism, but instead allows for compatibility during
the transition as features are added/removed.
Each feature has a user-presentable short name and a short description. Their
values are tracked in a struct features_t.
We start with one feature stderr_nocaret, but it's not hooked up yet.
As it turns out, for some terminals backspace is \b but only when
preceded by \e.
All this makes about as much sense as the english language.
Fixes#4955.
This was done in share/config.fish, but leads to surprising results if
that isn't read - e.g. because someone just built fish in the git
directory to test it without installing.
It's also not something that is any more or less complicated.
For compatibility, keep it in config.fish as well for the time being.
The previous completion generation was broken for several reasons:
* ./foo would break detection of suffix due to the leading . being
interpreted an extension marker,
* ./foo would be completed as foo, which would be excluded from
matching inrcomplete.cpp
Using `git for-each-ref` both simplifies the code (no need to deal
with detached heads anymore) and speeds it up.
With 1600 branches, the time goes from ~48ms to ~16ms.
- fix capitalization
- shorten descriptions
- implement subcommand shortcuts
- add arg completion for 'limit' and 'depth' switches
- improve arg completion for list subcommand in case of -p switch
bower was calling `__fish_should_complete_args`, the old name for
`__fish_should_complete_switches.`
yarn was parsing bower.json instead of package.json.
To be used by completions to directly determine whether it is either
possible or preferable to complete a switch (instead of a subcommand),
(presuming that switches must come before subcommands).
* __fish_can_complete_switches: we are in a position where a switch may
be placed.
* __fish_should_complete_switches: we're in a position to accept a
switch and the current token starts with `-` so we have no choice but
to do so.
Selectively reverts 156d4fb9b9.
`all-the-package-names` is still used to generate completions for `npm`
if it is installed, but it is not manually installed nor updated. It is
now the user's responsibility to do both, and it must be installed
globally.
`npm search` was _way_ too slow to be used for dynamic completions, so
using a cached list of all avaialable NPM packages to match against.
This is a bit brave for a fish completion, but the npm package
`all-the-package-names` has a list of, well, all the package names
avaialable for installation via the default npm registry. Installing a
copy locally to $HOME/.cache/fish/npm_completions and using that to
search for packages matching the tokenized command line.
Preference would be to call `__update_atpm` in the background, but that
emits an ugly "job has completed" message..
Should also use this for completions for `yarn add`.
Instead, attempt to extract the message that _would_ be displayed on
execution of `./configure --help` by relying on some markers present in
autoconf-generated configure files.
As measured with 'hyperfine' on a laptop running in reduced frequency
power savings mode, `fish -c "__fish_parse_configure ./configure"`
runtime dropped from ~1.25s to ~0.8ms, which is inline with the
previously observed ~350ms execution time for `./configure --help`.
fish's own startup time is approximately 75ms before parsing begins.
Still very slow, but much better.
This relies on the new `read --line/-L` support as an entire parser for
the output of `./configure --help` was written in fishscript. Also
doesn't work without 72f32e6d8a7905b064680ec4b578c41dea62bf84.
The completion script is slow... a function of both the autotools
configure script itself being written in a shell script combined with a
fishscript output parser.
fish's own `./configure --help` takes around 350ms to execute, while
`__fish_parse_configure ./configure` (which runs that behind the scenes)
takes around 660ms to run, all-in-all - a not insignificant overhead.
Output can be cached (based off of ./configure hash or mtime) in the
future if this is a big deal.
complete.cpp strips the path from commands before parsing for
completions, meaning that when we called `path_get_path()` against
`cmd`, if `./cmd` were typed in at the command line but `cmd` does not
exist in the PATH, then the command would incorrectly be flagged as not
present and the completions would be skipped.
This is also faster when an absolute/relative path is used for a
command, as we now search with the original path which skips searching
PATH directories unnecessarily.
Found when debugging why completions for `./configure` wouldn't work.
`read` with IFS empty was expected to set all parameters after the first
n filled variables to an empty string, but that was inconsistent with
the behavior of `read` everywhere else.
I'm not sure why fish differed from the spec with regards to the
behavior in the event of an empty IFS: we eschew IFS where possible, yet
here we adopt non-standard behavior splitting on every (unicode)
character instead of not splitting at all with IFS empty. We still do
that, but now the unset variables are treated as they normally would be,
i.e. cleared and not set to an empty string (which is what an empty
value between two IFS separators would contain).
The default completions that autojump ships with for fish are broken
(emitting output like "1\___\#...") as they use hackes to work around
the previous lack of `complete -k`. The history-based autojump
completions fully replace it.
The job expansion wrapper was swallowing `-n` (and presumably `-e` and
others) when that was the literal argument we needed to emit. Using
`printf %s ...` instead.
This brings back expansion of `%n` where `n` is a job id, but not as a
general parser syntax. This makes `jobs -p %n` work, which can be used
as part of the job control command chain, i.e.
```
cat &
fg (jobs -p %1)
```
fg/bg/wait can either be wrapped in a function to call `jobs -p` for
`%n` arguments, or they can be updated to take `%n` arguments
themselves.
The order of this list does not need to be strictly maintained any
longer.
Benchmarked with `hyperfine` as follows, where `bench1` is the existing
approach of binary search and `bench2` is the new unordered_set code,
(executed under bash because fish would always return non-zero). The
benchmark code checks each argv to see if it is a builtin keyword (both
return the same result):
```
hyperfine './bench1 $(shuf /usr/share/dict/words)' './bench2 $(shuf /usr/share/dict/words)'
Benchmark #1: ./bench1 $(shuf /usr/share/dict/words)
Time (mean ± σ): 68.4 ms ± 3.0 ms [User: 28.8 ms, System: 38.9 ms]
Range (min … max): 60.4 ms … 75.4 ms
Benchmark #2: ./bench2 $(shuf /usr/share/dict/words)
Time (mean ± σ): 61.4 ms ± 2.3 ms [User: 23.1 ms, System: 39.8 ms]
Range (min … max): 58.1 ms … 67.1 ms
Summary
'./bench2 $(shuf /usr/share/dict/words)' ran
1.11x faster than './bench1 $(shuf /usr/share/dict/words)'
```
Now the description includes the variable scope, `set [-e] -[Ugl]`
completions only provide variables matching that scope, and completions
that shouldn't be modified are hidden from the user. Completions that
are often modified but rarely unset (`fish_*` variables) are omitted
from `set -e` completions.
A new helper function `__fish_seen_argument` has been added that makes
it easy to only provied completions for a specific flag.
Launch `cmd.exe /c "start URL"` under WSL for both `fish_config` and
`help`. This works around #4299 but does not address the underlying
issue (#1132).
Prior to this fix, the fish universal variables file claimed that
changes to it would be overwritten. This no longer true and has not
been true for a long time. Remove that warning.
This switches the universal variables file from a machine-specific
name to the fixed '.config/fish/fish_universal_variables'. The old file
name is migrated if necessary.
Fixes#1912
This removes the caret as a shorthand for redirecting stderr.
Note that stderr may be redirected to a file via 2>/some/path...
and may be redirected with a pipe via 2>|.
Fixes#4394
The previous commit caused the tests to fail since env_remove() was
returning a blanket `!0` when a variable couldn't be unset because it
didn't exist in the first place. This caused the wrong message to be
emitted since the code clashed with a return code for `env_set()`.
Added `ENV_NOT_FOUND` to signify that the variable requested unset
didn't exist in the first place, but _not_ printing the error message
currently so as not to break existing behavior before checking if this
is something we want.
Variables set in if and while conditions are in the enclosing block, not
the if/while statement block. For example:
if set -l var (somecommand) ; end
echo $var
will now work as expected.
Fixes#4820. Fixes#1212.
fish reads paths out of /etc/paths.d. Prior to adbaddf it did
this on every shell invocation; with adbaddf it does so on only login
shells. This change wasn't justified so let's revert this behavior.
Currently, there are two possibilities for holes in the background:
- When there are two candidates with the same meaning (a long and a
short option or two candidates with the same description)
- When a candidate does not have a description (meaning the color
won't continue after it)
This changes both so the background just goes on.
In addition, it avoids making the background multiple times.
Fixes#4866.
The official fish documentation makes no mention of how `string split`
treats empty tokens, e.g. splitting 'key1##key2' on '#' or (more
confusingly) splitting '/path' on '/'. With this commit, `string split`
now has an option to exclude zero-length substrings from the resulting
array with a new `--no-empty/-n`. The default behavior of preserving
empty entries is kept so as to avoid breakage.
The two unicode glyphs used to represent missing new lines and redacted
characters for secure entry are both not present in the glyph tables of
the default font under Windows (Consolas and Lucida Console), use an
alternative glyph instead.
The "return" symbol is replaced with a pilcrow (¶) and the "redacted
character" symbol is replaced with a bullet (•). Both of these are
well-defined in almost all fonts as they're very old symbols. This
change only takes place if -DWSL is supplied by the build toolchain.
Note: this means a Windows SSH client connecting to a fish remote
instance on a non-Windows machine will still use the (unavailable)
default glyphs instead.
(and /etc/paths.d/*)
Do so by emulating the behavior of /usr/libexec/path_helper for login
shells, matching the behavior in /etc/profile.
Also add a path_helper command to reproduce the behavior of
/usr/libexec/path_helper for fish.
This also handles setting MANPATH if necessary.
Fixes issue #4336
* Completion for conda, the package manager
* Make the list of platforms a private variable
* Add commands activate and deactivate
* Avoid clobbering a user-defined function __
* Use Use __fish_seen_subcommand_from to identify subcommand
And treat the case of the first argument as a special case
with function __fish_conda_fist_arg
* Factor out create from loop for option --name
* Fix typo (missing parenthesis in description)
* Start from a blank state by removing completions from conda configuration script
* Make wcwidth configurable
This adds the cmake option "INTERNAL_WCWIDTH" (to be set to "ON" or
"OFF") and the configure option --[en,dis]able-internal-wcwidth.
Both default to enabling our fallback, but can be set to use the system's wcwidth again.
Sequel to #4554.
See #4571, #4539, #4609.
On my system, this would fix#4306.
The newly added `:` command is implemented as a function (to avoid
increasing complexity by making it a builtin), but it is saved to a path
that does not match its filename (since its name is somewhat of a
special character that might cause problems during installation).
Directly probing the `colon` function for autoload causes `:` to be
correctly loaded, so doing just that after function paths are loaded
upon startup.
This is a hack since the CPP code shouldn't really be aware of
individual functions, perhaps there is a better way of doing this.
no-op function for compatibility with sh, bash, and others.
Often used to insert a comment into a chain of commands without having
it eat up the remainder of the line, handy in Makefiles.
Fixes an issue introduced in 4414d5c888
where functions loaded from custom directories are not detected as being
valid for purposes of determining whether or not completions should be
called.
Restore localization to tokenizer error strings Work around #4810 by retrieving localizations at runtime to avoid issues possibly caused by inserting into the static unordered_map during static initialization. Closes#810.
Work around #4810 by retrieving localizations at runtime to avoid issues
possibly caused by inserting into the static unordered_map during static
initialization.
Closes#810.
- Cache translations instead of calling `gettext` once per file
- Only do the ":/" thing if the file isn't in $PWD/**
For a git repo created like
```fish
git init
touch a(seq 0 1000)b
```
this changes the time from about 2s to 0.3s.
`git rm --cached` is often used to delete a file that no longer exists
in the working tree but remains in git's index. `git ls-files` will list
files that are in the HEAD, which is exactly what we want. Local files
not in `HEAD` can't be deleted from git anyway.
Line continuations (i.e. escaped new lines) now make sense again. With
the smart pipe support (pipes continue on to next line) recently added,
this hack to have continuations ignore comments makes no sense.
This is valid code:
```fish
echo hello |
# comment here
tr -d 'l'
```
this isn't:
```fish
echo hello | \
# comment here
tr -d 'l'
```
Reverts @snnw's 318daaffb2Closes#2928. Closes#2929.
The tool subcommand had a "-f" flag to disallow file completions which is wrong: most of the tools there require a file/directory argument.
Since we're here, also limit "go tool compile" to only match Go source files.
From the discussion in #3802, handling spaces within braces more
gracefully. Leading and trailing whitespace that isn't quoted or escaped
is stripped, whitespace in the middle is preserved. Any whitespace
encountered within expansion tokens is treated as a single space,
similar to how programming languages that don't hard break tokens/quotes
on line endings would.
cmake can (and should) be used to invoke the build/install command,
instead of directly calling `ninja` or `make`, via the `--build DIR
[--target TARGET]` syntax.
This will use the native BSD bmake build system instead of the previous
hack which spawned an instance of `gmake` (GNU Make) if installed to
perform the build.
This addresses the discussion regarding the dependency on `hostname` and
the addition of a `$hostname` variable to replace it.
`$hostname` is a read-only, GLOBAL_ENV, non-electrified, lowercased,
non-exported variable that is read once at the start of a fish session.
The finer points of this can be debated endlessly, but this is a shared
starting point that any changes can build on (ref #4422).
Regarding performance: @krader1961 brought up some good points in #4422
regarding potential DNS timeouts (but they really don't apply except if
the host name is not hardcoded in resolv.conf, which quickly manifests
with a cascade of errors on most *nix systems in all cases), but note
that gethostname() was already being called by fish so that would be
more of a future optimization than a "must" at this point.
The value is not electrified or tied and is read-only. It isn't cached
in the get_hostname_identifier() function as the ENV_GLOBAL $hostname
will cache it for its duration.
The behavior of `gethostname` in case of an insufficient buffer is
library and version dependent. Work around this by using a big enough
buffer then truncating the output to our desired max length.
The 0th index of the array was tested inside the loop instead of just
once outside it.
Also explain `input_mapping_is_match` control code behavior and
reasoning and simplify control flow.
Removed misleading statement about read requiring an argument, as the
note about read's new behavior when no arguments are provided covers
that and is less confusing.
In similar vein to how fish_default_key_bindings works, parameters
passed to the function are automatically passed to bind upstream.
Additionally, -s is automatically added if no parameters had been
specified to prevent startup error messages. See 46d1334.
Closes#4494
Drops the % notation for process expansion. The existing notation was a
mess and expanded jobs, process ids, and process names via dark magic.
With this change, % is no longer a special character and can be used
unescaped with impunity.
The variables %self and %last, referring to fish's own pid and the pid
of the last backgrounded job respectively, have been replaced with $pid
and $last_pid. These are read-only variables, protected against being
redefined by the user.
Author's note: I would have personally preferred $fish_pid instead of
$pid but since we debated changing $version to $fish_version and then
reverted that change (with much acrimony), it makes no sense to break
with that precedent here. Additionally, $fish_last_pid is quite wordy.
Closes#4230. Closes#1202.
When number is infinite, not a number, larger than LONG_MAX or smaller
than LONG_MIN, print a corresponding error and return STATUS_CMD_ERROR.
This should fix the worst of the problems, by at least making them clear.
Fixes#4479.
Fixes#4768.
This allows prompts to react to $COLUMNS by e.g. omitting some parts.
We still fallback to a ">" prompt if that's still not short enough,
but now the user has a way of making a nicer prompt.
Fixes#904.
Fixes#4381.
If the head is not a valid, existent command, do not load and run custom
completion sources. This applies to both the autosuggestion provider and
manual user completions. File-based completions will still be offered.
Supersedes #4782 and #4783. Closes#4783. Closes#4782. Closes#2365.
This promotes "and" and "or" from a type of statement to "job
decorators," as a possible prefix on a job. The point is to rationalize
how they interact with && and ||.
In the new world 'and' and 'or' apply to a entire job conjunction, i.e.
they have "lower precedence." Example:
if [ $age -ge 0 ] && [ $age -le 18 ]
or [ $age -ge 75 ] && [ $age -le 100 ]
echo "Child or senior"
end
This fixes a variety of issues related to building the documentation
with CMake. In particular it cleans up the dependency management and
fixes some issues where the documentation build was using generated
files from the source directory.
Now parses package.json and uses results to provide a list of possible
completions to `yarn remove`. There may be other subcommands that could
benefit from this.
Could have parsed yarn output, but yarn is slow and packages.json format
is generally standard since it's machine-generated json.
Can be used to retrieve a list of parent paths, useful for searching
ancestors recursively via their absolute paths. Paths are returned from
deepest to shallowest, starting from the path passed in. Paths are not
validated for performance reasons. (Usually the input to
__fish_parent_directories would be (pwd) or (dir $file).)
This should speed things up on slower PCs given that the vast majority
of shell commands are simple jobs consisting of a single command without
any pipelines, in which case there's no need for a keepalive process at
all. Applies to WSL only.
As a temporary workaround for the behavior described in
Microsoft/WSL#2997 wherein WSL does not correctly assign the spawned
child its own PID as its PGID, explicitly set the PGID for the newly
spawned process.
fish's cmake install routines were attempting to create system
directories that already existed, an operation for which the permissions
to do so may not be available (e.g. /usr/local/share/pkgconfig)
This commit first checks if a directory exists before creating it.
This replaces muparser with tinyexpr, which
- Saves about 4000 lines
- Removes functionality we do not use, making it so we can document _all_ of it
- Should be more palatable to distributions
This now reports "TOO_MANY_ARGS" instead of no error (and triggering
an assertion).
We might want to add a new error type or report the missing operator
before, but this is okay for now.
This turns a bunch of ifs on their heads.
We often see this pattern in te:
```c
if (s->type != SOME_TYPE) {
// error handling
} else {
// normal code
}
```
Only, since we want to return the first error, we do
```c
if (s->type == SOME_TYPE) {
// normal code
} else if (s->type != TOK_ERROR) {
// Add a new error - if it already has type error
// this should already be handled.
}
```
One big issue is the comma operator, that means arity-1 functions can
take an arbitrary number of arguments. E.g.
```fish
math "sin(5,9)"
```
will return the value of sin for _9_, since this is read as "5 COMMA
9".
We no longer use muparser, but tinyexpr.
tinyexpr does not have:
- "Statistical functions" like min, max, avg
- Multiple expressions separated with ","
[ci skip]
This enables some limited use of arguments for wrapping completions. The
simplest example is that complete gco -w 'git checkout' now works like
you would want: `gco <tab>` now invokes git's completions with the
`checkout` argument prepended.
Fixes#1976
Previously, in
ls ^a bcd
(with "^" as the cursor), kill-word would delete the "a" and then go
on, remove the space and the "bcd".
With this, it will only kill the "a".
Fixes#4747.
Homebrew and other systems set the path for the extra completion,
function and configuration directories outside the writeable prefix.
Mirror the autotools build in trying to create these directories, but
not causing the whole install to fail if this operation in unsuccessful.
This is part of an effort to improve fish's Unicode handling. This commit
attempts to grapple with the fact that, certain characters (principally
emoji) were considered to have a wcwidth of 1 in Unicode 8, but a width of
2 in Unicode 9.
The system wcwidth() here cannot be trusted; terminal emulators do not
respect it. iTerm2 even allows this to be set in preferences.
This commit introduces a new function is_width_2_in_Uni9_but_1_in_Uni8() to
detect characters of version-ambiguous width. For these characters, it
returns a width guessed based on the value of TERM_PROGRAM and
TERM_VERSION, defaulting to 1. This value can be overridden by setting the
value of a new variable fish_emoji_width (presumably either to 1 or 2).
Fixes#4539, #2652.
`argparse`, `read`, `set`, `status`, `test` and `[` now can't be used
as function names anymore.
This is because (except for `test` and `[`) there is no way to wrap these properly, so any
function called that will be broken anyway.
For `test` (and `[`), there is nothing that can be added and there
have been confused users who created a function that then broke
everything.
Fixes#3000.
Prior to this fix, each redirection type was a separate token_type.
Unify these under a single type TOK_REDIRECT and break the redirection
type out into a new sub-type redirection_type_t.
The custom command for fish.pc had a dependency on FBVF, but it appears
that the relative path to FBVF was incorrect and with CMake 3.10.1 under
FreeBSD this was consistently causing the build to fail if
../build_tools/git_version_gen.sh hadn't (coincidentally, I think?)
already run.
Explicitly set the dependency path for FBVF to the binary directory.
The custom command for fish.pc had a dependency on FBVF, but there was
no cmake rule for the generation of the FBVF file. With CMake 3.10.1
under FreeBSD, this was consistently causing the build to fail if
../build_tools/git_version_gen.sh hadn't (coincidentally, I think?)
already run.
Prior to this the tokenizer ran "one ahead", where tokenizer_t::next()
would in fact return the last-parsed token. Switch to parsing on demand
instead of running one ahead; this is simpler and prepares for tokenizer
changes.
Turns out the process-exit is only ever used in conjunction with
`%self`. Make that explicit by just adding a new "fish_exit" event,
and deprecate the general process-exit machinery.
Fixes#4700.
The previous attempt to support newlines after pipes changed the lexer to
swallow newlines after encountering a pipe. This has two problems that are
difficult to fix:
1. comments cannot be placed after the pipe
2. fish_indent won't know about the newlines, so it will erase them
Address these problems by removing the lexer behavior, and replacing it
with a new parser symbol "optional_newlines" allowing the newlines to be
reflected directly in the fish grammar.
Prior to this fix, if you attempt to complete from inside a quote and the
completion contained an entity that cannot be represented inside quotes
(i.e. \n \r \t \b), the result would be a broken mess of quotes. Rewrite
the implementation so that it exits the quotes, emits the correct unquoted
escape, and then re-enters the quotes.
Properly escape literal tildes in tab completion results. Currently we
always escape tildes in unquoted arguments; in the future we may escape
only leading tildes.
Fixes#2274
Prior to this fix, autoloads like function and completion autoloads
would check their path variable (like fish_function_path) on every
autoload request. Switch to invalidating it in response to the variable
changing.
This improves time on a microbenchmark:
for i in (seq 50000)
setenv test_env val$i
end
from ~11 seconds to ~6.5 seconds.
The job control functions were a bit messy, in particular
`set_child_group`'s name would imply that all it does is set the child
group, but in reality it used to set the child group (via `setpgid`),
set the job's pgrp if it hasn't been set, and possibly assign control of
the terminal to the newly-created job.
These have been split into separate functions. Now `set_child_group`
does just (and only) that, `maybe_assign_terminal` might assign the
terminal to the new pgrp, and `on_process_created` is used to set the
job properties the first time an external process is created. This might
also speed things up (but probably not noticeably) as there are no more
repeated calls to `getpgrp()` if JOB_CONTROL is not set.
Additionally, this closes#4715 by no longer unconditionally calling
`setpgid` on all new processes, including those created by `posix_spawn`
which does not need this since the child's pgrep is set at in the
arguments to that API call.
This merges a set of changes that switch functions from executing source
to executing an already parsed tree (the same tree used when the function
is defined). This speeds up function execution, reduces memory usage, and
avoids annoying double parsing.
A simple microbenchmark of function execution:
for i in (seq 10000)
setenv test_env val$i
end
time improves from 1.63 to 1.32 seconds.
This switches function execution from the function's source code to
its stored node and pstree. This means we no longer have to re-parse
the function every time we execute it.
The idea is that we can return the shared pointer directly, avoiding
lots of annoying little getter functions that each need to take locks.
It also helps to pull together the data structures used to initialize
functions versus store them.
This concerns block nodes with redirections, like
begin ... end | grep ...
Prior to this fix, we passed in a pointer to the node. Switch to passing
in the tnode and parsed source ref. This improves type safety and better
aligns with the function-node plans.
Prior to this fix, functions stored a string representation of their
contents. Switch them to storing a parsed source reference and the
tnode of the contents. This is part of an effort to avoid reparsing
a function's contents every time it executes.
Add a fish-specific wrapper around std::mutex that records whether it is
locked in a bool. This is to make ASSERT_IS_LOCKED() simpler (it can just
check the boolean instead of relying on try_lock) which will make Coverity
Scan happier.
Some details: Coverity Scan was complaining about an apparent double-unlock
because it's unaware of the semantics of try_lock(). Specifically fish
asserts that a lock is locked by asserting that try_lock fails; if it
succeeds fish prints an error and then unlocks the lock (so as not to leave
it locked). This unlock is of course correct, but it confused Coverity Scan.
Use wcstring/string instead of a character array. The variable
`term_env` was not being freed before the function exited.
Fixes defect 7520324 in coverity scan.
This merges a sequence of commits that undoes the SIGCONT orchestration
used for WSL compatibility. The essential problem is this: In Unix and
Linux, exited processes are still valid until it is reaped; you can, say,
make an exited process a group leader. But in Windows, when a process
exits, it is gone, and most syscalls (other than, say, waitpid) fail for
it. This is known as the WSL Rick Grimes problem.
This manifests as various race conditions in WSL between a parent operating
on a child, and the child exiting. Prior to this merge, these were
addressed by having the child wait for the parent to send it a SIGCONT.
This resolved the race.
This merge removes this approach and replaces it with a simpler mechanism
that leverages the existing keepalive machinery. A keepalive process is
created for all platforms when we have a pipeline that contains a builtin.
This is necessary to keep the whole process group alive. The fix is, on
WSL, we always create a keepalive and make it the group leader. Because the
keepalive does not call exec and its lifetime is bound to a C++ stack
frame, it is easy to resolve the race.
This improves performance a bit (except on WSL), since child processes no
longer have to synchronize with the parent process, but the big win is
simplicity. This removes the notion of the single global stopped child, of
which there could only be one, and which had be resumed at the right
time(s), of which there were several.
keepalive processes are typically killed by the main shell process.
However if the main shell exits the keepalive may linger. In WSL
keepalives are used more often, and the lingering keepalives are both
leaks and prevent the tests from finishing.
Have keepalives poll for their parent process ID and exit when it
changes, so they can clean themselves up. The polling frequency can be
low.
Have WSL use a keepalive whenever the first process is external.
This works around the fact that WSL prohibits setting an exited
process as the group leader.
* 🚀
* prepare to merge into fish-shell
* split into different files
* remove deprecated option
* captitalize descriptions
* make shorter description for ansible
* update ansible-playbook (and ansible for consistency)
* update version on vault and galaxy
When the pager wants to use the full screen to show many options, it reserves
space at the top to see the command. Previously it pretended the command was a
prompt and engaged the prompt layout mechanism to compute these lines. Instead
let's juts count newlines since escape sequences within commands are very rare.
There were several issues with the way that the include tests for curses.h
were being done that were ultimately causing fish to use the headers from
ncurses but link against curses on platforms that provide an actual
libcurses.so that isn't just a symlink to libncurses.so
In particular, the old code was first testing for curses's cureses.h and then
falling back to libncurses's implementation of the same - but that logic was
reversed when it came to including term.h, in which case it was testing for
the ncurses term.h and falling back to the curses.h header. Long story short,
while cmake will link against libcurses.so if both libcurses.so and
libncurses.so are present (unless CURSES_NEED_NCURSES evaluates to TRUE, but
that makes ncurses a hard requirement), but we were brining in some of the
defines from the ncurses headers, causing SIGSEGV panics when fish ultimately
tried to access variables that weren't exported or were mapped to undefined
areas of memory in the other library.
Additionally it is an error to include termios.h prior to including the plain
Jane curses.h (not ncurses/curses.h), causing errors about unimplemented types
SGTTY/chtype. So far as I can tell, both curses.h and ncurses/curses.h pull in
termios.h themselves so it shouldn't even be necessary to manually include it,
but I have just moved its #include below that of curses.h
The non-ncurses version of term.h requires that curses.h be first
included. Only very recent versions of CMake include a LANGUAGE
option to CHECK_INCLUDE_FILES, so we aren't using it and specifying
CXX here..
This never worked properly (since a branch that only exists locally
would also be offered) and is dog-slow.
When we come up with a better way to do it we can readd it.
When git prints a path like "share/completions/git.fish", that's
relative to the root of the repo. So we need to either remove
everything from the $PWD (if the path is inside the $PWD), or prepend
a ":/", which is git-speak for "relative to the root".
This was removed by mistake in the recent switch to `git status`.
Fixes#4688.
This special cases expansion of $history variables, so that slicing
history no longer needs to construct the entire history array. Speedup
is around 100x in my test.
Fixes#4650
Prior to this fix, if the user typed normal characters while the
completion pager was shown, it would begin searching. This feature was
not well liked, so we are going to instead just append the characters as
normal and disable paging. Control-S can be used to toggle the search
field.
Fixes#2249
CheckFunctionExists checks for C linkage only, and recommends the use of
CheckSymbolExists in the documentation. This improves the detection of
C++ features, as opposed to C features.
try_get_child() was taking the address of a reference; clang was thereby
assuming it could not be null and so was dropping the null check. Ensure
we do not dereference a null pointer.
Fixes#4678
* git completions: Parse git status --porcelain
This is much faster on large repositories, as it allows us to do a lot
more with a single git call.
It also makes it easy to add descriptions to distinguish modified
files from untracked ones.
TBD is if all commands now have the right kinds of files.
[ci skip]
`-v` is a non-standard GNU-only extension to `awk`, its usage in the
generation of the fish.pc script breaks on non-GNU platforms (such as
FreeBSD and presumably macOS).
Using `sed` with only standard posix commands instead.
This was a symbol that represented either an argument or a redirection.
This was only used as part of argument_or_redirection_list.
It's simpler to just have these types be alternatives in the list type.
This merges a set of changes to improve the type safety of the fish parse
tree, in preparation for modifying fish grammar in 3.0. It expresses the
fish grammar via a new file parse_grammar.h. It then adds a new type
tnode_t parametrized on grammar elements, with typesafe access to its
children.
The idea here is to make it easy to change the fish grammar, and have the
compiler report code locations that must to be updated.
Merge branch 'threeparse'
Some of these were failing on Travis quite often, and this is probably
the result of too tight a window.
E.g. one emacs test (transpose words, default timeout, short delay)
waited 250ms to enter something else, with a timeout of 300ms. That
meant a window of 50ms.
* [PO][FR]fix translation
"key" was being translated to "fonction". ("function")
Based the new wording on the above translation.
* [PO][FR]fix translation
"directory" was being translated to "fonction". ("function")
* [PO][FR]fix translation
"Permission denied" was being translated to "Nom de fonction illégal". ("Illegal function name")
I took the new translation from strerror.
* [PO][FR]fix translation
"Introduction" was being translated to "Instruction illégale". ("illegal instruction")
* $ make po/fr.po
* #4655: changes requested by @PenegalECI
* fix some automatically generated translations
uniformly translate "logging" to "journalisation".
This reverts commit 36a2f2cc01.
This attempted to modify RPATH when building with Ninja, but the CMake if
statement wasn't actually valid so this wasn't doing anything. This check
couldn't really be tested - let's make sure not to accumulate build system
rules that we don't understand.
Some dotfile users like to add directories to PATH that point at
non-existent directories (because those directories exist on other
machines). Stop warning in that case, unless those directories contain
a colon, in which case it's probably a user error.
Changed cd completion to differentiate between cd autosuggest and cd tab
completion. cd autosuggest will find deepest unique hierarchy and cd tab
completion will not.
Issue #4402
This untangles the CMake versioning issues (I hope) as discussed in #4626.
Note most of the advice found on the Internet about how to inject git
versions into CMake is just wrong.
The behavior we want is to unconditionally run the script
build_tools/git_version_gen.sh at build time (i.e. when you invoke ninja or
make, and not when you invoke cmake, which is build system generation time).
This script is careful to only update the FISH-BUILD-VERSION-FILE if the
contents have changed, to avoid spurious rebuilding dependencies of
FISH-BUILD-VERSION-FILE. Assuming the git version hasn't changed, the script
will run, but not update FISH-BUILD-VERSION-FILE, and therefore
fish_version.o will not have to be rebuilt.
This might normally rebuild more than is necessary even if the timestamp is
not updated, because ninja computes the dependency chain ahead of time. But
Ninja also supports the 'restat' option for just this case, and CMake is rad
and exposes this via BYPRODUCTS. So mark FISH-BUILD-VERSION-FILE as a
byproduct and make the script always update a dummy file
(fish-build-version-witness.txt). Note this is the use case for which
BYPRODUCTS is designed.
We also have fish_version.cpp #include "FISH-BUILD-VERSION-FILE", and do a
semi-silly thing and make FISH-BUILD-VERSION-FILE valid C++ (so there's just
one version file). This means we have to filter out the quotes in other
cases..
This reverts commit 25839b8c36.
This was an attempt to simplify the version generation, but it
computed the version at build sytem generation time rather than
at build time, requiring another run of CMake to update it.
Correctly generate FISH_BUILD_VERSION for use in fish_version.h/cpp and
fish.pc to allow `fish --version` and `echo $version` to work again.
Not needing the same convoluted measures used by Makefile builds to
prevent the regeneration of the fish version file when it hasn't
changed.
Purposely created a new `cmake_git_version_gen.sh` file so that the old
`git_version_gen.sh` remains compatible with the existing Makefile build
script. Same reason why `fish.pc.in` was not modified to use a lowercase
variable name to match the CMAKE variable of the same name.
Closes#4626
If `doxygen` isn't installed, the man files aren't built and that's
quite ok. The cmake `install` target was presuming the man files would
always be present and the install stage was failing if they weren't
built.
CMake originally links build artifacts/results so that they can run from
the target directory. As a result, it must first relink the binaries
before installation so that they can run from the installation target
directory, typically done in the preinstall stage. Ninja does not have a
preinstall stage, and the CMake code that generates the build.ninja file
does not take that into account [0].
Setting `CMAKE_BUILD_WITH_INSTALL_RPATH` [1] makes it originally link
the files with the RPATH settings for the final destination directory,
meaning that relinking is no longer needed.
Technically setting the RPATH is not required for the `fish` binary as
we do not have any relative dependencies; this is the output of
`ldd ./build/fish`:
```
linux-vdso.so.1 => (0x00007ffffacdc000)
libncurses.so.5 => /lib/x86_64-linux-gnu/libncurses.so.5
(0x00007f6632350000)
libtinfo.so.5 => /lib/x86_64-linux-gnu/libtinfo.so.5
(0x00007f6632120000)
libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2
(0x00007f6631f00000)
libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6
(0x00007f6631b70000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6
(0x00007f6631860000)
libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1
(0x00007f6631630000)
libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0
(0x00007f6631410000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6
(0x00007f6631040000)
/lib64/ld-linux-x86-64.so.2 (0x00007f6632600000)
```
However, since the bug only exists when the build generator is set to
ninja, the workaround is only activated for that specific build
generator to prevent any future problems.
[0]: https://cmake.org/Bug/print_bug_page.php?bug_id=13934
[1]: https://cmake.org/cmake/help/v3.0/variable/CMAKE_BUILD_WITH_INSTALL_RPATH.html
`git push REMOTE :BRANCH` deletes remote branch BRANCH from remote
REMOTE. Should only kick in when the pattern matches, hopefully didn't
break anything else!
A large portion of time was spent constructing strings and passing
them to debug(). Turn debug into a macro so that the strings are only
constructed if they're going to be printed.
The psub tests create a fifo and launch a background job to write to it.
However fifos have this obnoxious behavior where opening the file blocks
until both sides are ready. In one of the tests we don't actually read
from the fifo we create, so the background job hangs, and the tests
never complete. Fix this by just reading from the fifo.
This adds a new class arg_iterator_t which encapsulates decisions about
whether to read arguments from stdin or argv. It also migrates the
unread bytes buffer from a static variable to an instance variable.
* Add eopkg support
Add support for eopkg in __fish_print_packages function, and
add new completion eopkg.fish in share/completions
* Sorry for the empty file
* Sorry for the empty file again
* Use builtin function for checking subcommand and options
* Fix description
* Use string function to replace grep and cut
* Add completion for search command
This was caused by it prepending "-s" to argv always,
and later checking $argv[1].
As it turns out, that is kinda superfluous, so we can just add "-s" to
the `bind` calls.
Also adjust the tests so the vi-bindings are enabled via the function,
which would have caught this.
Fixes#4494.
Profiling with callgrind revealed that about 60% of the time in a `something | string match` call
was actually spent in `string_get_arg_stdin()`,
because it was calling `read` one byte at a time.
This makes it read in chunks similar to builtin read.
This increases performance for `getent hosts | string match -v '0.0.0.0*'` from about 300ms to about 30ms (i.e. 90%).
At that point it's _actually_ quicker than `grep`.
To improve performance even more, we'd have to cut down on str2wcstring.
Fixes#4604.
Muparser Exceptectomy
This removes large pieces of muParser that fish does not use, such as its optimizer. It also switches muParser from throwing exceptions to propagating errors explicitly.
This is a very strange design that determines whether initialization
needs to be performed by reassigning a function pointer. A misguided
optimization? Just check explicitly.
Instead of throwing an exception, simply return false. It is too
complicated to thread the error return through this function and
ParserInt is unused by fish anyways.
The optimizer adds a fair amount of complexity in muparser with no
benefit to fish, since fish is not going to use complicated expressions
or cache parsed expressions.
To help remove exception handling, we will need to have a type that
has visibility into both ParserError and value_type. We're going to
put this type in muParserDef.h. Remove the error header and fold its
contents into muParserDef.h.
This was a silly data structure that didn't carry its weight.
Replace it with a wrapper around std::vector that doesn't explicitly
throw exceptions. It's unclear if muparser relied on the exception
throwing behavior of ParserStack, and it seems there's no way to find
out except removing it and seeing what breaks.
The tests pass for what that's worth!
Turns out "__fish_git_staged_files" does the same thing as "__fish_git_modified_files --staged".
Also use "--staged" instead of "--cached", which is a more
understandable synonym.
Many thanks to @thomcc on gitter.
Prior to this fix, a "bare variable" in math like 'x + 1' would be
looked up in the environment, i.e. equivalent to '$x + 1'. This appears
to have been done for performance. However this breaks the orthogonality
of fish; performance is not a sufficient justification to give math this
level of built-in power, especially because the performance of math is
not a bottleneck. The implementation is also ugly.
Remove this feature so that variables must be prefixed with the dollar
sign and undergo normal variable expansion. Reading 'git grep' output
does not show any uses of this in fish functions or completions.
Also added to changelog.
Fixes#4393
Before this change, if a command failed, this was indicated by the "$"
at the end of the prompt turning red.
With this change in place, if a command fails, the exit code of the
failing command is displayed in [square brackets].
Running "cut" multiple times in a loop has an adverse performance
impact on first use, especially on slow systems. Using builtin "read"
for the same purpose is faster and cleaner.
Command name continues twice in man page.
Current version's example:
NAME
andand - conditionally execute a command
Fixed version:
NAME
and - conditionally execute a command
for various completions.
This makes the code a bit nicer, removes one of the
__fish_print_hostnames calls (which are slow) and a sed call, thereby
improving performance by about 33% (600ms to 400ms).
Fixes#4511.
The previous hack used to work around an OS X issue/bug where launching
a URL with a #fragment appended would drop the fragment by using
`osascript` does not seem to work any more. Append the section name as a
query string (in addition to, not instead of #section) and then use some
basic javascript appended to the user doc HTML template to parse that
and jump to the correct section (if the section was dropped).
Closes#4480
fish was indiscriminately calling `rustc -Z help` in the autocompletion
script for `rustc`, but `-Z` (and its `-Z help` output completions) is
only available when using the nightly compiler.
Note that this isn't a perfect fix since if you try to use those command
line options now added to the autocompletions list without using the
nightly toolchain, `rustc` will still throw an error. But at least this
way we don't cause random errors about `-Z` not being available to
appear any time someone tries to use `rustc` from the fish command prompt.
This adds a new library fishlib, which the CMake build builds.
This library is linked by the tests, fish, and fish_indent, so
that object files do not have to be built separately for each
of them.
This adds support for creating the FISH-BUILD-VERSION-FILE in the CMake
build. A FISH-BUILD-VERSION-FILE is created in the CMake directory
and only updated when necessary.
As part of factoring out the documentation building parts of the fish
build, add a new file build_user_doc.sh that builds the user_doc directory.
Invoke it from both the Makefile and CMake build.
As part of factoring out the documentation building parts of the fish
build, add a new file build_index_hdr.sh that builds the index.hdr
file. Invoke it from both the Makefile and CMake build.
As part of factoring out the documentation building parts of the fish
build, add a new file build_index_hdr.sh that builds the index.hdr
file. Invoke it from both the Makefile and CMake build.
As part of factoring out the documentation building parts of the fish
build, add a new file build_commands_hdr.sh that builds the commands.hdr
file. Invoke it from both the Makefile and CMake build.
This adds a new script build_tools/build_lexicon_filter.sh
that builds the lexicon filter. It is factored out from the Makefile,
and both the Makefile and CMake build invoke it.
E.g. if "foo" is in CDPATH, and both "foo/bar" and "./bar" exist, `cd
bar` will go to ./bar.
The completions described "bar" as going to "foo/bar" ("CDPATH foo").
This fixes it by checking for ./bar's existence.
See #4475.
macOS 10.5 and earlier do not support the convention of returning
a dynamically allocated string, plus this seems like an unnecessary
malloc. Always allocate a buffer for realpath() to write into.
(cherry picked from commit 05c0cb713d)
These messages are automatically generated as if `-w` were specified
at the gmake command line. The `--no-print-directory` option supresses
these messages.
(cherry picked from commit b7f1103088)
For some reason on Solaris the previous code was refusing to compile
with an error (regarding the declaration of stdout in the opts struct)
error: declaration of ‘__iob’ as array of references
The obvious guess that it had something to do with the name of the
variable in question proved true; renaming it from `stdout` to
`opts.stdout` allows the build to go through.
These messages are automatically generated as if `-w` were specified
at the gmake command line. The `--no-print-directory` option supresses
these messages.
macOS 10.5 and earlier do not support the convention of returning
a dynamically allocated string, plus this seems like an unnecessary
malloc. Always allocate a buffer for realpath() to write into.
It seems that `parse_cmd_opts` does not correctly handle no arguments,
and so argc was being decremented to -1 causing uninitialized memory
access when argv[0] was dereferenced at a later point.
"Use the fish_update_completions command.", the answer to "How do I update man page completions?", was also found at the end of the answer to "How do I get the exit status of a command?"
(cherry picked from commit 48797974d3)
No longer using `-` to indicate reading to stdout. Use lack of arguments
as stdout indicator. This prevents mixing of variables with stdout
reading and makes it clear that stdout may not be mixed with delimiters
or array mode.
Added an option to read to stdout via `read -`. While it may seem
useless at first blush, it lets you do things like include
mysql -p(read --silent) ...
Without needing to save to a local variable and then echo it back.
Kicks in when `-` is provided as the variable name to read to. This is
in keeping with the de facto syntax for reading/writing from/to
stdin/stdout instead of a file in, e.g., tar, cat, and other standard
unix utilities.
"Use the fish_update_completions command.", the answer to "How do I update man page completions?", was also found at the end of the answer to "How do I get the exit status of a command?"
\b does not match "end of spaces" but rather "start of a-z/0-9" and so
does not match the start of string '-c'. Match (and then re-insert) a
literal ' ' as part of the pattern instead.
(cherry picked from commit b61c4f1cbc)
\b does not match "end of spaces" but rather "start of a-z/0-9" and so
does not match the start of string '-c'. Match (and then re-insert) a
literal ' ' as part of the pattern instead.
Work around bug pypa/pip#4755
Don't expect all users to be running a version of pip2/3 that includes
the fix (once it's upstreamed). Will continue to work if/when pip2/3
emit the correct output. pip is already very slow at printing the
completions (see #4448) so the `sed` call overhead is neglible.
Work around bug pypa/pip#4755
Don't expect all users to be running a version of pip2/3 that includes
the fix (once it's upstreamed). Will continue to work if/when pip2/3
emit the correct output. pip is already very slow at printing the
completions (see #4448) so the `sed` call overhead is neglible.
* Add pip completion
* We call native pip completion for fish if pip is installed
* Add pipenv completion
* We call pipenv native completion if pipenv is installed
* Applied changes as requested by @floam
* Changed usage of `test (command -v)` for just `command -sq`
* Add completions for pip2/3
* In some systems pip is not aliased and we have pip2 and pip3
* In those cases, we just load the completions for those commands
* Separate pip2/3 completions in their own file as requested by @floam
* Add pip completion
* We call native pip completion for fish if pip is installed
* Add pipenv completion
* We call pipenv native completion if pipenv is installed
* Applied changes as requested by @floam
* Changed usage of `test (command -v)` for just `command -sq`
* Add completions for pip2/3
* In some systems pip is not aliased and we have pip2 and pip3
* In those cases, we just load the completions for those commands
* Separate pip2/3 completions in their own file as requested by @floam
This silences binding errors due to keys not found in the current
termcap config in the default fish bindings.
Closes#4188, #4431, and obviates the original fix for #1155
It was necessary to re-implement builtin_bind as a class in order to
avoid passing around the options array from function to function and
as adding an opts parameter to `get_terminfo_sequence` would require
otps to be passed to all other builtin_bind_ functions so they could, in
turn, pass it to `get_terminfo_sequence`.
On powerpc64 (big-endian platform) one test failed as:
Testing file printf.in ... fail
Output differs for file printf.in. Diff follows:
--- printf.tmp.out 2017-10-02 18:14:17.740000000 -0700
+++ printf.out 2017-10-02 18:11:59.370000000 -0700
@@ -1,5 +1,5 @@
Hello 1 2 3.000000 4.000000 5 6
-a B 0 18446744073709551615
+a B 10 18446744073709551615
It happens due to roughly the following code:
swprintf(..., L"%o", (long long)8);
Here mismatch happens between "%o" (requires 32-bit value)
and 'long long' (requires 64-bit value).
The fix turns it effectively to:
swprintf(..., L"%llo", (long long)8);
as it was previously done for 'x', 'd' and other int-like types.
Makes tests pass on powerpc64.
Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>
This
- Offers more candidates
- Is more reactive (it'll always incorporate "--state=" and "--type="
- Is faster (about 800ms to about 120ms)
- Needs fewer function files
All __fish_systemctl_* functions except __fish_systemctl_services have
been removed.
Squashed commit of the following:
commit fb252e6e10
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Tue Sep 26 15:52:23 2017 -0500
CHANGELOG.md: kdeconnect-cli, not kdecomplete
commit e031d91c19
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Tue Sep 26 15:49:59 2017 -0500
fixup! Updated changelog with info about all new and updated completions since 2.6.0
commit 6366a67c21
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Tue Sep 26 15:36:37 2017 -0500
fixup! Updated changelog with info about all new and updated completions since 2.6.0
commit 281be31eb3
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Tue Sep 26 15:21:01 2017 -0500
Updated changelog with info about all new and updated completions since 2.6.0
Decided to move doc_src/fish.lss to share/lynx.lss, which just makes
more sense all around. Accordingly, now using {$__fish_datadir} instead
of {$__fish_help_dir} in help.fish.
Makefile now installs the custom lss on make install
(cherry picked from commit 338311af1e)
Lynx uses a very naïve method of applying styles to HTML elements by
hashing the element type and the class name to generate a map of
hash:style. After the hash is calculated, Lynx does not go back and
check whether or not the actual string values match the LSS properties.
See the following links on the Links mailing list:
* http://lists.gnu.org/archive/html/lynx-dev/2015-12/msg00037.html
* http://lists.gnu.org/archive/html/lynx-dev/2015-12/msg00039.html
This patch copies the default Lynx stylesheet but removes highlighting
and other styles that would result in unreadable text (due to not enough
contrast with the background color), and if the `help` builtin detects
that the best web browser to use is Lynx, it instructs it to use this
modified stylesheet.
(cherry picked from commit 8b858f2fcc)
If the $DISPLAY environment variable is not set, xdg-open should not be
used to load the web browser. Just because it is installed does not mean
that the user exclusively runs in an X session.
Needed for Lynx detection to work around #4170
(cherry picked from commit 2b425ad221)
Decided to move doc_src/fish.lss to share/lynx.lss, which just makes
more sense all around. Accordingly, now using {$__fish_datadir} instead
of {$__fish_help_dir} in help.fish.
Makefile now installs the custom lss on make install
Lynx uses a very naïve method of applying styles to HTML elements by
hashing the element type and the class name to generate a map of
hash:style. After the hash is calculated, Lynx does not go back and
check whether or not the actual string values match the LSS properties.
See the following links on the Links mailing list:
* http://lists.gnu.org/archive/html/lynx-dev/2015-12/msg00037.html
* http://lists.gnu.org/archive/html/lynx-dev/2015-12/msg00039.html
This patch copies the default Lynx stylesheet but removes highlighting
and other styles that would result in unreadable text (due to not enough
contrast with the background color), and if the `help` builtin detects
that the best web browser to use is Lynx, it instructs it to use this
modified stylesheet.
If the $DISPLAY environment variable is not set, xdg-open should not be
used to load the web browser. Just because it is installed does not mean
that the user exclusively runs in an X session.
Needed for Lynx detection to work around #4170
The POSIX standard specifies that a buffer should be supplied to
getcwd(), not doing so is undefined (or rather, platform-defined)
behavior. This was causing the getcwd errors on illumos (though not seen
on Solaris 11) reported in #3340Closes#3340
(cherry picked from commit b495c68f28)
The POSIX standard specifies that a buffer should be supplied to
getcwd(), not doing so is undefined (or rather, platform-defined)
behavior. This was causing the getcwd errors on illumos (though not seen
on Solaris 11) reported in #3340Closes#3340
Thanks to @ThomasAH, as per #4378. Tested on many platforms (OS X,
FreeBSD, Linux, and Solaris). Works with IPv4 and IPv6 as well as
host names and loopback addresses.
(cherry picked from commit 3b3bcc998e)
Thanks to @ThomasAH, as per #4378. Tested on many platforms (OS X,
FreeBSD, Linux, and Solaris). Works with IPv4 and IPv6 as well as
host names and loopback addresses.
Took care of remaining issues preventing fish from building on Solaris.
Mainly caused by some assumptions that certain defines are POSIX when
they are not (`NAME_MAX`).
Moved `NAME_MAX` defines to common.h - for some reason, it was being
defined in a cpp file (`env_universal_common.cpp`) even though it is used
in multiple source files.
Now compiles on Solaris 11 with GNU Make. Still some warnings because
fish was written with GNU getopt in mind and the Solaris version doesn't
use `const char *` but rather just `char *` for getopt values, but it
builds nevertheless.
Assuming this closes#3340
(cherry picked from commit ffebe74885)
Took care of remaining issues preventing fish from building on Solaris.
Mainly caused by some assumptions that certain defines are POSIX when
they are not (`NAME_MAX`).
Moved `NAME_MAX` defines to common.h - for some reason, it was being
defined in a cpp file (`env_universal_common.cpp`) even though it is used
in multiple source files.
Now compiles on Solaris 11 with GNU Make. Still some warnings because
fish was written with GNU getopt in mind and the Solaris version doesn't
use `const char *` but rather just `char *` for getopt values, but it
builds nevertheless.
Assuming this closes#3340
We had pid_status defined as a pid_t instance, which was fine since on
most platforms pid_t is an alias for int. However, that is not
universally the case and waitpid takes an int *, not a pid_t *.
After cc35241a6e, BSD users can just call
make normally and have it redirect the build/install/test/whatever to
GNU Make.
(cherry picked from commit 3604522bf2)
Smarter BSDmakefile that automatically calls gmake to build the targets,
even including `-j` if provided. README.md can be simplified to remove
`gmake` references from build instructions for BSD users.
(cherry picked from commit cc35241a6e)
Smarter BSDmakefile that automatically calls gmake to build the targets,
even including `-j` if provided. README.md can be simplified to remove
`gmake` references from build instructions for BSD users.
A completion may have zero length; in this case the length of the
prefix was omitted and the completion was not visible. Correct the
calculation to account for zero-width completions.
Fixes#4424
A completion may have zero length; in this case the length of the
prefix was omitted and the completion was not visible. Correct the
calculation to account for zero-width completions.
Fixes#4424
Use \uXXXX consistently for unicode code points
(cherry picked from commit 6b2e84be0e)
Backporting to 2.7.0 branch just to try and keep changes between master
and this branch as minimal as possible.
Drawing prompt in repo with text=auto attribute and mixed line endings in files was spawning crlf conversion warnings to terminal from unsilenced stderr of git diff
A recent discussion involving whether `can_be_encoded()` was broken
caused me to notice that we are inconsistent about whether Unicode code
points are specified using `\xXXXX` or `\uXXXX` notation. Which is
harmless but silly and potentially confusing.
j does not have any "logical" source of completions, but it almost often
called with arguments that have been seen before (since it is used to
jump to favorite/recent directories). We can search the history for
possible completions and use those.
This is an example of the behavior mentioned in #4344 as a possible
enhancement for fish 3.0, where completions can be provided from history
if none are otherwise found.
This flag was only documented for a few weeks before being renamed
`--show-time` and has been deprecated for a long time. Fish 3.0 is a good
opportunity to remove it.
Instead of treating the search term as a literal string to be matched
treat it as a glob. This allows the user to get a more useful set of
results by using the `*` glob character in the search term.
Partial fix for #3136
* Implement `history search --reverse`
It should be possible to have `history search` output ordered oldest to
newest like nearly every other shell including bash, ksh, zsh, and csh.
We can't make this the default because too many people expect the
current behavior. This simply makes it possible for people to define
their own abbreviations or functions that provide behavior they are
likely used to if they are transitioning to fish from another shell.
This also fixes a bug in the `history` function with respect to how it
handles the `-n` / `--max` flag.
Fixes#4354
* Fix comment for format_history_record()
On BSD platforms, a BSD-specific BSDmakefile is searched for and used
before any generic Makefile. We can use this to emit an informational
message directing the user to use GNU Make instead of relying on the
user's recognizing of random build failures on syntax errors as a sign
to switch to GNU Make.
(Random fact: this same trick also applies to GNU Make, which searches
for a GNUmakefile before using Makefile)
On BSD platforms, a BSD-specific BSDmakefile is searched for and used
before any generic Makefile. We can use this to emit an informational
message directing the user to use GNU Make instead of relying on the
user's recognizing of random build failures on syntax errors as a sign
to switch to GNU Make.
(Random fact: this same trick also applies to GNU Make, which searches
for a GNUmakefile before using Makefile)
commit e07f1d59c06094846db8ce59f65d4790b222fffa
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sun Sep 10 21:54:45 2017 -0500
Use git branch and git branch --remote for checkout completions
commit 9e1632236be065e051e306b11082ca4e9c7a0ee1
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sun Sep 10 11:27:30 2017 -0500
Correct classification of remote and local branches
To prevent any breakage, no changes were made to __fish_git_branches,
instead its output was filtered into __fish_git_remote_branches and
__fish_git_local_branches, the two of which are now used to provide
completions for "git checkout ..."
Fixes#4395Closes#4396
It seems that under python3, s3cmd emits its output as a long list (like
ls -l) with or without the --long parameter to "s3cmd ls s3://...".
This patch includes only s3://* paths from that output as completions.
commit e07f1d59c06094846db8ce59f65d4790b222fffa
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sun Sep 10 21:54:45 2017 -0500
Use git branch and git branch --remote for checkout completions
commit 9e1632236be065e051e306b11082ca4e9c7a0ee1
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sun Sep 10 11:27:30 2017 -0500
Correct classification of remote and local branches
To prevent any breakage, no changes were made to __fish_git_branches,
instead its output was filtered into __fish_git_remote_branches and
__fish_git_local_branches, the two of which are now used to provide
completions for "git checkout ..."
Fixes#4395Closes#4396
It seems that under python3, s3cmd emits its output as a long list (like
ls -l) with or without the --long parameter to "s3cmd ls s3://...".
This patch includes only s3://* paths from that output as completions.
Addresses the main concern of #3830 by preserving the internal ordering
of tag/branch listings generated by git. Fixes mixing of remote and
local branches in completions.
Does not address the concern of having local branches on top, remote
branches after, and tags at the bottom - I don't believe we have that
functionality available to us yet. #361 only implemented sort within a
category of completions, but there is no category "weight" unless I'm
mistaken.
As discussed in #3805, this patch disables assigning fish to its own
process group at startup. This was trialled in #4349 alongside other
pgrp fixes which introduced additional problems, but this particular fix
seems to be OK.
Fixes#3805 and works around Microsoft/BashOnWindows#1653
* Hoist `for` loop control var to enclosing scope
It should be possible to reference the last value assigned to a `for`
loop control var when the loop terminates. This makes it easier to detect
if we broke out of the loop among other things. This change makes fish
`for` loops behave like most other shells.
Fixes#1935
* Remove redundant line
This now includes hosts with custom ports (and other hosts with the
same key), and explicitly excludes negated hosts, those with a
wildcard and those with an `@`-marker (e.g. `@revoked`)
It's also possibly a bit quicker because the ordering is better,
especially for files with many comments.
* Add repo completion for zypper
* Replace sed with string in __fish_print_zypp_repos
* Move function into completion script
* Update zypper completion
add subcommand packages to __fish_zypper_repo_commands
(cherry picked from commit 81becc5f6b)
* Add repo completion for zypper
* Replace sed with string in __fish_print_zypp_repos
* Move function into completion script
* Update zypper completion
add subcommand packages to __fish_zypper_repo_commands
The type cached_esc_sequences_t caches escape sequences, and is tasked
with finding an escape sequence that prefixes a given string. Before
this fix, it did so by storing the lengths of cached escape sequences,
and searching for substrings of that length. The new implementation
instead stores all cached escape sequences in a sorted vector, and uses
binary search to find the shortest escape sequence that is a prefix of
the input. This is a substantial simplification that also reduces
allocations.
56d9134534 contained an LRU cache plus
changes to the documentation; 95162ef19d
reverted both.
This commit re-adds the documentation changes, which are still correct.
This eliminates the "missing" notion of env_var_t. Instead
env_get returns a maybe_t<env_var_t>, which forces callers to
handle the possibility that the variable is missing.
maybe_t is an implementation of the Maybe/Optional type, allowing
for an optional value to be stored. This will enable a more
principled approach for functions that return values or failure,
such as env_get.
This commit backs out certain optimizations around setting environment
variables, and replaces them with move semantics. env_set accepts a
list, by value, permitting callers to use std::move to transfer
ownership.
Commit f872f25f introduced a freed memory access regression on line 460
of env.cpp, where an environment variable was converted to a temporary
string, the .c_str() address of which was stored while the string
temporary was destroyed.
This commit keeps a reference to the original string lying around so
that the c_str() pointer does not point to freed memory.
Valgrind warns that the sometimes uninitialized sigaction.sa_flags field
is sometimes used when passed to the signal handler.
This patch explicitly zeros out the sigaction.sa_flags field at creation
time.
Valgrind warns that the sometimes uninitialized sigaction.sa_flags field
is sometimes used when passed to the signal handler.
This patch explicitly zeros out the sigaction.sa_flags field at creation
time.
cherry-picked from krader1961/fish-shell commit b69df4fe72
Fixes#4353 (regression in indexing of history contents) and introduces
new unit tests to catch bad $history indexing in the future.
Using bare vars is more efficient because it makes the builtin `math`
expression cache more useful. That's because if you prefix each var with
a dollar-sign then the fish parser expands it before `math` is run.
Something like `math x + 1` can be cached since the expression is the
same each time it is run. But if you do `math $x + 1` and x==1 then you're
effectively executing `math 1 + 1`. And if x==2 the next time then you're
running `math 2 + 1`. Which makes the expression cache much less effective.
This implements an LRU cache of recently seen math expressions. When
executing math inside loops and the like this can provide a 33% decrease
in the time to execute the `math` command.
Remove our `math` function that wraps `bc`. Our math builtin is now good
enough that it can be the default implementation.
Another step in resolving #3157.
We need our `math` builtin to behave like `bc` with respect to rounding
floating point values to integer to avoid breaking to many existing
uses. So when scale is zero round down to the nearest integer.
Another change for #3157.
The MuParser supports the concept of multiple expressions separated by
commas. This implements support for that so that you can do things like
this:
set results (math '1+1, 4*2, 9^2')
This is the second baby step in resolving #3157. Implement a bare minimum
builtin `math` command. This is solely to ensure that fish can be built
and run in the Travis build environments. This is okay since anyone running
`builtin math` today is already getting an error response.
Also, more work is needed to support bare var references, multiple result
values, etc.
First step in fixing issue #3157 is to check-in the source code and hook
it into our build system.
The inclusion of the MuParser source adds the MIT License to those that
apply to fish. Update our documentation to reflect that fact.
The MuParser documentation is at
http://beltoforion.de/article.php?a=muparser. The source was downloaded
from https://github.com/beltoforion/muparser/releases. It is also hosted
on Github, https://github.com/beltoforion/muparser/. I did not download
it from Github because that source contained just a couple of cleanup
changes which don't affect its behavior.
Using a read-only variable like `status` as a for loop control variable
has never worked. But without this change you simply get non-sensical
behavior that leaves you scratching your head in puzzlement. This change
replaces the non-sensical behavior with an explicit error message.
Fixes#4342
Recent changes to switch to unordered sets/maps can cause the order in
which items are returned to be non-deterministic. This change ensures
that the argparse "Mutually exclusive flags" error message to be
deterministic with respect to the order of the interpolated values.
Make setting fish vars more efficient by avoiding creating a
wcstring_list_t for the case where we're setting one value. For the case
where we're passing a list of values swap it with the list in the var
rather than copying it. This makes the benchmark in #4200 approximately
6% faster.
Since we are including XXHash32/64 anyway for the wchar_t* hashing,
we might as well use it.
Use arch-specific hash size and xxhash for all wcstring hashing
Instead of using XXHash64 for all platforms, use the 32-bit version
when running on 32-bit platforms where XXHash64 is significantly slower
than XXHash32 (and the additional precision will not be used).
Additionally, manually specify wcstring_hash as hashing method for
non-const wcstring unordered_set/map instances (the const varieties
don't have an in-library hash and so already use our xxhash-based
specialization when calling std::hash<const wcstring>).
commit 50f414a45d58fcab664ff662dd27befcfa0fdd95
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sat Aug 19 13:43:35 2017 -0500
Converted file_id_t set to unordered_set with custom hash
commit 83ef2dd7cc1bc3e4fdf0b2d3546d6811326cc3c9
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sat Aug 19 13:43:14 2017 -0500
Converted remaining set<wcstring> to unordered_set<wcstring>
commit 053da88f933f27505b3cf4810402e2a2be070203
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sat Aug 19 13:29:21 2017 -0500
Switched function sets to unordered_set
commit d469742a14ac99599022a9258cda8255178826b5
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sat Aug 19 13:21:32 2017 -0500
Converted list of modified variables to an unordered set
commit 5c06f866beeafb23878b1a932c7cd2558412c283
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sat Aug 19 13:15:20 2017 -0500
Convert const_string_set_t to std::unordered_set
As it is a readonly-list of raw character pointer strings (not
wcstring), this necessitated the addition of a hashing function since
the C++ standard library does not come with a char pointer hash
function.
To that end, a zlib-licensed [0] port of the excellent, lightweight
XXHash family of 32- and 64-bit hashing algorithms in the form of a C++
header-only include library has been included. XXHash32/64 is pretty
much universally the fastest hashing library for general purpose
applications, and has been thoroughly vetted and is used in countless
open source projects. The single-header version of this library makes it
a lot simpler to include in the fish project, and the license
compatibility with fish' GPLv2 and the zero-lib nature should make it an
easy decision.
std::unordered_set brings a massive speedup as compared to the default
std::set, and the further use of the fast XXHash library to provide the
string hashing should make all forms of string lookups in fish
significantly faster (to a user-noticeable extent).
0: http://create.stephan-brumme.com/about.html
commit 30d7710be8f0c23a4d42f7e713fcb7850f99036e
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sat Aug 19 12:29:39 2017 -0500
Using std::unordered_set for completions backing store
While the completions shown to the user are sorted, their storage in
memory does not need to be since they are re-sorted before they are
shown in completions.cpp.
commit 695e83331d7a60ba188e57f6ea0d9b6da54860c6
Author: Mahmoud Al-Qudsi <mqudsi@neosmart.net>
Date: Sat Aug 19 12:06:53 2017 -0500
Updated is_loading to use unordered_set
Per the discussion with @faho in #4332, replaced some custom completion
state detection functions with standard __fish_* functions used in other
completion sources.
(cherry picked from commit f706081ea4)
No longer auto-generated. Everything has been summarized. Supressing
file completions for initial command, providing list of valid initial
commands, filtering --options by subcommand.
(cherry picked from commit 539acd9fc5)
No longer using RAII wrappers around pthread_mutex_t and pthread_cond_t
in favor of the C++11 std::mutex, std::recursive_mutex, and
std::condition_variable data types.
The `react_to_variable_change()` function is called whenever a fish var
is set. Even as a consequence of statements like `for x in a b c`. It is
therefore critical that that function be as fast as possible. Especially
when setting the var doesn't have any side-effects which is true something
like 99.9999% of the time.
This change reduces the overhead of `react_to_variable_change()` to
unmeasurable levels. Making the synthetic benchmark in issue #4341
36% faster.
Fixes#4341
Internally fish should store vars as a vector of elements. The current
flat string representation is a holdover from when the code was written
in C.
Fixes#4200
Per the discussion with @faho in #4332, replaced some custom completion
state detection functions with standard __fish_* functions used in other
completion sources.
No longer auto-generated. Everything has been summarized. Supressing
file completions for initial command, providing list of valid initial
commands, filtering --options by subcommand.
This patch adds completion for the update subcommand, that is, when the
user types in `composer update <tab>`.
The code depends on python for the json parsing. I'm not sure if this
is appropriate or if there is a fish-native way to parse json data.
Use suggestions for remove subcommand.
Add suggestions for why, why-not and depends.
Add why/why-not suggestion.
This patch adds completion for the update subcommand, that is, when the
user types in `composer update <tab>`.
The code depends on python for the json parsing. I'm not sure if this
is appropriate or if there is a fish-native way to parse json data.
Use suggestions for remove subcommand.
Add suggestions for why, why-not and depends.
Add why/why-not suggestion.
A semi-empty var is one with a single empty string element. The
`env_var_t::empty()` method returns true for such vars but we want
`set --show` to report that it has a single empty element.
A semi-empty var is one with a single empty string element. The
`env_var_t::empty()` method returns true for such vars but we want
`set --show` to report that it has a single empty element.
This reverts functional changes in commit
ea3e9698df.
* Annotated tags only should be used for releases - see #3572 for
examples of where we want to use lightweight tags.
See also git-tag(1) on the purpose of annotated and lightweight tags.
* Version numbers are numbers and should not start with a branch name.
The commit ID is embedded in the version and uniquely identifies the
history. `fish --version` and `echo $FISH_VERSION` contain this
information.
(cherry picked from commit dcb39bfa86)
This reverts functional changes in commit
3bef4a3c1f.
* Annotated tags only should be used for releases - see #3572 for
examples of where we want to use lightweight tags.
See also git-tag(1) on the purpose of annotated and lightweight tags.
* Version numbers are numbers and should not start with a branch name.
The commit ID is embedded in the version and uniquely identifies the
history. `fish --version` and `echo $FISH_VERSION` contain this
information.
Now that we're working on the 3.0.0 major release it is more important
than ever that fish binaries built by developers have version strings
which clearly communicate where they came from.
Now that we're working on the 3.0.0 major release it is more important
than ever that fish binaries built by developers have version strings
which clearly communicate where they came from.
An optional feature that suggests you install Python is okay;
core-dumping is not.
The note on tests was about fish development tests, not the `test`
builtin for conditional syntax.
Specifically mention git, hg, and svn in the VCS section.
getopt doesn't work very well in the BSDs, and getent has plenty of
fallbacks to replace it when it's not available.
<https://github.com/terrycloth/fish-shell/commit/
47a768ceeaef1d702624802d83338edbcc0f377c#commitcomment-23613921>
Newer versions of GCC and Clang are not satisfied by a cast to void,
this fix is adapted from glibc's solution.
New wrapper function ignore_result should be used when a function with
explicit _unused_attribute_ wrapper is called whose result will not be
handled.
This reverts commit ee15f1b987.
The test relies on undefined behaviour (checking for errno in the
absence of an error condition) and was broken on OpenBSD.
Closes#4184.
Make the `env_var_t::missing_var()` object a singleton rather than a
dynamically constructed object. This requires some discipline in its use
since C++ doesn't directly support immutable objects. But it is slightly
more efficient and helps identify code that incorrectly mutates `env_var_t`
objects that should not be modified.
It's bugged me forever that the scope is the second arg to `env_get()`
but not `env_set()`. And since I'll be introducing some helper functions
that wrap `env_set()` now is a good time to change the order of its
arguments.
Specifically closes#4313.
Not being as agressive in what we ignore/blacklist, but can be revisited
easily in the future to add more characters to the argument blacklist.
By far the most common problem with universal variables being overridden
by global variables is other values being imported from the environment;
the `set -q; or set -gx` is much more of an edge case.
My previous change to eliminate `class var_entry_t` caused me to notice
that `env_get()` turned a set but empty var into a missing var. Which
is wrong. Fixing that brought to light several other pieces of code that
were wrong as a consequence of the aforementioned bug.
Another step to fixing issue #4200.
* Fix clearing abandoned line with VTE
With VTE-based terminals, resizing currently causes multi-line prompts
to go weird.
This changes the sequence we use to clear the line to one suggested by
a VTE
developer (https://bugzilla.gnome.org/show_bug.cgi?id=763390#c4).
It changes nothing in konsole 17.04.3 and urxvt 9.22, but they already
work.
Note that this does not fix the case where output did not end in a
newline, but that doesn't seem to be up to us. Also, it only affects
those lines.
Fixes#2320.
* Use terminfo definition instead of hardcoding
Thanks to @ixjlyons.
getopt doesn't work very well in the BSDs, and getent has plenty of
fallbacks to replace it when it's not available.
<https://github.com/terrycloth/fish-shell/commit/
47a768ceeaef1d702624802d83338edbcc0f377c#commitcomment-23613921>
Newer versions of GCC and Clang are not satisfied by a cast to void,
this fix is adapted from glibc's solution.
New wrapper function ignore_result should be used when a function with
explicit _unused_attribute_ wrapper is called whose result will not be
handled.
This reverts commit ee15f1b987.
The test relies on undefined behaviour (checking for errno in the
absence of an error condition) and was broken on OpenBSD.
Closes#4184.
getopt doesn't work very well in the BSDs, and getent has plenty of
fallbacks to replace it when it's not available.
<https://github.com/terrycloth/fish-shell/commit/
47a768ceeaef1d702624802d83338edbcc0f377c#commitcomment-23613921>
Newer versions of GCC and Clang are not satisfied by a cast to void,
this fix is adapted from glibc's solution.
New wrapper function ignore_result should be used when a function with
explicit _unused_attribute_ wrapper is called whose result will not be
handled.
This reverts commit ee15f1b987.
The test relies on undefined behaviour (checking for errno in the
absence of an error condition) and was broken on OpenBSD.
Closes#4184.
Make the `env_var_t::missing_var()` object a singleton rather than a
dynamically constructed object. This requires some discipline in its use
since C++ doesn't directly support immutable objects. But it is slightly
more efficient and helps identify code that incorrectly mutates `env_var_t`
objects that should not be modified.
It's bugged me forever that the scope is the second arg to `env_get()`
but not `env_set()`. And since I'll be introducing some helper functions
that wrap `env_set()` now is a good time to change the order of its
arguments.
Specifically closes#4313.
Not being as agressive in what we ignore/blacklist, but can be revisited
easily in the future to add more characters to the argument blacklist.
Specifically closes#4313.
Not being as agressive in what we ignore/blacklist, but can be revisited
easily in the future to add more characters to the argument blacklist.
* Clarify dependencies: required vs optional, and build vs runtime.
A first pass at updating the dependency documentation, based on the
discussion in this thread:
https://github.com/fish-shell/fish-shell/issues/2062
* Clarify notes on dependency errors, tests, and VCS integration.
An optional feature that suggests you install Python is okay;
core-dumping is not.
The note on tests was about fish development tests, not the `test`
builtin for conditional syntax.
Specifically mention git, hg, and svn in the VCS section.
An optional feature that suggests you install Python is okay;
core-dumping is not.
The note on tests was about fish development tests, not the `test`
builtin for conditional syntax.
Specifically mention git, hg, and svn in the VCS section.
My previous change to eliminate `class var_entry_t` caused me to notice
that `env_get()` turned a set but empty var into a missing var. Which
is wrong. Fixing that brought to light several other pieces of code that
were wrong as a consequence of the aforementioned bug.
Another step to fixing issue #4200.
By far the most common problem with universal variables being overridden
by global variables is other values being imported from the environment;
the `set -q; or set -gx` is much more of an edge case.
* Fix clearing abandoned line with VTE
With VTE-based terminals, resizing currently causes multi-line prompts
to go weird.
This changes the sequence we use to clear the line to one suggested by
a VTE
developer (https://bugzilla.gnome.org/show_bug.cgi?id=763390#c4).
It changes nothing in konsole 17.04.3 and urxvt 9.22, but they already
work.
Note that this does not fix the case where output did not end in a
newline, but that doesn't seem to be up to us. Also, it only affects
those lines.
Fixes#2320.
* Use terminfo definition instead of hardcoding
Thanks to @ixjlyons.
These completions never actually worked and always fell back to the
builtin path completion. But a recent fix means that these now keep the
fallback from happening resulting in no completions for these commands.
Doing `set -U var` when a global named `var` exists can result in
confusing behavior. Try to limit the confusion by improving the warning
we write. Also, only write the warning if interactive.
Fixes#4267
The process_t pointer sent to setup_child_process can actually be 0
without it being failure, as that is what fish sends when `exec` is run
(in the case of INTERNAL_EXEC).
This was causing exec to fail.
There is no more race condition between parent and child with
regards to setting the process groups. Each child sets it for themselves
and then blocks indefinitely until the parent does what it needs to for
them (having waited for them to set their process groups). They are not
SIGCONT'd until the next process in the chain (if any) starts so that
that process can join their process group and open the pipes.
In the last commit, we introduced an indiscriminate if !EXTERNAL check
that unblocks a previously SIGSTOP'd command (if any) to allow the main
loop in exec_job to read from it without deadlocking (since builtins and
functions read directly from input as an optimization, sometimes).
Now only unblocking where a fork will not happen to ensure that if a
builtin ends up forking, that fork'd process is guaranteed to be able to
join the previous process' process group and access its output pipes.
Setting the process group in a fork/exec scenario is a well-documented
race condition in pretty much any job control mechanism [0] [1]. The
Wikipedia article contradicts the glibc article and suggests that the
best approach is for the parent to wait for the child to become the
process group leader, while the glibc article suggests that both should
make it so (which is what fish did previously). However, I'm running
into cases where tcsetpgrp is causing an EPERM error, which it isn't
documented to do except if the session id for the calling process
differs from that of the target process group (which is never the case
in fish since they are all part of the same session), which should cause
a _different_ error (SIGTTOU to be sent to all members of the calling
process' group).
In all cases, this is easily remedied by checking if the process group
in question is already in control of the terimnal. There's still the
off-chance that in the time between we check that and the time that the
command completes that situation may have changed, but the parent
process is supposed to ignore the result of this call if it errors out.
[0]: https://en.wikipedia.org/wiki/Process_group
[1]: https://www.gnu.org/software/libc/manual/html_node/Launching-Jobs.html
We were having child processes SIGSTOP themselves immediately after
setting their process group and before launching their intended targets,
but they were not necessarily stopped by the time the next command was
being executed (so the opposite of the original race condition where
they might have finished executing by the time the next command came
around), and as a result when we sent them SIGCONT, that could never
reach. Now using waitpid to synchronize the SIGSTOP/SIGCONT between the
two.
If we had a good, unnamed inter-process event/semaphore, we could use
that to have a child process conditionally stop itself if the next
command in the job chain hadn't yet been started / setup, but this is
probably a lot more straightforward and less-confusing, which isn't a
bad thing.
Additionally, there was a bug caused by the fact that the main exec_job
loop actually blocks to read from previous commands in the job if the
current command is a built-in that doesn't need to fork.
With this waitpid code, I was able to finally add the SIGSTOP code to
all the fork'd processes in the main exec_job loop without introducing
deadlocks; it turns out that they should be treated just like the main
EXTERNAL fork, but they tend to execute faster causing the same deadlock
described above to occur more readily.
The only thing I'm not sure about is whether we should execute
unblock_pid undconditionally for all !EXTERNAL commands. It makes more
sense to *only* do that if a blocking read were about to be done in the
main loop, otherwise the original race condition could still appear
(though it is probably mitigated by whatever duration the SIGSTOP lasted
for, even if it is SIGCONT'd before the next command tries to join the
process group).
I hadn't realized that the for loop is called multiple times for a given
"single input" (anything that doesn't include semicolons, etc) to fish,
and so processes were being blocked but blocked_pid was lost by the time
that the next job (which was reading from the last process in the
previous job) came around.
Now using a static variable to store the last blocked PID. AFAICT, this
main job control loop is always executed from the same process and
thread, so this shouldn't need to be wrapped in atomics/mutexes, etc.
This code should be more portable, and certainly cleaner. We are
currently always sending SIGCONT to the last process (if it was part of
a job chain) regardless of whether it called SIGSTOP on itself or not,
which should be fine.
Need to explore whether or not the other forks in src/exec.cpp need to
be SIGSTOP'd on run or only the one that we included in this patch.
I'm not sure if this happens on all platforms, but under WSL with the
existing codebase, processes in the job chain that pipe their
stdout/stderr to the next process in the job could terminate before the
next job started (on fast enough machines for quick enough jobs).
This caused issues like #4235 and possibly #3952, at least for external
commands. What was happening is that the first process was finishing
before the second process was fully set up. fish would then try to
assign (in both the child and the parent) the process group id belonging
to the process group leader to the new process; and if the first process
had already terminated, it would have ended its process group with it as
well before that happened.
I'm not sure if there was already a mechanism in place for ensuring that
a process remains running at least as long as it takes for the next
process in the chain to join its group, etc., but if that code was
there, it wasn't working in my test setup (WSL).
This patch definitely needs some review; I'm not sure how I should
handle non-external commands (and external commands executed via
posix_spawn). I don't know if they are affected by the race condition in
the first place, but when I tried to add the same "wait for next command
in chain to run before unblocking" that would cause black screens
requiring ctrl+c to bypass.
The "unblock previous command" code was originally run by the next child
to be forked, but was then moved to the shell code instead, making it
more-centrally located and less error-prone.
Note that additional headers may be required for the mmap system call on
other platforms.
This is the first step to implementing issue #4200 is to stop subclassing
env_var_t from wcstring. Not too surprisingly doing this identified
several places that were incorrectly treating env_var_t and wcstring as
interchangeable types. I'm not talking about those places that passed
an env_var_t instance to a function that takes a wcstring. I'm talking
about doing things like assigning the former to the latter type, relying
on the implicit conversion, and thus losing information.
We also rename `env_get_string()` to `env_get()` for symmetry with
`env_set()` and to make it clear the function does not return a string.
I decided this was just too useful not to include in our final fish 2.x
release. And since it does not modify any existing behavior it is safe
to include at this late date in the process of creating 2.7.
This makes command substitutions impose the same limit on the amount
of data they accept as the `read` builtin. It does not limit output of
external commands or builtins in other contexts.
Fixes#3822
This adds a new capability to the `set` command. It is similar to
running `set` with no other arguments but provides far more detail about
each variable. Such as whether it is set in each of the local, global,
and universal scopes. And the values in each scope. You can also ask for
specific variables to be shown.
Fixes#4265
Rewrite the `abbr` function to store each abbreviation in a separate
variable. This greatly improves the efficiency. For the common case
it is 5x faster. For pathological cases it is upwards of 100x faster.
Most people should be able to unconditionally define abbreviations in
their config.fish without a noticable slow down.
Fixes#4048
This takes a string that is then split upon like `string split`.
Unlike $IFS, the string is used as one piece, not a set of characters.
There is still a fallback to IFS if no delimiter is given, that
behaves exactly as before.
Fixes#4156.
This silences warnings from the compiler about ignoring return value of
‘ssize_t write(int, const void*, size_t)’, declared with attribute
warn_unused_result [-Wunused-result].
The class `completer_t` declares `complete_special_cd`, an unused method. I searched the entire source tree and this declaration seems to be the only instance of `complete_special_cd`. There is no definition or uses which likely means this is dead code.
PR #3691 made most calls to `signal_block()` and `signal_unblock()`
no-ops unless a magic env var is set when fish starts running. It's
been seven months since that change was made and no problems have been
reported. This finishes that work by removing those no-op function calls
and support for the magic env var in our next major release (which won't
happen till at least six months from now).
The primary motivation for --keep-order for `complete` was to support
something like commit history completions, which are returned by git in
reverse chronological order and make no sense alphabetically (they are
SHA1 hashes).
See https://github.com/fish-shell/fish-shell/issues/361 for more info.
Introduce a -k/--keep-order switch to `complete` that can be used to
prevent fish from sorting/re-ordering the results provided by a completion
source.
In addition, this patch does so without doing away with deduplication
of completions by introducing a new unique_unsorted(..) helper function
that removes duplicates in-place without affecting the general order of
the vector/container.
Note that the code now uses a stable sort for completions, since the
behavior of is_naturally_less_than as of this patch now means that the
results are not necessarily _actually_ identical just because that function
repeatedly returns false for any ordering of any given two elements.
Fixes#361
* Make npm run-script completion faster with `jq`
When jq is available, it's actually faster to invoke jq and parse the `package.json`
invoking the `npm` command.
Also, prior to this commit, both `__fish_complete_npm` and `__fish_npm_run` were being run
whenever completions for `npm run` subcommand was being used, which was actually making
repetitive work (invoking npm command twice). This pull request is supposed to make completion
without `jq` faster as well
* Refactor npm.fish for code reutilization
Created function to handle both cases of npm run completion parse, with or without `jq` completion.
* Remove unecessary blank line
This completes the refactoring of the `set` builtin. It also removes a
seemingly never used feature of the `set` command. It also eliminates all
the lint warnings about this module.
Fixes#4236
When reporting whether a boolean flag was seen report the actual flags
rather than a summary count. For example, if you have option spec `h/help`
and we parse `-h --help -h` don't do the equivalent of `set _flag_h 3`
do `set _flag_h -h --help -h`.
Partial fix for #4226
When executing a function, local-exported (`set -lx`) variables
previously were not accessible at all. This is weird e.g. in case of
aliases, since
```fish
set -lx PAGER cat
git something # which will call $PAGER
```
would not work if `git` were a function, even if that ends up calling
`command git`.
Now, we copy these variables, so functions get a local-exported copy.
```fish
function x
echo $var
set var wurst
echo $var
end
set -lx var banana
x # prints "banana" and "wurst"
echo $var # prints "banana"
```
One weirdness here is that, if a variable is both local and global,
the local-copy takes precedence:
```fish
set -gx var banana
set -lx var pineapple
echo $var # prints "pineapple"
x # from above, prints "pineapple" and "wurst"
echo $var # still prints "pineapple"
set -el var # deletes local version
echo $var # "banana" again
```
I don't think there is any more consistent way to handle this - the
local version is the one that is accessed first, so it should also be
written to first.
Global-exported variables are _not_ copied, instead they still offer
full read-write access.
When reporting whether a boolean flag was seen report the actual flags
rather than a summary count. For example, if you have option spec `h/help`
and we parse `-h --help -h` don't do the equivalent of `set _flag_h 3`
do `set _flag_h -h --help -h`.
Partial fix for #4226
In the rare case that we don't inherit $HOME _and_ can't read it from
/etc/passwd, this makes it so instead of triggering an assert() $HOME
is set to the empty list.
Tilde-expansion expands to nothing in such a case (and a string-empty
$HOME), `cd` errors out.
Fixes#4229.
Fish 2.6.0 introduced a regression that keeps setting
`fish_escape_delay_ms` as a uvar from working. This also fixes a related
problem: callbacks generated from the initial loading of universal vars
were not being acted on.
Fixes#4196
Also stop special-casing `printf` as if it were a syntactical keyword
with respect to handling `printf --help`. It should use the same pattern
as every other builtin command.
The code for reporting parser errors needs a major overhaul. But rather
than do that I'm going to add another hack in the hope that this doesn't
introduce yet another problem.
Fixes#4221
The recent change to switch `psub` to use `argparse` caused it to use
a fifo by default because it inadvertently fixed a long standing bug in
the fish script. This changes the behavior back to `psub --file` being
the default behavior and introduces a `--fifo` flag. It also updates the
documentation to make it clearer when and why `--fifo` mode should not
be used.
Fixes#4222
While updating the `history` function to use `argparse` I realized it is
useful to define an option that can be used in three ways. First by
using the short flag; e.g., `-n NNN`. Second by using the long flag;
e.g., `--max NNN`. Third, as an implicit int flag; e.g., `-NNN`. This
use case is now supported by a spec of the form `n#max`.
A recent regression to the `alias` command points out the need for more
unit tests of its behavior. I also decided to use it as an opportunity
to normalize the output of just `alias` to list aliases.
The previous change to use `argparse` for parity with every other
builtin and function introduced a regression. Invocations that start
with a negative number can fail because the negative value looks like an
invalid flag.
This implements support for numeric flags without an associated short or
long flag name. This pattern is used by many commands. For example `head
-3 /a/file` to emit the first three lines of the file.
Fixes#4214
This implements a `fish_opt` command that provides a way for people
to create option specs for the `argparse` command as an alternative to
creating such strings by hand.
Fixes#4190
We've needed a fishy way to parse flags and arguments given to scripts
and functions for a very long time. In particular a manner that provides
the same behavior implemented by builtin commands. The long term goal is
to support DocOpt. But since it is unclear when that will happen so this
implements a `argparse` command. So named as homage to the excellent
Python module of the same name.
Fixes#4190
This fixes a stupid bug in my previous commit to standardize on a new
`list_to_array_val()` function. This adds a unit test to keep this from
regressing.
This is the first step in implementing a better abstraction for handling
fish script vars in the C++ code. It implements a new function (with two
signatures) to provide a standard method for construct the flag string
representation of a fish script array.
Partial fix for #4200
The count command should not treat any flag specially. Not even `-h` and
`--help`. It should simply return a count of the number of arguments it
received.
Fixes#4189
Completion strings, especially the description, might contain characters,
such as backspace, which make it impossible to calculate the width of
the string.
Fixes#4179
Var `___git_ps_color_suffix_done` is supposed to be
`___fish_git_prompt_color_suffix_done`. This bug was found by an
experimental change to detect the use of undefined variables (#4163).
Similarly, we should simply test whether `__fish_git_prompt_showcolorhints`
is set rather than set to a non-empty string.
The `read` command `-m` and `--mode-name` vars are now deprecated and do
nothing other than result in a warning message. The `read` command now
honors the `FISH_HISTORY` var that is used to control where commands are
read from and written to. You can set that var to the empty string to
suppress the use of both history files. Or you can set it to a history
session ID in which case that will limit the `read` history that is
available.
Fixes#1504
Don't import the bash history if the user has specified that a non-default
fish history file should be used. Also, rename the var that specifies
the fish history session ID from `FISH_HISTFILE` to `FISH_HISTORY`.
Fixes#4172
Using the FISH_HISTFILE variable will let people customise the session
to use for the history file. The resulting history file is:
`$XDG_DATA_HOME/fish/name_history`
Where `name` is the name of the session. The default value is `fish`
which results in the current history file.
If it's set to an empty string, the history will not be stored to a
file.
Fixes#102
Because the 'getopt' library differs between systems, it's likely
that there will be different output. This is the case between the
GNU-based Linux and the BSD-based Darwin, for the 'getopt' library,
it seems. It causes the tests to produce different results.
To allow us to test, and check for regressions, on the different
platforms, the invocation code has been updated to allow a
system-specific suffix to be used on the test files. If this suffix
is found, the test will also be flagged as being system-specific
which should ensure the change in behaviour is noted.
The Travis macOS test systems do not appear to have colordiff present, so any
failures would mean that no output would be shown. This may also be a
problem for the other test scripts as well, but the invocation tests are
the ones being affected here.
We change our behaviour to downgrade to the plain diff tool if colordiff is
not present.
The invocation tests were not especially clear on how they should be
used, without reading the code. And who really wants to do that? So,
a description of what the test does (and thus what each file is) is
now present in the file prologue comment.
Some more of the invocations are tested in this change:
- bad switches
- errors in configuration files
- regular command, configuration and init command ordering
- persistence of variables over command invocation.
- interactive and login switch use
- terminal exit code return
- version request
There are sure to be other invocations that should be tested, but
these give a fair number of them a go.
The new '-C' initial command needs some tests, and as there are no
tests just yet for the command invocation, this change adds a harness
and calls it from the high-level tests in the Makefile.
The tests are similar in style to the other high level tests, in that
we capture the output and compare it to that which we expect. The
harness itself is written in bash - sorry - because we're testing the
fish shell's invocation, and trying to do that with the fish we've
just built wouldn't actually make for a very useful test when things
go wrong.
The 'tests/invocation.sh' script can be executed manually, or as part
of the 'make test' target, to make it easy to use both as part of the
development and as part of automation.
The harness has only been tested on linux with bash 4.3.11, and requires
grep and sed. Although not tested with OS X, I believe I have avoided
the syntax which is inconsistent.
The tests added here cover just the initial command's basic execution,
and when it is mixed with the regular '-c' command.
In order to allow the execution of commands before dropping to an
interactive prompt, a new switch, '-C' or '--init-command' has been
added to those switches that we accept.
The documentation has been updated correspondingly.
The original code only supported a single command list to be executed,
and this command list terminates the shell when it completes. To allow
the new command list to preceed the original one, both have been
wrapped in a new container class 'command_line_switches_t'. This is
then passed around in place of the list of strings we used previously.
I had considered moving the interactive, login and other command line
switch states into this container, but doing so would change far more
of the code, moving the structure to be available globally, and I
wasn't confident of the impact. However, this might be a useful thing
to do in the future.
A new function, run_command_list, was lifted from the prior execution
code, and re-used for both the initial command and the regular command
execution.
We now have a builtin that can do URL escaping so use it. I can't find
any uses of our private `__fish_urlencode` function in any Oh-My-Fish or
Fisherman code so remove it.
We need a way to encode arbitrary strings into valid fish variable
names. It would also be nice if we could convert strings to valid URLs
without using the slow and hard to understand `__fish_urlencode` function.
In particular, eliminating the need to manipulate the locale.
Fixes#4150
As part of addressing #1310 I decided it makes more sense to replace
`current-function` with just `function`, etc., because I'm going to add
flags to let the user specify which stack level they are interested in.
With the default being zero or the "current" level.
This just removes every invalid index.
That means with `set foo a b c` and the "show" function from tests/expand.in:
- `show $foo[-5..-1]` prints "3 a b c"
- `show $foo[-10..1]` prints "1 a"
- `show $foo[2..5]` prints "2 b c"
- `show $foo[1 3 7 2]` prints "3 a c b"
and similar for command substitutions.
Fixes#826.
This is another step to resolving issue #1310. It makes
`fish_breakpoint_prompt` a replacement for `fish_prompt` if it is defined
and we're presenting a prompt in the context of a `breakpoint` command.
This implements `status is-breakpoint` that returns true if the current
shell prompt is displayed in the context of a `breakpoint` command.
This also fixes several bugs. Most notably making `breakpoint` a no-op if
the shell isn't interactive. Also, typing `breakpoint` at an interactive
prompt should be an error rather than creating a new nested debugging
context.
Partial fix for #1310
This does several things. It fixes `builtin_function()` so that errors it
emits are displayed. As part of doing that I've removed the unnecessary
`out_err` parameter to make the interface like every other builtin.
This also fixes a regression introduced by #4000 which was attempting to
fix a bug introduced by #3649.
Fixes#4139
Running the tests on travis revealed that some compilers (or at least
with some options) call the wrong struct constructor if there is more
than one struct with the same name but differing definitions.
Hoist the code for parsing flags out of each individual subcommand and
into a function shared by all the subcommands. This reduces duplication
and potential for error. More importantly it makes the code that
actually implements the subcommand more prominent.
When 2.6.0 was released some people reported that the third-party `rbenv`
and `pyenv` commands were incorrectly depending on our `setenv` function
not behaving exactly like the csh command of the same name. Specifically,
our version had a bug. It allowed more than one value. It no longer
does so after it was rewritten so that the three auto-split vars were
correctly handled.
See issue #4103
The Haiku stdio library has a bug. If we set stdout to unbuffered and it
is attached to a tty it discards wide output. Given how we interact with
the tty it should be safe to replace the problematic `fputwc()` calls
with simple `write()` calls. This does depend on the rest of the fish
code that writes to the tty to ultimately call write() which is true at
this time and should remain true in the future.
Fixes#4100
This change does several things. First, it works around a quirk of the
`xgetttext` command that only recognizes description strings in even
numbered position on the command. Second, it allows descriptions
introduced by the `-d` short flag to be recognized.
More importantly, it normalizes the strings so that `xgettext` correctly
extracts them into the *.po file. Prior to this change many fish script
strings were ignored due to how they were written (e.g., single versus
double quotes).
Fixes#4073
This came up in the context of issue #4068. This change makes it more
likely that the correct translation from english to another language
will be done for the "Job ... has {ended,stopped}" message.
Users continue to be surprised that fish auto splits/joins three env
vars but not other similar vars. Mention this in the tutorial to make it
less likely new users are surprised by this behavior.
Fixes#4009
The problem was overlooking a `break` statement when refactoring a
`switch` block into a simpler `if...else...` block. This fixes the
behavior of the `history-token-search-backward` function and its forward
searching analog.
Fixes#4065
This started out as a refactoring to eliminate the lint warnings. Adding
unit tests revealed the current implementation does not behave as
implied. So this is a complete rewrite of the implementation. With the
addition of unit tests so that it doesn't break in the future and anyone
who thinks this new version behaves wrong can update the unit tests to
help ensure we're testing for the correct behavior.
Fixes#4027
It still performs the assignment even if the command substitution
returned unsuccessfully - `set foo (echo bar; false)` returns 1 but
sets $foo to bar.
Also use `type -p` instead of `which`.
* Added Magento2 CLI completions
This is the completion file for the Magento2 CLI application I use on my servers. It has an additional feature tho, I'm not sure if it fits into the fish completion philosophy:
If you provide limited access credentials, it will connect to the MySQL database and provide additional suggestions, such as available users, themes or indexers in the database. If this file is never touched, those suggestions simply won't show up. I, personally, find them to be pretty useful, though.
Should I remove those database suggestions before creating a PR?
* Removed functions using MySQL, updated formatting
* Several smaller fixes
* Improved descriptions
Tried to shorten the text as much as possible and removed unnecessary characters
Some platforms do not correctly define `struct dirent` so that its
`d_name` member is long enough for the longest file name. Work around
such broken definitions.
Fixes#4030
The LRU cache wants to store references from nodes back into the
lookup map, so that it is efficient to remove a node from the
map. However certain compilers refuse to form a std::map::iterator
with an incomplete type. Fix this by storing a pointer to the key
instead of the iterator.
2017-05-15 09:11:40 -07:00
1573 changed files with 680288 additions and 158632 deletions
This release of fish fixes a major issue discovered in fish 3.1.1:
- Commands such as `fzf` and `enhancd`, when used with `eval`, would hang. `eval` buffered output too aggressively, which has been fixed (#6955).
If you are upgrading from version 3.0.0 or before, please also review the release notes for 3.1.1, 3.1.0 and 3.1b1 (included below).
---
# fish 3.1.1 (released April 27, 2020)
This release of fish fixes a number of major issues discovered in fish 3.1.0.
- Commands which involve `. ( ... | psub)` now work correctly, as a bug in the `function --on-job-exit` option has been fixed (#6613).
- Conflicts between upstream packages for ripgrep and bat, and the fish packages, have been resolved (#5822).
- Starting fish in a directory without read access, such as via `su`, no longer crashes (#6597).
- Glob ordering changes which were introduced in 3.1.0 have been reverted, returning the order of globs to the previous state (#6593).
- Redirections using the deprecated caret syntax to a file descriptor (eg `^&2`) work correctly (#6591).
- Redirections that append to a file descriptor (eg `2>>&1`) work correctly (#6614).
- Building fish on macOS (#6602) or with new versions of GCC (#6604, #6609) is now successful.
-`time` is now correctly listed in the output of `builtin -n`, and `time --help` works correctly (#6598).
- Exported universal variables now update properly (#6612).
-`status current-command` gives the expected output when used with an environment override - that is, `F=B status current-command` returns `status` instead of `F=B` (#6635).
-`test` no longer crashes when used with "`nan`" or "`inf`" arguments, erroring out instead (#6655).
- Copying from the end of the command line no longer crashes fish (#6680).
-`read` no longer removes multiple separators when splitting a variable into a list, restoring the previous behaviour from fish 3.0 and before (#6650).
- Functions using `--on-job-exit` and `--on-process-exit` work reliably again (#6679).
- Functions using `--on-signal INT` work reliably in interactive sessions, as they did in fish 2.7 and before (#6649). These handlers have never worked in non-interactive sessions, and making them work is an ongoing process.
- Functions using `--on-variable` work reliably with variables which are set implicitly (rather than with `set`), such as "`fish_bind_mode`" and "`PWD`" (#6653).
- 256 colors are properly enabled under certain conditions that were incorrectly detected in fish 3.1.0 (`$TERM` begins with xterm, does not include "`256color`", and `$TERM_PROGRAM` is not set) (#6701).
- The Mercurial (`hg`) prompt no longer produces an error when the current working directory is removed (#6699). Also, for performance reasons it shows only basic information by default; to restore the detailed status, set `$fish_prompt_hg_show_informative_status`.
- The VCS prompt, `fish_vcs_prompt`, no longer displays Subversion (`svn`) status by default, due to the potential slowness of this operation (#6681).
- Pasting of commands has been sped up (#6713).
- Using extended Unicode characters, such as emoji, in a non-Unicode capable locale (such as the `C` or `POSIX` locale) no longer renders all output blank (#6736).
-`help` prefers to use `xdg-open`, avoiding the use of `open` on Debian systems where this command is actually `openvt` (#6739).
- Command lines starting with a space, which are not saved in history, now do not get autosuggestions. This fixes an issue with Midnight Commander integration (#6763), but may be changed in a future version.
- Copying to the clipboard no longer inserts a newline at the end of the content, matching fish 2.7 and earlier (#6927).
-`fzf` in complex pipes no longer hangs. More generally, code run as part of command substitutions or `eval` will no longer have separate process groups. (#6624, #6806).
This release also includes:
- a number of changes to improve macOS compatibility with code signing and notarization;
- a number of improvements to completions; and
- a number of content and formatting improvements to the documentation.
If you are upgrading from version 3.0.0 or before, please also review the release notes for 3.1.0 and 3.1b1 (included below).
## Errata for fish 3.1
A new builtin, `time`, was introduced in the fish 3.1 releases. This builtin is a reserved word (like `test`, `function`, and others) because of the way it is implemented, and functions can no longer be named `time`. This was not clear in the fish 3.1b1 changelog.
---
# fish 3.1.0 (released February 12, 2020)
Compared to the beta release of fish 3.1b1, fish version 3.1.0:
- fixes a regression where spaces after a brace were removed despite brace expansion not occurring (#6564)
- fixes a number of problems in compiling and testing on Cygwin (#6549) and Solaris-derived systems such as Illumos (#6553, #6554, #6555, #6556, and #6558);
- fixes the process for building macOS packages;
- fixes a regression where excessive error messages are printed if Unicode characters are emitted in non-Unicode-capable locales (#6584); and
- contains some improvements to the documentation and a small number of completions.
If you are upgrading from version 3.0.0 or before, please also review the release notes for 3.1b1 (included below).
---
# fish 3.1b1 (released January 26, 2020)
## Notable improvements and fixes
- A new `$pipestatus` variable contains a list of exit statuses of the previous job, for each of the separate commands in a pipeline (#5632).
- fish no longer buffers pipes to the last function in a pipeline, improving many cases where pipes appeared to block or hang (#1396).
- An overhaul of error messages for builtin commands, including a removal of the overwhelming usage summary, more readable stack traces (#3404, #5434), and stack traces for `test` (aka `[`) (#5771).
- fish's debugging arguments have been significantly improved. The `--debug-level` option has been removed, and a new `--debug` option replaces it. This option accepts various categories, which may be listed via `fish --print-debug-categories` (#5879). A new `--debug-output` option allows for redirection of debug output.
-`string` has a new `collect` subcommand for use in command substitutions, producing a single output instead of splitting on new lines (similar to `"$(cmd)"` in other shells) (#159).
- The fish manual, tutorial and FAQ are now available in `man` format as `fish-doc`, `fish-tutorial` and `fish-faq` respectively (#5521).
- Like other shells, `cd` now always looks for its argument in the current directory as a last resort, even if the `CDPATH` variable does not include it or "." (#4484).
- fish now correctly handles `CDPATH` entries that start with `..` (#6220) or contain `./` (#5887).
- The `fish_trace` variable may be set to trace execution (#3427). This performs a similar role as `set -x` in other shells.
- fish uses the temporary directory determined by the system, rather than relying on `/tmp` (#3845).
- The fish Web configuration tool (`fish_config`) prints a list of commands it is executing, to help understanding and debugging (#5584).
- Major performance improvements when pasting (#5866), executing lots of commands (#5905), importing history from bash (#6295), and when completing variables that might match `$history` (#6288).
### Syntax changes and new commands
- A new builtin command, `time`, which allows timing of fish functions and builtins as well as external commands (#117).
- Brace expansion now only takes place if the braces include a "," or a variable expansion, meaning common commands such as `git reset HEAD@{0}` do not require escaping (#5869).
- New redirections `&>` and `&|` may be used to redirect or pipe stdout, and also redirect stderr to stdout (#6192).
-`switch` now allows arguments that expand to nothing, like empty variables (#5677).
- The `VAR=val cmd` syntax can now be used to run a command in a modified environment (#6287).
-`and` is no longer recognised as a command, so that nonsensical constructs like `and and and` produce a syntax error (#6089).
-`math`'s exponent operator, '`^`', was previously left-associative, but now uses the more commonly-used right-associative behaviour (#6280). This means that `math '3^0.5^2'` was previously calculated as '(3^0.5)^2', but is now calculated as '3^(0.5^2)'.
- In fish 3.0, the variable used with `for` loops inside command substitutions could leak into enclosing scopes; this was an inadvertent behaviour change and has been reverted (#6480).
### Scripting improvements
-`string split0` now returns 0 if it split something (#5701).
- In the interest of consistency, `builtin -q` and `command -q` can now be used to query if a builtin or command exists (#5631).
-`math` now accepts `--scale=max` for the maximum scale (#5579).
-`builtin $var` now works correctly, allowing a variable as the builtin name (#5639).
-`cd` understands the `--` argument to make it possible to change to directories starting with a hyphen (#6071).
-`complete --do-complete` now also does fuzzy matches (#5467).
-`complete --do-complete` can be used inside completions, allowing limited recursion (#3474).
-`count` now also counts lines fed on standard input (#5744).
-`eval` produces an exit status of 0 when given no arguments, like other shells (#5692).
-`printf` prints what it can when input hasn't been fully converted to a number, but still prints an error (#5532).
-`complete -C foo` now works as expected, rather than requiring `complete -Cfoo`.
-`complete` has a new `--force-files` option, to re-enable file completions. This allows `sudo -E` and `pacman -Qo` to complete correctly (#5646).
-`argparse` now defaults to showing the current function name (instead of `argparse`) in its errors, making `--name` often superfluous (#5835).
-`argparse` has a new `--ignore-unknown` option to keep unrecognized options, allowing multiple argparse passes to parse options (#5367).
-`argparse` correctly handles flag value validation of options that only have short names (#5864).
-`read -S` (short option of `--shell`) is recognised correctly (#5660).
-`read` understands `--list`, which acts like `--array` in reading all arguments into a list inside a single variable, but is better named (#5846).
-`read` has a new option, `--tokenize`, which splits a string into variables according to the shell's tokenization rules, considering quoting, escaping, and so on (#3823).
-`read` interacts more correctly with the deprecated `$IFS` variable, in particular removing multiple separators when splitting a variable into a list (#6406), matching other shells.
-`fish_indent` now handles semicolons better, including leaving them in place for `; and` and `; or` instead of breaking the line (#5859).
-`fish_indent --write` now supports multiple file arguments, indenting them in turn.
- The default read limit has been increased to 100MiB (#5267).
-`math` now also understands `x` for multiplication, provided it is followed by whitespace (#5906).
-`math` reports the right error when incorrect syntax is used inside parentheses (#6063), and warns when unsupported logical operations are used (#6096).
-`functions --erase` now also prevents fish from autoloading a function for the first time (#5951).
-`jobs --last` returns 0 to indicate success when a job is found (#6104).
-`commandline -p` and `commandline -j` now split on `&&` and `||` in addition to `;` and `&` (#6214).
- A bug where `string split` would drop empty strings if the output was only empty strings has been fixed (#5987).
-`eval` no long creates a new local variable scope, but affects variables in the scope it is called from (#4443). `source` still creates a new local scope.
-`abbr` has a new `--query` option to check for the existence of an abbreviation.
- Local values for `fish_complete_path` and `fish_function_path` are now ignored; only their global values are respected.
- Syntax error reports now display a marker in the correct position (#5812).
- Empty universal variables may now be exported (#5992).
- Exported universal variables are no longer imported into the global scope, preventing shadowing. This makes it easier to change such variables for all fish sessions and avoids breakage when the value is a list of multiple elements (#5258).
- A bug where `for` could use invalid variable names has been fixed (#5800).
- A bug where local variables would not be exported to functions has been fixed (#6153).
- The null command (`:`) now always exits successfully, rather than passing through the previous exit status (#6022).
- The output of `functions FUNCTION` matches the declaration of the function, correctly including comments or blank lines (#5285), and correctly includes any `--wraps` flags (#1625).
-`type` supports a new option, `--short`, which suppress function expansion (#6403).
-`type --path` with a function argument will now output the path to the file containing the definition of that function, if it exists.
-`type --force-path` with an argument that cannot be found now correctly outputs nothing, as documented (#6411).
- The `$hostname` variable is no longer truncated to 32 characters (#5758).
- Line numbers in function backtraces are calculated correctly (#6350).
- A new `fish_cancel` event is emitted when the command line is cancelled, which is useful for terminal integration (#5973).
### Interactive improvements
- New Base16 color options are available through the Web-based configuration (#6504).
- fish only parses `/etc/paths` on macOS in login shells, matching the bash implementation (#5637) and avoiding changes to path ordering in child shells (#5456). It now ignores blank lines like the bash implementation (#5809).
- The locale is now reloaded when the `LOCPATH` variable is changed (#5815).
-`read` no longer keeps a history, making it suitable for operations that shouldn't end up there, like password entry (#5904).
-`dirh` outputs its stack in the correct order (#5477), and behaves as documented when universal variables are used for its stack (#5797).
-`funced` and the edit-commandline-in-buffer bindings did not work in fish 3.0 when the `$EDITOR` variable contained spaces; this has been corrected (#5625).
- Builtins now pipe their help output to a pager automatically (#6227).
-`set_color` now colors the `--print-colors` output in the matching colors if it is going to a terminal.
- fish now underlines every valid entered path instead of just the last one (#5872).
- When syntax highlighting a string with an unclosed quote, only the quote itself will be shown as an error, instead of the whole argument.
- Syntax highlighting works correctly with variables as commands (#5658) and redirections to close file descriptors (#6092).
-`help` works properly on Windows Subsytem for Linux (#5759, #6338).
- A bug where `disown` could crash the shell has been fixed (#5720).
- fish will not autosuggest files ending with `~` unless there are no other candidates, as these are generally backup files (#985).
- Escape in the pager works correctly (#5818).
- Key bindings that call `fg` no longer leave the terminal in a broken state (#2114).
- Brackets (#5831) and filenames containing `$` (#6060) are completed with appropriate escaping.
- The output of `complete` and `functions` is now colorized in interactive terminals.
- The Web-based configuration handles aliases that include single quotes correctly (#6120), and launches correctly under Termux (#6248) and OpenBSD (#6522).
-`function` now correctly validates parameters for `--argument-names` as valid variable names (#6147) and correctly parses options following `--argument-names`, as in "`--argument-names foo --description bar`" (#6186).
- History newly imported from bash includes command lines using `&&` or `||`.
- The automatic generation of completions from manual pages is better described in job and process listings, and no longer produces a warning when exiting fish (#6269).
- In private mode, setting `$fish_greeting` to an empty string before starting the private session will prevent the warning about history not being saved from being printed (#6299).
- In the interactive editor, a line break (Enter) inside unclosed brackets will insert a new line, rather than executing the command and producing an error (#6316).
- Ctrl-C always repaints the prompt (#6394).
- When run interactively from another program (such as Python), fish will correctly start a new process group, like other shells (#5909).
- Job identifiers (for example, for background jobs) are assigned more logically (#6053).
- A bug where history would appear truncated if an empty command was executed was fixed (#6032).
#### New or improved bindings
- Pasting strips leading spaces to avoid pasted commands being omitted from the history (#4327).
- Shift-Left and Shift-Right now default to moving backwards and forwards by one bigword (words separated by whitespace) (#1505).
- The default escape delay (to differentiate between the escape key and an alt-combination) has been reduced to 30ms, down from 300ms for the default mode and 100ms for Vi mode (#3904).
- The `forward-bigword` binding now interacts correctly with autosuggestions (#5336).
- The `fish_clipboard_*` functions support Wayland by using [`wl-clipboard`](https://github.com/bugaevc/wl-clipboard) (#5450).
- The `nextd` and `prevd` functions no longer print "Hit end of history", instead using a bell. They correctly store working directories containing symbolic links (#6395).
- If a `fish_mode_prompt` function exists, Vi mode will only execute it on mode-switch instead of the entire prompt. This should make it much more responsive with slow prompts (#5783).
- The path-component bindings (like Ctrl-w) now also stop at ":" and "@", because those are used to denote user and host in commands such as `ssh` (#5841).
- The NULL character can now be bound via `bind -k nul`. Terminals often generate this character via control-space. (#3189).
- A new readline command `expand-abbr` can be used to trigger abbreviation expansion (#5762).
- A new readline command, `delete-or-exit`, removes a character to the right of the cursor or exits the shell if the command line is empty (moving this functionality out of the `delete-or-exit` function).
- The `self-insert` readline command will now insert the binding sequence, if not empty.
- A new binding to prepend `sudo`, bound to Alt-S by default (#6140).
- The Alt-W binding to describe a command should now work better with multiline prompts (#6110)
- The Alt-H binding to open a command's man page now tries to ignore `sudo` (#6122).
- A new pair of bind functions, `history-prefix-search-backward` (and `forward`), was introduced (#6143).
- Vi mode now supports R to enter replace mode (#6342), and `d0` to delete the current line (#6292).
- In Vi mode, hitting Enter in replace-one mode no longer erases the prompt (#6298).
- Selections in Vi mode are inclusive, matching the actual behaviour of Vi (#5770).
#### Improved prompts
- The Git prompt in informative mode now shows the number of stashes if enabled.
- The Git prompt now has an option (`$__fish_git_prompt_use_informative_chars`) to use the (more modern) informative characters without enabling informative mode.
- The default prompt now also features VCS integration and will color the host if running via SSH (#6375).
- The default and example prompts print the pipe status if an earlier command in the pipe fails.
- The default and example prompts try to resolve exit statuses to signal names when appropriate.
#### Improved terminal output
- New `fish_pager_color_` options have been added to control more elements of the pager's colors (#5524).
- Better detection and support for using fish from various system consoles, where limited colors and special characters are supported (#5552).
- fish now tries to guess if the system supports Unicode 9 (and displays emoji as wide), eliminating the need to set `$fish_emoji_width` in most cases (#5722).
- Improvements to the display of wide characters, particularly Korean characters and emoji (#5583, #5729).
- The Vi mode cursor is correctly redrawn when regaining focus under terminals that report focus (eg tmux) (#4788).
- Variables that control background colors (such as `fish_pager_color_search_match`) can now use `--reverse`.
#### Completions
- Added completions for
-`aws`
-`bat` (#6052)
-`bosh` (#5700)
-`btrfs`
-`camcontrol`
-`cf` (#5700)
-`chronyc` (#6496)
-`code` (#6205)
-`cryptsetup` (#6488)
-`csc` and `csi` (#6016)
-`cwebp` (#6034)
-`cygpath` and `cygstart` (#6239)
-`epkginfo` (#5829)
-`ffmpeg`, `ffplay`, and `ffprobe` (#5922)
-`fsharpc` and `fsharpi` (#6016)
-`fzf` (#6178)
-`g++` (#6217)
-`gpg1` (#6139)
-`gpg2` (#6062)
-`grub-mkrescue` (#6182)
-`hledger` (#6043)
-`hwinfo` (#6496)
-`irb` (#6260)
-`iw` (#6232)
-`kak`
-`keepassxc-cli` (#6505)
-`keybase` (#6410)
-`loginctl` (#6501)
-`lz4`, `lz4c` and `lz4cat` (#6364)
-`mariner` (#5718)
-`nethack` (#6240)
-`patool` (#6083)
-`phpunit` (#6197)
-`plutil` (#6301)
-`pzstd` (#6364)
-`qubes-gpg-client` (#6067)
-`resolvectl` (#6501)
-`rg`
-`rustup`
-`sfdx` (#6149)
-`speedtest` and `speedtest-cli` (#5840)
-`src` (#6026)
-`tokei` (#6085)
-`tsc` (#6016)
-`unlz4` (#6364)
-`unzstd` (#6364)
-`vbc` (#6016)
-`zpaq` (#6245)
-`zstd`, `zstdcat`, `zstdgrep`, `zstdless` and `zstdmt` (#6364)
- Lots of improvements to completions.
- Selecting short options which also have a long name from the completion pager is possible (#5634).
- Tab completion will no longer add trailing spaces if they already exist (#6107).
- Completion of subcommands to builtins like `and` or `not` now works correctly (#6249).
- Completion of arguments to short options works correctly when multiple short options are used together (#332).
- Activating completion in the middle of an invalid completion does not move the cursor any more, making it easier to fix a mistake (#4124).
- Completion in empty commandlines now lists all available commands.
- Functions listed as completions could previously leak parts of the function as other completions; this has been fixed.
### Deprecations and removed features
- The vcs-prompt functions have been promoted to names without double-underscore, so __fish_git_prompt is now fish_git_prompt, __fish_vcs_prompt is now fish_vcs_prompt, __fish_hg_prompt is now fish_hg_prompt and __fish_svn_prompt is now fish_svn_prompt. Shims at the old names have been added, and the variables have kept their old names (#5586).
-`string replace` has an additional round of escaping in the replacement expression, so escaping backslashes requires many escapes (eg `string replace -ra '([ab])' '\\\\\\\$1' a`). The new feature flag `regex-easyesc` can be used to disable this, so that the same effect can be achieved with `string replace -ra '([ab])' '\\\\$1' a` (#5556). As a reminder, the intention behind feature flags is that this will eventually become the default and then only option, so scripts should be updated.
- The `fish_vi_mode` function, deprecated in fish 2.3, has been removed. Use `fish_vi_key_bindings` instead (#6372).
### For distributors and developers
- fish 3.0 introduced a CMake-based build system. In fish 3.1, both the Autotools-based build and legacy Xcode build system have been removed, leaving only the CMake build system. All distributors and developers must install CMake.
- fish now depends on the common `tee` external command, for the `psub` process substitution function.
- The documentation is now built with Sphinx. The old Doxygen-based documentation system has been removed. Developers, and distributors who wish to rebuild the documentation, must install Sphinx.
- The `INTERNAL_WCWIDTH` build option has been removed, as fish now always uses an internal `wcwidth` function. It has a number of configuration options that make it more suitable for general use (#5777).
- mandoc can now be used to format the output from `--help` if `nroff` is not installed, reducing the number of external dependencies on systems with `mandoc` installed (#5489).
- Some bugs preventing building on Solaris-derived systems such as Illumos were fixed (#5458, #5461, #5611).
- Completions for `npm`, `bower` and `yarn` no longer require the `jq` utility for full functionality, but will use Python instead if it is available.
- The paths for completions, functions and configuration snippets have been extended. On systems that define `XDG_DATA_DIRS`, each of the directories in this variable are searched in the subdirectories `fish/vendor_completions.d`, `fish/vendor_functions.d`, and `fish/vendor_conf.d` respectively. On systems that do not define this variable in the environment, the vendor directories are searched for in both the installation prefix and the default "extra" directory, which now defaults to `/usr/local` (#5029).
---
# fish 3.0.2 (released February 19, 2019)
This release of fish fixes an issue discovered in fish 3.0.1.
### Fixes and improvements
- The PWD environment variable is now ignored if it does not resolve to the true working directory, fixing strange behaviour in terminals started by editors and IDEs (#5647).
If you are upgrading from version 2.7.1 or before, please also review the release notes for 3.0.1,
3.0.0 and 3.0b1 (included below).
---
# fish 3.0.1 (released February 11, 2019)
This release of fish fixes a number of major issues discovered in fish 3.0.0.
### Fixes and improvements
-`exec` does not complain about running foreground jobs when called (#5449).
- while loops now evaluate to the last executed command in the loop body (or zero if the body was empty), matching POSIX semantics (#4982).
-`read --silent` no longer echoes to the tty when run from a non-interactive script (#5519).
- On macOS, path entries with spaces in `/etc/paths` and `/etc/paths.d` now correctly set path entries with spaces. Likewise, `MANPATH` is correctly set from `/etc/manpaths` and `/etc/manpaths.d` (#5481).
- fish starts correctly under Cygwin/MSYS2 (#5426).
- The `pager-toggle-search` binding (Ctrl-S by default) will now activate the search field, even when the pager is not focused.
- The error when a command is not found is now printed a single time, instead of once per argument (#5588).
- Fixes and improvements to the git completions, including printing correct paths with older git versions, fuzzy matching again, reducing unnecessary offers of root paths (starting with `:/`) (#5578, #5574, #5476), and ignoring shell aliases, so enterprising users can set up the wrapping command (via `set -g __fish_git_alias_$command $whatitwraps`) (#5412).
- Significant performance improvements to core shell functions (#5447) and to the `kill` completions (#5541).
- Starting in symbolically-linked working directories works correctly (#5525).
- The default `fish_title` function no longer contains extra spaces (#5517).
- The `nim` prompt now works correctly when chosen in the Web-based configuration (#5490).
-`string` now prints help to stdout, like other builtins (#5495).
- Killing the terminal while fish is in vi normal mode will no longer send it spinning and eating CPU. (#5528)
- A number of crashes have been fixed (#5550, #5548, #5479, #5453).
- Improvements to the documentation and certain completions.
### Known issues
There is one significant known issue that was not corrected before the release:
- fish does not run correctly under Windows Services for Linux before Windows 10 version 1809/17763, and the message warning of this may not be displayed (#5619).
If you are upgrading from version 2.7.1 or before, please also review the release notes for 3.0.0 and 3.0b1 (included below).
---
# fish 3.0.0 (released December 28, 2018)
fish 3 is a major release, which introduces some breaking changes alongside improved functionality. Although most existing scripts will continue to work, they should be reviewed against the list contained in the 3.0b1 release notes below.
Compared to the beta release of fish 3.0b1, fish version 3.0.0:
- builds correctly against musl libc (#5407)
- handles huge numeric arguments to `test` correctly (#5414)
- removes the history colouring introduced in 3.0b1, which did not always work correctly
There is one significant known issue which was not able to be corrected before the release:
- fish 3.0.0 builds on Cygwin (#5423), but does not run correctly (#5426) and will result in a hanging terminal when started. Cygwin users are encouraged to continue using 2.7.1 until a release which corrects this is available.
If you are upgrading from version 2.7.1 or before, please also review the release notes for 3.0b1 (included below).
---
# fish 3.0b1 (released December 11, 2018)
fish 3 is a major release, which introduces some breaking changes alongside improved functionality. Although most existing scripts will continue to work, they should be reviewed against the list below.
## Notable non-backward compatible changes
- Process and job expansion has largely been removed. `%` will no longer perform these expansions, except for `%self` for the PID of the current shell. Additionally, job management commands (`disown`, `wait`, `bg`, `fg` and `kill`) will expand job specifiers starting with `%` (#4230, #1202).
-`set x[1] x[2] a b`, to set multiple elements of an array at once, is no longer valid syntax (#4236).
- A literal `{}` now expands to itself, rather than nothing. This makes working with `find -exec` easier (#1109, #4632).
- Literally accessing a zero-index is now illegal syntax and is caught by the parser (#4862). (fish indices start at 1)
- Successive commas in brace expansions are handled in less surprising manner. For example, `{,,,}` expands to four empty strings rather than an empty string, a comma and an empty string again (#3002, #4632).
-`for` loop control variables are no longer local to the `for` block (#1935).
- Variables set in `if` and `while` conditions are available outside the block (#4820).
- Local exported (`set -lx`) vars are now visible to functions (#1091).
- The new `math` builtin (see below) does not support logical expressions; `test` should be used instead (#4777).
- Range expansion will now behave sensibly when given a single positive and negative index (`$foo[5..-1]` or `$foo[-1..5]`), clamping to the last valid index without changing direction if the list has fewer elements than expected.
-`read` now uses `-s` as short for `--silent` (à la `bash`); `--shell`'s abbreviation (formerly `-s`) is now `-S` instead (#4490).
-`cd` no longer resolves symlinks. fish now maintains a virtual path, matching other shells (#3350).
-`source` now requires an explicit `-` as the filename to read from the terminal (#2633).
- Arguments to `end` are now errors, instead of being silently ignored.
- The names `argparse`, `read`, `set`, `status`, `test` and `[` are now reserved and not allowed as function names. This prevents users unintentionally breaking stuff (#3000).
- The `fish_user_abbreviations` variable is no longer used; abbreviations will be migrated to the new storage format automatically.
- The `FISH_READ_BYTE_LIMIT` variable is now called `fish_byte_limit` (#4414).
- Environment variables are no longer split into arrays based on the record separator character on startup. Instead, variables are not split, unless their name ends in PATH, in which case they are split on colons (#436).
- The `history` builtin's `--with-time` option has been removed; this has been deprecated in favor of `--show-time` since 2.7.0 (#4403).
- The internal variables `__fish_datadir` and `__fish_sysconfdir` are now known as `__fish_data_dir` and `__fish_sysconf_dir` respectively.
## Deprecations
With the release of fish 3, a number of features have been marked for removal in the future. All users are encouraged to explore alternatives. A small number of these features are currently behind feature flags, which are turned on at present but may be turned off by default in the future.
A new feature flags mechanism is added for staging deprecations and breaking changes. Feature flags may be specified at launch with `fish --features ...` or by setting the universal `fish_features` variable. (#4940)
- The use of the `IFS` variable for `read` is deprecated; `IFS` will be ignored in the future (#4156). Use the `read --delimiter` option instead.
- The `function --on-process-exit` switch will be removed in future (#4700). Use the `fish_exit` event instead: `function --on-event fish_exit`.
-`$_` is deprecated and will removed in the future (#813). Use `status current-command` in a command substitution instead.
-`^` as a redirection deprecated and will be removed in the future. (#4394). Use `2>` to redirect stderr. This is controlled by the `stderr-nocaret` feature flag.
-`?` as a glob (wildcard) is deprecated and will be removed in the future (#4520). This is controlled by the `qmark-noglob` feature flag.
## Notable fixes and improvements
### Syntax changes and new commands
- fish now supports `&&` (like `and`), `||` (like `or`), and `!` (like `not`), for better migration from POSIX-compliant shells (#4620).
- Variables may be used as commands (#154).
- fish may be started in private mode via `fish --private`. Private mode fish sessions do not have access to the history file and any commands evaluated in private mode are not persisted for future sessions. A session variable `$fish_private_mode` can be queried to detect private mode and adjust the behavior of scripts accordingly to respect the user's wish for privacy.
- A new `wait` command for waiting on backgrounded processes (#4498).
-`math` is now a builtin rather than a wrapper around `bc` (#3157). Floating point computations is now used by default, and can be controlled with the new `--scale` option (#4478).
- Setting `$PATH` no longer warns on non-existent directories, allowing for a single $PATH to be shared across machines (eg via dotfiles) (#2969).
-`while` sets `$status` to a non-zero value if the loop is not executed (#4982).
- Command substitution output is now limited to 10 MB by default, controlled by the `fish_read_limit` variable (#3822). Notably, this is larger than most operating systems' argument size limit, so trying to pass argument lists this size to external commands has never worked.
- The machine hostname, where available, is now exposed as the `$hostname` reserved variable. This removes the dependency on the `hostname` executable (#4422).
- Bare `bind` invocations in config.fish now work. The `fish_user_key_bindings` function is no longer necessary, but will still be executed if it exists (#5191).
-`$fish_pid` and `$last_pid` are available as replacements for `%self` and `%last`.
### New features in commands
-`alias` has a new `--save` option to save the generated function immediately (#4878).
-`bind` has a new `--silent` option to ignore bind requests for named keys not available under the current terminal (#4188, #4431).
-`complete` has a new `--keep-order` option to show the provided or dynamically-generated argument list in the same order as specified, rather than alphabetically (#361).
-`exec` prompts for confirmation if background jobs are running.
-`funced` has a new `--save` option to automatically save the edited function after successfully editing (#4668).
-`functions` has a new ` --handlers` option to show functions registered as event handlers (#4694).
-`history search` supports globs for wildcard searching (#3136) and has a new `--reverse` option to show entries from oldest to newest (#4375).
-`jobs` has a new `--quiet` option to silence the output.
-`read` has a new `--delimiter` option for splitting input into arrays (#4256).
-`read` writes directly to stdout if called without arguments (#4407).
-`read` can now read individual lines into separate variables without consuming the input in its entirety via the new `/--line` option.
-`set` has new `--append` and `--prepend` options (#1326).
-`string match` with an empty pattern and `--entire` in glob mode now matches everything instead of nothing (#4971).
-`string split` supports a new `--no-empty` option to exclude empty strings from the result (#4779).
-`string` has new subcommands `split0` and `join0` for working with NUL-delimited output.
-`string` no longer stops processing text after NUL characters (#4605)
-`string escape` has a new `--style regex` option for escaping strings to be matched literally in `string` regex operations.
-`test` now supports floating point values in numeric comparisons.
### Interactive improvements
- A pipe at the end of a line now allows the job to continue on the next line (#1285).
- Italics and dim support out of the box on macOS for Terminal.app and iTerm (#4436).
-`cd` tab completions no longer descend into the deepest unambiguous path (#4649).
- Pager navigation has been improved. Most notably, moving down now wraps around, moving up from the commandline now jumps to the last element and moving right and left now reverse each other even when wrapping around (#4680).
- Typing normal characters while the completion pager is active no longer shows the search field. Instead it enters them into the command line, and ends paging (#2249).
- A new input binding `pager-toggle-search` toggles the search field in the completions pager on and off. By default, this is bound to Ctrl-S.
- Searching in the pager now does a full fuzzy search (#5213).
- The pager will now show the full command instead of just its last line if the number of completions is large (#4702).
- Abbreviations can be tab-completed (#3233).
- Tildes in file names are now properly escaped in completions (#2274).
- Wrapping completions (from `complete --wraps` or `function --wraps`) can now inject arguments. For example, `complete gco --wraps 'git checkout'` now works properly (#1976). The `alias` function has been updated to respect this behavior.
- Path completions now support expansions, meaning expressions like `python ~/<TAB>` now provides file suggestions just like any other relative or absolute path. (This includes support for other expansions, too.)
- Autosuggestions try to avoid arguments that are already present in the command line.
- Notifications about crashed processes are now always shown, even in command substitutions (#4962).
- The screen is no longer reset after a BEL, fixing graphical glitches (#3693).
- vi-mode now supports ';' and ',' motions. This introduces new {forward,backward}-jump-till and repeat-jump{,-reverse} bind functions (#5140).
- The `*y` vi-mode binding now works (#5100).
- True color is now enabled in neovim by default (#2792).
- Terminal size variables (`$COLUMNS`/`$LINES`) are now updated before `fish_prompt` is called, allowing the prompt to react (#904).
- Multi-line prompts no longer repeat when the terminal is resized (#2320).
-`xclip` support has been added to the clipboard integration (#5020).
- The Alt-P keybinding paginates the last command if the command line is empty.
-`$cmd_duration` is no longer reset when no command is executed (#5011).
- Deleting a one-character word no longer erases the next word as well (#4747).
- Token history search (Alt-Up) omits duplicate entries (#4795).
- The `fish_escape_delay_ms` timeout, allowing the use of the escape key both on its own and as part of a control sequence, was applied to all control characters; this has been reduced to just the escape key.
- Completing a function shows the description properly (#5206).
- Added completions for
-`ansible`, including `ansible-galaxy`, `ansible-playbook` and `ansible-vault` (#4697)
- Lots of improvements to completions (especially `darcs` (#5112), `git`, `hg` and `sudo`).
- Completions for `yarn` and `npm` now require the `all-the-package-names` NPM package for full functionality.
- Completions for `bower` and `yarn` now require the `jq` utility for full functionality.
- Improved French translations.
### Other fixes and improvements
- Significant performance improvements to `abbr` (#4048), setting variables (#4200, #4341), executing functions, globs (#4579), `string` reading from standard input (#4610), and slicing history (in particular, `$history[1]` for the last executed command).
- Fish's internal wcwidth function has been updated to deal with newer Unicode, and the width of some characters can be configured via the `fish_ambiguous_width` (#5149) and `fish_emoji_width` (#2652) variables. Alternatively, a new build-time option INTERNAL_WCWIDTH can be used to use the system's wcwidth instead (#4816).
-`functions` correctly supports `-d` as the short form of `--description`. (#5105)
-`/etc/paths` is now parsed like macOS' bash `path_helper`, fixing $PATH order (#4336, #4852) on macOS.
- Using a read-only variable in a `for` loop produces an error, rather than silently producing incorrect results (#4342).
- The universal variables filename no longer contains the hostname or MAC address. It is now at the fixed location `.config/fish/fish_variables` (#1912).
- Exported variables in the global or universal scope no longer have their exported status affected by local variables (#2611).
- Major rework of terminal and job handling to eliminate bugs (#3805, #3952, #4178, #4235, #4238, #4540, #4929, #5210).
- Improvements to the manual page completion generator (#2937, #4313).
-`suspend --force` now works correctly (#4672).
- Pressing Ctrl-C while running a script now reliably terminates fish (#5253).
### For distributors and developers
- fish ships with a new build system based on CMake. CMake 3.2 is the minimum required version. Although the autotools-based Makefile and the Xcode project are still shipped with this release, they will be removed in the near future. All distributors and developers are encouraged to migrate to the CMake build.
- Build scripts for most platforms no longer require bash, using the standard sh instead.
- The `hostname` command is no longer required for fish to operate.
--
# fish 2.7.1 (released December 23, 2017)
This release of fish fixes an issue where iTerm 2 on macOS would display a warning about paste bracketing being left on when starting a new fish session (#4521).
If you are upgrading from version 2.6.0 or before, please also review the release notes for 2.7.0 and 2.7b1 (included below).
--
# fish 2.7.0 (released November 23, 2017)
There are no major changes between 2.7b1 and 2.7.0. If you are upgrading from version 2.6.0 or before, please also review the release notes for 2.7b1 (included below).
Xcode builds and macOS packages could not be produced with 2.7b1, but this is fixed in 2.7.0.
--
# fish 2.7b1 (released October 31, 2017)
## Notable improvements
- A new `cdh` (change directory using recent history) command provides a more friendly alternative to prevd/nextd and pushd/popd (#2847).
- A new `argparse` command is available to allow fish script to parse arguments with the same behavior as builtin commands. This also includes the `fish_opt` helper command. (#4190).
- Invalid array indexes are now silently ignored (#826, #4127).
- Improvements to the debugging facility, including a prompt specific to the debugger (`fish_breakpoint_prompt`) and a `status is-breakpoint` subcommand (#1310).
-`string` supports new `lower` and `upper` subcommands, for altering the case of strings (#4080). The case changing is not locale-aware yet.- `string escape` has a new `--style=xxx` flag where `xxx` can be `script`, `var`, or `url` (#4150), and can be reversed with `string unescape` (#3543).
- History can now be split into sessions with the `fish_history` variable, or not saved to disk at all (#102).
- Read history is now controlled by the `fish_history` variable rather than the `--mode-name` flag (#1504).
-`command` now supports an `--all` flag to report all directories with the command. `which` is no longer a runtime dependency (#2778).
- fish can run commands before starting an interactive session using the new `--init-command`/`-C` options (#4164).
-`set` has a new `--show` option to show lots of information about variables (#4265).
## Other significant changes
- The `COLUMNS` and `LINES` environment variables are now correctly set the first time `fish_prompt` is run (#4141).
-`complete`'s `--no-files` option works as intended (#112).
-`echo -h` now correctly echoes `-h` in line with other shells (#4120).
- The `export` compatibility function now returns zero on success, rather than always returning 1 (#4435).
- Stop converting empty elements in MANPATH to "." (#4158). The behavior being changed was introduced in fish 2.6.0.
-`count -h` and `count --help` now return 1 rather than produce command help output (#4189).
- An attempt to `read` which stops because too much data is available still defines the variables given as parameters (#4180).
- A regression in fish 2.4.0 which prevented `pushd +1` from working has been fixed (#4091).
- A regression in fish 2.6.0 where multiple `read` commands in non-interactive scripts were broken has been fixed (#4206).
- A regression in fish 2.6.0 involving universal variables with side-effects at startup such as `set -U fish_escape_delay_ms 10` has been fixed (#4196).
- Added completions for:
-`as` (#4130)
-`cdh` (#2847)
-`dhcpd` (#4115)
-`ezjail-admin` (#4324)
- Fabric's `fab` (#4153)
-`grub-file` (#4119)
-`grub-install` (#4119)
-`jest` (#4142)
-`kdeconnect-cli`
-`magneto` (#4043, #4108)
-`mdadm` (#4198)
-`passwd` (#4209)
-`pip` and `pipenv` (#4448)
-`s3cmd` (#4332)
-`sbt` (#4347)
-`snap` (#4215)
- Sublime Text 3's `subl` (#4277)
- Lots of improvements to completions.
- Updated Chinese and French translations.
- Improved completions for:
-`apt`
-`cd` (#4061)
-`composer` (#4295)
-`eopkg`
-`flatpak` (#4456)
-`git` (#4117, #4147, #4329, #4368)
-`gphoto2`
-`killall` (#4052)
-`ln`
-`npm` (#4241)
-`ssh` (#4377)
-`tail`
-`xdg-mime` (#4333)
-`zypper` (#4325)
---
# fish 2.6.0 (released June 3, 2017)
Since the beta release of fish 2.6b1, fish version 2.6.0 contains a number of minor fixes, new completions for `magneto` (#4043), and improvements to the documentation.
@@ -29,6 +619,7 @@ If you are upgrading from version 2.5.0 or before, please also review the releas
-`read` will read up to 10 MiB by default, leaving the target variable empty and exiting with status 122 if the line is too long. You can set a different limit with the `FISH_READ_BYTE_LIMIT` variable.
-`read` supports a new `--silent` option to hide the characters typed (#838), for when reading sensitive data from the terminal. `read` also now accepts simple strings for the prompt (rather than scripts) with the new `-P` and `--prompt-str` options (#802).
-`export` and `setenv` now understand colon-separated `PATH`, `CDPATH` and `MANPATH` variables.
-`setenv` is no longer a simple alias for `set -gx` and will complain, just like the csh version, if given more than one value (#4103).
-`bind` supports a new `--list-modes` option (#3872).
-`bg` will check all of its arguments before backgrounding any jobs; any invalid arguments will cause a failure, but non-existent (eg recently exited) jobs are ignored (#3909).
-`funced` warns if the function being edited has not been modified (#3961).
@@ -36,7 +627,7 @@ If you are upgrading from version 2.5.0 or before, please also review the releas
-`status` supports a new `current-function` subcommand to print the current function name (#1743).
-`string` supports a new `repeat` subcommand (#3864). `string match` supports a new `--entire` option to emit the entire line matched by a pattern (#3957). `string replace` supports a new `--filter` option to only emit lines which underwent a replacement (#3348).
-`test` supports the `-k` option to test for sticky bits (#733).
-`umask` understands symbolic modes (#738).
-`umask` understands symbolic modes (#738).
- Empty components in the `CDPATH`, `MANPATH` and `PATH` variables are now converted to "." (#2106, #3914).
- New versions of ncurses (6.0 and up) wipe terminal scrollback buffers with certain commands; the `C-l` binding tries to avoid this (#2855).
- Some systems' `su` implementations do not set the `USER` environment variable; it is now reset for root users (#3916).
This document provides guidelines for making changes to the fish-shell project. This includes rules for how to format the code, naming conventions, etcetera. Generally known as the style of the code. It also includes recommended best practices such as creating a Travis-CI account so you can verify your changes pass all the tests before making a pull-request.
This document provides guidelines for making changes to the fish-shell project. This includes rules for how to format the code, naming conventions, etcetera. Generally known as the style of the code. It also includes recommended best practices such as creating a TravisCI account so you can verify that your changes pass all the tests before making a pullrequest.
See the bottom of this document for help on installing the linting and style reformatting tools discussed in the following sections.
Fish source should limit the C++ features it uses to those available in C++11. It should not use exceptions.
Before introducing a new dependency, please make it optional with graceful failure if possible. Add
any new dependencies to the README.md under the *Running* and/or *Building* sections.
## Versioning
The fish version is constructed by the *build_tools/git_version_gen.sh* script. For developers the version is the branch name plus the output of `git describe --always --dirty`. Normally the main part of the version will be the closest annotated tag. Which itself is usually the most recent release number (e.g., `2.6.0`).
## Include What You Use
You should not depend on symbols being visible to a `*.cpp` module from `#include` statements inside another header file. In other words if your module does `#include "common.h"` and that header does `#include "signal.h"` your module should not assume the sub-include is present. It should instead directly `#include "signal.h"` if it needs any symbol from that header. That makes the actual dependencies much clearer. It also makes it easy to modify the headers included by a specific header file without having to worry that will break any module (or header) that includes a particular header.
To help enforce this rule the `make lint` (and `make lint-all`) command will run the [include-what-you-use](http://include-what-you-use.org/) tool. You can find the IWYU project on [github](https://github.com/include-what-you-use/include-what-you-use).
To help enforce this rule the `make lint` (and `make lint-all`) command will run the [include-what-you-use](https://include-what-you-use.org/) tool. You can find the IWYU project on [github](https://github.com/include-what-you-use/include-what-you-use).
To install the tool on OS X you'll need to add a [formula](https://github.com/jasonmp85/homebrew-iwyu) then install it:
@@ -20,7 +26,11 @@ brew tap jasonmp85/iwyu
brew install iwyu
```
On Ubuntu you can install it via `sudo apt-get install iwyu`.
On Ubuntu you can install it via `apt-get`:
```
sudo apt-get install iwyu
```
## Lint Free Code
@@ -28,10 +38,12 @@ Automated analysis tools like cppcheck and oclint can point out potential bugs o
Ultimately we want lint free code. However, at the moment a lot of cleanup is required to reach that goal. For now simply try to avoid introducing new lint.
To make linting the code easy there are two make targets: `lint` and `lint-all`. The latter does just what the name implies. The former will lint any modified but not committed `*.cpp` files. If there is no uncommitted work it will lint the files in the most recent commit.
To make linting the code easy there are two make targets: `lint` and `lint-all`. The latter does exactly what the name implies. The former will lint any modified but not committed `*.cpp` files. If there is no uncommitted work it will lint the files in the most recent commit.
Fish has custom cppcheck rules in the file `.cppcheck.rule`. These help catch mistakes such as using `wcwidth()` rather than `fish_wcwidth()`. Please add a new rule if you find similar mistakes being made.
Fish also depends on `diff` and `expect` for its tests.
### Dealing With Lint Warnings
You are strongly encouraged to address a lint warning by refactoring the code, changing variable names, or whatever action is implied by the warning.
@@ -57,17 +69,17 @@ Suppressing oclint warnings is more complicated to describe so I'll refer you to
The following sections discuss the specific rules for the style that should be used when writing fish code. To ensure your changes conform to the style rules you simply need to run
```
make style
build_tools/style.fish
```
before commiting your change. That will run `git-clang-format` to rewrite just the lines you're modifying.
before committing your change. That will run `git-clang-format` to rewrite only the lines you're modifying.
If you've already committed your changes that's okay since it will then check the files in the most recent commit. This can be useful after you've merged someone elses change and want to check that it's style is acceptable. However, in that case it will run `clang-format` to ensure the entire file, not just the lines modified by the commit, conform to the style.
If you've already committed your changes that's okay since it will then check the files in the most recent commit. This can be useful after you've merged another person's change and want to check that it's style is acceptable. However, in that case it will run `clang-format` to ensure the entire file, not just the lines modified by the commit, conform to the style.
If you want to check the style of the entire code base run
```
make style-all
build_tools/style.fish --all
```
That command will refuse to restyle any files if you have uncommitted changes.
@@ -94,7 +106,25 @@ If you use Emacs: TBD
### Configuring Your Editor for Fish Scripts
If you use ViM: TBD
If you use ViM: Install [vim-fish](https://github.com/dag/vim-fish), make sure you have syntax and filetype functionality in `~/.vimrc`:
```
syntax enable
filetype plugin indent on
```
Then turn on some options for nicer display of fish scripts in `~/.vim/ftplugin/fish.vim`:
```
" Set up :make to use fish for syntax checking.
compiler fish
" Set this to have long lines wrap inside comments.
setlocal textwidth=79
" Enable folding of block structures in fish.
setlocal foldmethod=expr
```
If you use Emacs: Install [fish-mode](https://github.com/wwwjfy/emacs-fish) (also available in melpa and melpa-stable) and `(setq-default indent-tabs-mode nil)` for it (via a hook or in `use-package`s ":init" block). It can also be made to run fish_indent via e.g.
@@ -113,11 +143,13 @@ code to ignore
// clang-format on
```
However, as I write this there are no places in the code where we use this and I can't think of any legitimate reasons for exempting blocks of code from clang-format.
## Fish Script Style Guide
1. All fish scripts, such as those in the *share/functions* and *tests* directories, should be formatted using the `fish_indent` command.
1. Function names should be all lowercase with undescores separating words. Private functions should begin with an underscore. The first word should be `fish` if the function is unique to fish.
1. Function names should be in all lowercase with words separated by underscores. Private functions should begin with an underscore. The first word should be `fish` if the function is unique to fish.
1. The first word of global variable names should generally be `fish` for public vars or `_fish` for private vars to minimize the possibility of name clashes with user defined vars.
@@ -127,40 +159,54 @@ code to ignore
1. The `clang-format` command is authoritative with respect to indentation, whitespace around operators, etc.
1. All names in code should be `small_snake_case`. No Hungarian notation is used. Classes and structs names should be followed by `_t`.
1. All names in code should be `small_snake_case`. No Hungarian notation is used. The names for classes and structs should be followed by `_t`.
1. Always attach braces to the surrounding context.
1. Indent with spaces, not tabs and use four spaces per indent.
1.Comments should always use the C++ style; i.e., each line of the comment should begin with a `//` and should be limited to 100 characters. Comments that do not begin a line should be separated from the previous text by two spaces.
1. Comments that document the purpose of a function or class should begin with three slashes, `///`, so that OS X Xcode (and possibly other IDE's) will extract the comment and show it in the "Quick Help" window when the cursor is on the symbol.
1.Document the purpose of a function or class with doxygen-style comment blocks. e.g.:
```
/**
* Sum numbers in a vector.
*
* @param values Container whose values are summed.
* @return sum of `values`, or 0.0 if `values` is empty.
*/
double sum(std::vector<double> & const values) {
...
}
*/
```
or
```
/// brief description of somefunction()
void somefunction() {
```
## Testing
The source code for fish includes a large collection of tests. If you are making any changes to fish, running these tests is mandatory to make sure the behaviour remains consistent and regressions are not introduced. Even if you don't run the tests they will be run via the [Travis CI](https://travis-ci.org/fish-shell/fish-shell) service.
The source code for fish includes a large collection of tests. If you are making any changes to fish, running these tests is mandatory to make sure the behaviour remains consistent and regressions are not introduced. Even if you don't run the tests on your machine, they will still be run via the [Travis CI](https://travis-ci.org/fish-shell/fish-shell) service.
You are strongly encouraged to add tests when changing the functionality of fish. Especially if you are fixing a bug to help ensure there are no regressions in the future (i.e., we don't reintroduce the bug).
You are strongly encouraged to add tests when changing the functionality of fish, especially if you are fixing a bug to help ensure there are no regressions in the future (i.e., we don't reintroduce the bug).
### Local testing
The tests can be run on your local computer on all operating systems.
Running the tests is only supported from the autotools build and not xcodebuild. On OS X, you will need to install autoconf — we suggest using [Homebrew](http://brew.sh/) to install these tools.
autoconf
./configure
make test # or "gmake test" on BSD
```
cmake path/to/fish-shell
make test
```
### Travis CI Build and Test
The Travis Continuous Integration services can be used to test your changes using multiple configurations. This is the same service that the fishshell project uses to ensure new changes haven't broken anything. Thus it is a really good idea that you leverage Travis CI before making a pull-request to avoid embarrasment at breaking the build.
The Travis Continuous Integration services can be used to test your changes using multiple configurations. This is the same service that the fish-shell project uses to ensure new changes haven't broken anything. Thus it is a really good idea that you leverage Travis CI before making a pullrequest to avoid potential embarrassment at breaking the build.
You will need to [fork the fish-shell repository on GitHub](https://help.github.com/articles/fork-a-repo/). Then setup Travis to test your changes before you make a pull-request:
You will need to [fork the fish-shell repository on GitHub](https://help.github.com/articles/fork-a-repo/), then setup Travis to test your changes before making a pullrequest.
1. [Sign in to Travis CI](https://travis-ci.org/auth) with your GitHub account, accepting the GitHub access permissions confirmation.
1. Once you're signed in, and your repositories are synchronised, go to your [profile page](https://travis-ci.org/profile) and enable the fish-shell repository.
1. Once you're signed in and your repositories are synchronized, go to your [profile page](https://travis-ci.org/profile) and enable the fish-shell repository.
1. Push your changes to GitHub.
You'll receive an email when the tests are complete telling you whether or not any tests failed.
@@ -200,9 +246,9 @@ fi
exit 0
```
This will check if the push is to the master branch and, if it is, will run `make test` and only allow the push if that succeeds. In some circumstances it might be advisable to circumvent it with `git push --no-verify`, but usually that should not be necessary.
This will check if the push is to the master branch and, if it is, only allow the push if running `make test` succeeds. In some circumstances it may be advisable to circumvent this check with `git push --no-verify`, but usually that isn't necessary.
To install the hook, put it in .git/hooks/pre-push and make it executable.
To install the hook, place the code in a new file `.git/hooks/pre-push` and make it executable.
### Coverity Scan
@@ -212,7 +258,7 @@ We use Coverity's static analysis tool which offers free access to open source p
### Installing the Linting Tools
To install the lint checkers on Mac OS X using HomeBrew:
To install the lint checkers on Mac OS X using Homebrew:
```
brew tap oclint/formulae
@@ -220,7 +266,7 @@ brew install oclint
brew install cppcheck
```
To install the lint checkers on Linux distros that use Apt:
To install the lint checkers on Debian-based Linux distributions:
Fish uses the GNU gettext library to translate messages from English to other languages.
All non-debug messages output for user consumption should be marked for translation. In C++, this requires the use of the `_` (underscore) macro:
```
streams.out.append_format(_(L"%ls: There are no jobs\n"), argv[0]);
```
All messages in fish script must be enclosed in single or double quote characters. They must also be translated via a subcommand. This means that the following are **not** valid:
```
echo (_ hello)
_ "goodbye"
```
Above should be written like this instead:
```
echo (_ "hello")
echo (_ "goodbye")
```
Note that you can use either single or double quotes to enclose the message to be translated. You can also optionally include spaces after the opening parentheses and once again before the closing parentheses.
Creating and updating translations requires the Gettext tools, including `xgettext`, `msgfmt` and `msgmerge`. Translation sources are stored in the `po` directory, named `LANG.po`, where `LANG` is the two letter ISO 639-1 language code of the target language (eg `de` for German).
To create a new translation, for example for German:
* generate a `messages.pot` file by running `build_tools/fish_xgettext.fish` from the source tree
* copy `messages.pot` to `po/LANG.po` ()
To update a translation:
* generate a `messages.pot` file by running `build_tools/fish_xgettext.fish` from the source tree
* update the existing translation by running `msgmerge --update --no-fuzzy-matching po/LANG.po messages.pot`
Many tools are available for editing translation files, including command-line and graphical user interface programs.
Be cautious about blindly updating an existing translation file. Trivial changes to an existing message (eg changing the punctuation) will cause existing translations to be removed, since the tools do literal string matching. Therefore, in general, you need to carefully review any recommended deletions.
Read the [translations wiki](https://github.com/fish-shell/fish-shell/wiki/Translations) for more information.
[fish](https://fishshell.com/) - the friendly interactive shell [](https://travis-ci.org/fish-shell/fish-shell)
================================================
fish is a smart and user-friendly command line shell for OS X, Linux, and the rest of the family. fish includes features like syntax highlighting, autosuggest-as-you-type, and fancy tab completions that just work, with no configuration required.
fish is a smart and user-friendly command line shell for macOS, Linux, and the rest of the family.
fish includes features like syntax highlighting, autosuggest-as-you-type, and fancy tab completions
that just work, with no configuration required.
For more on fish's design philosophy, see the [design document](https://fishshell.com/docs/current/design.html).
@@ -11,88 +13,141 @@ fish generally works like other shells, like bash or zsh. A few important differ
Detailed user documentation is available by running `help` within fish, and also at <https://fishshell.com/docs/current/index.html>
## Building
You can quickly play with fish right in your browser by clicking the button below:
fish requires a C++11 compiler. It builds successfully with g++ 4.8 or later, or with clang 3.3 or later.
[](https://rootnroll.com/d/fish-shell/)
fish can be built using autotools or Xcode. autoconf 2.60 or later, as well as automake 1.13 or later, are required to build from git versions. These are not required to build from released tarballs.
## Getting fish
fish depends on a curses implementation, such as ncurses. The headers and libraries are required for building.
### macOS
fish requires PCRE2 due to the regular expression support contained in the `string` builtin. A copy is included with the source code, and will be used automatically if it does not already exist on your system.
fish can be installed:
fish requires gettext for translation support.
* using [Homebrew](http://brew.sh/): `brew install fish`
* using [MacPorts](https://www.macports.org/): `sudo port install fish`
* using the [installer from fishshell.com](https://fishshell.com/)
* as a [standalone app from fishshell.com](https://fishshell.com/)
Building the documentation requires Doxygen 1.8.7 or newer.
### Packages for Linux
### Autotools Build
Packages for Debian, Fedora, openSUSE, and Red Hat Enterprise Linux/CentOS are available from the
PPA](https://launchpad.net/~fish-shell/+archive/ubuntu/release-3), and can be installed using the
following commands:
### Xcode Development Build
```
sudo apt-add-repository ppa:fish-shell/release-3
sudo apt-get update
sudo apt-get install fish
```
* Build the `base` target in Xcode
* Run the fish executable, for example, in `DerivedData/fish/Build/Products/Debug/base/bin/fish`
Instructions for other distributions may be found at [fishshell.com](https://fishshell.com).
### Xcode Build and Install
### Windows
xcodebuild install
sudo ditto /tmp/fish.dst /
- On Windows 10, fish can be installed under the WSL Windows Subsystem for Linux with `sudo apt install fish` or from source with the instructions below.
- Fish can also be installed on all versions of Windows using [Cygwin](https://cygwin.com/) (from the **Shells** category).
## Help, it didn't build!
### Building from source
If fish reports that it could not find curses, try installing a curses development package and build again.
If packages are not available for your platform, GPG-signed tarballs are available from
[fishshell.com](https://fishshell.com/) and [fish-shell on
GitHub](https://github.com/fish-shell/fish-shell/releases). See the *Building* section for instructions.
Once installed, run `fish` from your current shell to try fish out!
On RedHat, CentOS, or Amazon EC2:
### Dependencies
sudo yum install ncurses-devel
Running fish requires:
## Runtime Dependencies
* curses or ncurses (preinstalled on most \*nix systems)
* some common \*nix system utilities (currently `mktemp`), in addition to the basic POSIX utilities (`cat`, `cut`, `dirname`, `ls`, `mkdir`, `mkfifo`, `rm`, `sort`, `tee`, `tr`, `uname` and `sed` at least, but the full coreutils plus find, sed and awk is preferred)
* gettext (library and `gettext` command), if compiled with translation support
fish requires a curses implementation, such as ncurses, to run.
The following optional features also have specific requirements:
fish requires PCRE2 due to the regular expression support contained in the `string` builtin. A bundled version will be compiled in automatically at build time if required.
* builtin commands that have the `--help` option or print usage messages require `ul` and either `nroff` or `mandoc` for display
* automated completion generation from manual pages requires Python (2.7+ or 3.3+) and possibly the
`backports.lzma` module for Python 2.7
* the `fish_config` web configuration tool requires Python (2.7+ or 3.3 +) and a web browser
* system clipboard integration (with the default Ctrl-V and Ctrl-X bindings) require either the
`xsel`, `xclip`, `wl-copy`/`wl-paste` or `pbcopy`/`pbpaste` utilities
* full completions for `yarn` and `npm` require the `all-the-package-names` NPM module
fish requires a number of utilities to operate, which should be present on any Unix, GNU/Linux or OS X system. These include (but are not limited to) hostname, grep, awk, sed, and getopt. fish also requires the bc program.
Translation support requires the gettext program.
Usage output for builtin functions is generated on-demand from the installed manpages using `nroff` and `ul`.
Some optional features of fish, such as the manual page completion parser and the web configuration tool, require Python.
In order to generate completions from man pages compressed with either lzma or xz, you may need to install an extra Python package. Python versions prior to 2.6 are not supported. To process lzma-compresed manpages, backports.lzma is needed for Python 3.2 or older. From version 3.3 onwards, Python already includes the required module.
## Packages for Linux
Instructions on how to find builds for several Linux distros are at <https://github.com/fish-shell/fish-shell/wiki/Nightly-builds>
## Switching to fish
### Switching to fish
If you wish to use fish as your default shell, use the following command:
chsh -s /usr/local/bin/fish
chsh will prompt you for your password, and change your default shell. Substitute "/usr/local/bin/fish" with whatever path to fish is in your /etc/shells file.
`chsh` will prompt you for your password and change your default shell. (Substitute `/usr/local/bin/fish` with whatever path fish was installed to, if it differs.) Log out, then log in again for the changes to take effect.
Use the following command if you didn't already add your fish path to /etc/shells.
Use the following command if fish isn't already added to `/etc/shells` to permit fish to be your login shell:
echo /usr/local/bin/fish | sudo tee -a /etc/shells
To switch your default shell back, you can run:
To switch your default shell back, you can run`chsh -s /bin/bash` (substituting `/bin/bash` with `/bin/tcsh` or `/bin/zsh` as appropriate).
chsh -s /bin/bash
## Building
Substitute /bin/bash with /bin/tcsh or /bin/zsh as appropriate.
### Dependencies
You may need to logout/login for the change (chsh) to take effect.
Compiling fish requires:
* a C++11 compiler (g++ 4.8 or later, or clang 3.3 or later)
* CMake (version 3.2 or later)
* a curses implementation such as ncurses (headers and libraries)
* PCRE2 (headers and libraries) - a copy is included with fish
* gettext (headers and libraries) - optional, for translation support
Sphinx is also optionally required to build the documentation from a cloned git repository.
### Building from source (all platforms) - Makefile generator
To install into `/usr/local`, run:
```bash
mkdir build;cd build
cmake ..
make
sudo make install
```
The install directory can be changed using the `-DCMAKE_INSTALL_PREFIX` parameter for `cmake`.
### Building from source (macOS) - Xcode
```bash
mkdir build;cd build
cmake .. -G Xcode
```
An Xcode project will now be available in the `build` subdirectory. You can open it with Xcode,
or run the following to build and install in `/usr/local`:
```bash
xcodebuild
xcodebuild -scheme install
```
The install directory can be changed using the `-DCMAKE_INSTALL_PREFIX` parameter for `cmake`.
### Help, it didn't build!
If fish reports that it could not find curses, try installing a curses development package and build again.
@@ -100,6 +155,6 @@ See the [Guide for Developers](CONTRIBUTING.md).
## Contact Us
Questions, comments, rants and raves can be posted to the official fish mailing list at <https://lists.sourceforge.net/lists/listinfo/fish-users> or join us on our [gitter.im channel](https://gitter.im/fish-shell/fish-shell) or IRC channel [#fish at irc.oftc.net](https://webchat.oftc.net/?channels=fish). Or use the [fish tag on Stackoverflow](https://stackoverflow.com/questions/tagged/fish) for questions related to fish script and the [fish tag on Superuser](https://superuser.com/questions/tagged/fish) for all other questions (e.g., customizing colors, changing key bindings).
Questions, comments, rants and raves can be posted to the official fish mailing list at <https://lists.sourceforge.net/lists/listinfo/fish-users> or join us on our [gitter.im channel](https://gitter.im/fish-shell/fish-shell). Or use the [fish tag on Stackoverflow](https://stackoverflow.com/questions/tagged/fish) for questions related to fish script and the [fish tag on Superuser](https://superuser.com/questions/tagged/fish) for all other questions (e.g., customizing colors, changing key bindings).
Found a bug? Have an awesome idea? Please open an issue on this github page.
Found a bug? Have an awesome idea? Please [open an issue](https://github.com/fish-shell/fish-shell/issues/new).
#Until now the makefile likely has been affecting our output, reset for upcoming warnings
tput sgr0
# Make sure INPUTDIR is found
iftest ! -d "$INPUTDIR";then
echo >&2"Could not find input directory '${INPUTDIR}'"
exit1
fi
# Make sure doxygen is found
DOXYGENPATH=`command -v doxygen`
iftest -z "$DOXYGENPATH";then
for i in /usr/local/bin/doxygen /opt/bin/doxygen /Applications/Doxygen.app/Contents/Resources/doxygen ~/Applications/Doxygen.app/Contents/Resources/doxygen ;do
iftest -f "$i";then
DOXYGENPATH="$i"
break
fi
done
fi
iftest -z "$DOXYGENPATH";then
echo >&2"doxygen is not installed, so documentation will not be built."
exit0
fi
# Check we have the lexicon filter
iftest -z "$INPUT_FILTER";then
echo >&2"Lexicon filter is not available. Continuing without."
INPUTFILTER=''
fi
# Determine where our output should go
if ! mkdir -p "${OUTPUTDIR}";then
echo"Could not create output directory '${OUTPUTDIR}'"
fi
# Make a temporary directory
TMPLOC=`mktemp -d -t fish_doc_build_XXXXXX`||{echo >&2"Could not build documentation because mktemp failed";exit 1;}
# Copy stuff to the temp directory
for i in "$INPUTDIR"/*.txt;do
INPUTFILE=$TMPLOC/`basename $i .txt`.doxygen
echo"/** \page"`basename $i .txt` > $INPUTFILE
cat $i >>$INPUTFILE
echo"*/" >>$INPUTFILE
done
# Make some extra stuff to pass to doxygen
# Input is kept as . because we cd to the input directory beforehand
# This prevents doxygen from generating "documentation" for intermediate directories
{cd"$PKGDIR/build"&& cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo -DMAC_CODESIGN_ID="${MAC_CODESIGN_ID}""$SRC_DIR"&& make -j 12&& envDESTDIR="$PKGDIR/root/" make install;}
# This is needed in order to get the really cool backtraces on Linux
#
AC_MSG_CHECKING([for -rdynamic linker flag])
prev_LDFLAGS="$LDFLAGS"
LDFLAGS="$LDFLAGS -rdynamic"
AC_LINK_IFELSE([AC_LANG_PROGRAM([[]],[[]])],
[
AC_MSG_RESULT([yes])
LDFLAGS_FISH="$LDFLAGS_FISH -rdynamic"
], [
AC_MSG_RESULT([no])
LDFLAGS_FISH="$LDFLAGS_FISH"
])
LDFLAGS="$prev_LDFLAGS"
#
# See if Linux procfs is present. This is used to get extra
# information about running processes.
#
AC_CHECK_FILES([/proc/self/stat])
# Disable curses macros that conflict with the STL
AC_DEFINE([NCURSES_NOMACROS], [1], [Define to 1 to disable ncurses macros that conflict with the STL])
AC_DEFINE([NOMACROS], [1], [Define to 1 to disable curses macros that conflict with the STL])
# Threading is excitingly broken on Solaris without adding -pthread to CXXFLAGS
# Only support GCC for now
dnl Ideally we would use the AX_PTHREAD macro here, but it's GPL3-licensed
dnl ACX_PTHREAD is way too old and seems to break the OS X build
dnl Both only check with AC_LANG(C) in any case
case $host_os in
solaris*)
CXXFLAGS="$CXXFLAGS -pthread"
CFLAGS="$CFLAGS -pthread"
;;
esac
#
# Check presense of various libraries. This is done on a per-binary
# level, since including various extra libraries in all binaries only
# because thay are used by some of them can cause extra bloat and
# slower compiles when developing fish.
#
# Check for os dependant libraries for all binaries.
AC_SEARCH_LIBS( nanosleep, rt, , [AC_MSG_ERROR([Cannot find the rt library, needed to build this package.] )] )
AC_SEARCH_LIBS( shm_open, rt, [AC_DEFINE([HAVE_SHM_OPEN], [1], [Define to 1 if the shm_open() function exists])] )
AC_SEARCH_LIBS( pthread_create, pthread, , [AC_MSG_ERROR([Cannot find the pthread library, needed to build this package.] )] )
AC_SEARCH_LIBS( setupterm, [ncurses tinfo curses], , [AC_MSG_ERROR([Could not find a curses implementation, needed to build fish. If this is Linux, try running 'sudo apt-get install libncurses5-dev' or 'sudo yum install ncurses-devel'])] )
# Although setupterm is linkable thanks to SEARCH_LIBS above, some
# builds of ncurses include the actual headers in a different package
#
AC_CHECK_DECL( [setupterm], , [AC_MSG_ERROR([Could not find a curses implementation, needed to build fish. If this is Linux, try running 'sudo apt-get install libncurses5-dev' or 'sudo yum install ncurses-devel'])], [
#if HAVE_NCURSES_H
#include <ncurses.h>
#elif HAVE_NCURSES_CURSES_H
#include <ncurses/curses.h>
#else
#include <curses.h>
#endif
#if HAVE_TERM_H
#include <term.h>
#elif HAVE_NCURSES_TERM_H
#include <ncurses/term.h>
#endif
] )
dnl AC_CHECK_FUNCS uses C linkage, but sometimes (Solaris!) the behaviour is
dnl different with C++.
AC_MSG_CHECKING([for wcsdup])
AC_TRY_LINK( [ #include <wchar.h> ],
[ wchar_t* foo = wcsdup(L""); ],
[ AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_WCSDUP, 1, Define to 1 if you have the `wcsdup' function.)
],
[AC_MSG_RESULT(no)],
)
AC_MSG_CHECKING([for std::wcsdup])
AC_TRY_LINK( [ #include <wchar.h> ],
[ wchar_t* foo = std::wcsdup(L""); ],
[ AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_STD__WCSDUP, 1, Define to 1 if you have the `std::wcsdup' function.)
],
[AC_MSG_RESULT(no)],
)
AC_MSG_CHECKING([for wcscasecmp])
AC_TRY_LINK( [ #include <wchar.h> ],
[ int foo = wcscasecmp(L"", L""); ],
[ AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_WCSCASECMP, 1, Define to 1 if you have the `wcscasecmp' function.)
],
[AC_MSG_RESULT(no)],
)
AC_MSG_CHECKING([for std::wcscasecmp])
AC_TRY_LINK( [ #include <wchar.h> ],
[ int foo = std::wcscasecmp(L"", L""); ],
[ AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_STD__WCSCASECMP, 1, Define to 1 if you have the `std::wcscasecmp' function.)
],
[AC_MSG_RESULT(no)],
)
AC_MSG_CHECKING([for wcsncasecmp])
AC_TRY_LINK( [ #include <wchar.h> ],
[ int foo = wcsncasecmp(L"", L"", 0); ],
[ AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_WCSNCASECMP, 1, Define to 1 if you have the `wcsncasecmp' function.)
],
[AC_MSG_RESULT(no)],
)
AC_MSG_CHECKING([for std::wcsncasecmp])
AC_TRY_LINK( [ #include <wchar.h> ],
[ int foo = std::wcsncasecmp(L"", L"", 0); ],
[ AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_STD__WCSNCASECMP, 1, Define to 1 if you have the `std::wcsncasecmp' function.)
],
[AC_MSG_RESULT(no)],
)
if test x$local_gettext != xno; then
AC_CHECK_FUNCS( gettext )
#
# The Makefile also needs to know if we have gettext, so it knows if
The fish documentation has been updated to support Doxygen 1.8.7+, and while the main benefit of this change is extensive Markdown support, the addition of a fish lexicon and syntax filter, combined with semantic markup rules allows for automatic formatting enhancements across the HTML user_docs and man pages.
Initially my motivation was to fix a problem with long options ([Issue #1557](https://github.com/fish-shell/fish-shell/issues/1557) on GitHub), but as I worked on fixing the issue I realised there was an opportunity to simplify, reinforce and clarify the current documentation, hopefully making further contribution easier and cleaner, while allowing the documentation examples to presented more clearly with less author effort.
While the documentation is pretty robust to variations in the documentation source, adherence to the following style guide will help keep the already excellent documention in good shape moving forward.
## Line breaks and wrapping
Contrary to the rest of the fish source code, the documentation greatly benefits from the use of long lines and soft wrapping. It allows paragraphs to be treated as complete blocks by Doxygen, means that the semantic filter can see complete lines when deciding on how to apply syntax highlighting, and means that man pages will consistently wrap to the width of the users console in advanced pagers, such as 'most'.
## Doxygen special commands and aliases
While Markdown syntax forms the basis of the documentation content, there are some exceptions that require the use of Doxygen special commands. On the whole, Doxygen commands should be avoided, especially inline word formatting such as \\c as this would allow Doxygen to make unhelpful assumptions, such as converting double dashes (\--) to n-dashes (–).
### Structure: \\page, \\section and \\subsection
Use of Doxygen sections markers are important, as these determine what will be eventually output as a web page, man page or included in the developer docs.
Currently the make process for the documentation is quite convoluted, but basically the HTML docs are produced from a single, compiled file, doc.h. This contains a number of \\page markers that produce the various pages used in the documentation. The format of a \\page mark is:
\page universally_unique_page_id Page title
The source files that contain the page markers are currently:
- __index.hdr.in__: Core documentation
- __commands.hdr.in__: Individual commands
- __tutorial.hdr__: Tutorial
- __design.hdr__: Design document
- __faq.hdr__: Frequently Asked Questions
- __license.hdr__: Fish and 3rd party licences
Unless there is a _VERY_ good reason and developer consensus, new pages should never be added.
The rest of the documentation is structured using \\section and \\subsection markers. Most of the source files (listed above) contain their full content, the exception being commands, which are separated out into source text files in the doc_src directory. These files are concatenated into one file, so each one starts with a \\section declaration. The synopsis, description and examples (if present) are declared as \\subsections. The format of these marks is practically identical to the page mark.
\section universally_unique_section_id Section title
\subsection universally_unique_subsection_id Subsection title
Each page, section and subsection id _must_ be unique across the whole of the documentation, otherwise Doxygen will issue a warning.
### Semantic markup: the \\fish .. \\endfish block
While Doxygen has support for \\code..\\endcode blocks with enhanced markup and syntax colouring, it only understands the core Doxygen languages: C, C++, Objective C, Java, PHP, Python, Tcl and Fortran. To enhance Fish's syntax presentation, use the special \\fish..\\endfish blocks instead.
Text placed in this block will be parsed by Doxygen using the included lexicon filter (see lexicon_filter.in) as a Doxygen input filter. The filter is built during make so that it can pick up information on builtins, functions and shell commands mentioned in completions and apply markup to keywords found inside the \\fish block.
Basically, preformatted plain text inside the \\fish block is fed through the filter and is returned marked up so that Doxygen aliases can convert it back to a presentable form, according to the output document type.
For instance:
`echo hello world`
is transformed into:
`@cmnd{echo} @args{hello} @args{world}`
which is then transformed by Doxygen into an HTML version (`make doc`):
In older browsers, it was easy to set the fonts used for the three basic type styles (serif, sans-serif and monospace). Modern browsers have removed these options in their respective quests for simplification, assuming the content author will provide suitable styles for the content in the site's CSS, or the end user will provide overriding styles manually. Doxygen's default styling is very simple and most users will just accept this default.
I've tried to use a sensible set of fonts in the documentation's CSS based on 'good' terminal fonts and as a result the firt preference font used throughout the documentation is '[DejaVu](http://dejavu-fonts.org)'. The rationale behaind this is that while DejaVu is getting a little long in the tooth, it still provides the most complete support across serif, sans-serif and monospace styles (giving a well balanced feel and consistent [x-height](http://en.wikipedia.org/wiki/X-height)), has the widest support for extended Unicode characters and has a free, permissive licenses (though it's still incompatible with GPLv2, though arguably less so than the SIL Open Font license, though this is a moot point when using it solely in the docs).
#### Fonts inside \\fish blocks and \`backticks\`
As the point of these contructs is to make fish's syntax clearer to the user, it makes sense to mimic what the user will see in the console, therefore any content is formatted using the monospaced style, specifically monospaced fonts are chosen in the following order:
1.__DejaVu Sans Mono__: Explained above. [[↓](http://dejavu-fonts.org)]
2.__Source Code Pro__: Monospaced code font, part of Adobe's free Edge Web Fonts. [[↓](https://edgewebfonts.adobe.com)]
5.__Consolas__: Modern Microsoft supplied console font.
6.__Monaco__: Apple supplied console font since 1984!
7.__Lucida Console__: Generic mono terminal font, standard in many OS's and distros.
8.__monospace__: Catchall style. Chooses default monospaced font, often Courier.
9.__fixed__: As above, more often used on mobile devices.
#### General Fonts
1.__DejaVu Sans__: As above.[[↓](http://dejavu-fonts.org)]
2.__Roboto__: Elegant Google free font and is Doxygen's default [[↓](http://www.google.com/fonts/specimen/Roboto)]
3.__Lucida Grande__: Default Apple OS X content font.
4.__Calibri__: Default Microsoft Office font (since 2007).
5.__Verdana__: Good general font found in a lot of OSs.
6.__Helvetica Neue__: Better spaced and balanced Helvetica/Arial variant.
7.__Helvetica__: Standard humanist typeface found almost everywhere.
8.__Arial__: Microsoft's Helvetica.
9.__sans-serif__: Catchall style. Chooses default sans-serif typeface, often Helvetica.
The ordering of the fonts is important as it's designed to allow the documentation to settle into a number of different identities according to the fonts available. If you have the complete DejaVu family installed, then the docs are presented using that, and if your Console is set up to use the same fonts, presentation will be completely consistent.
On OS X, with nothing extra installed, the docs will default to Menlo and Lucida Grande giving a Mac feel. Under Windows, it will default to using Consolas and Calibri on recent versions, giving a modern Windows style.
#### Other sources:
- [Font Squirrel](http://www.fontsquirrel.com): Good source of open source font packages.
### Choosing a CLI style: using a \\fish{style} block
By default, when output as HTML, a \\fish block uses syntax colouring suited to the style of the documentation rather than trying to mimic the terminal. The block has a light, bordered background and a colour scheme that 'suggests' what the user would see in a console.
Additional stying can be applied adding a style declaration:
\fish{additional_style [another_style...]}
...
\endfish
This will translate to classes applied to the `<div>` tag, like so:
<div class="fish additional_style another_style">
...
</div>
The various classes are defined in `doc_src/user_doc.css` and new style can be simply added
The documentation currently defines a couple of additional styles:
- __cli-dark__: Used in the _tutorial_ and _FAQ_ to simulate a dark background terminal, with fish's default colours (slightly tweaked for legibility in the browser).
- __synopsis__: A simple colour theme helpful for displaying the logical 'summary' of a command's syntax, options and structure.
## Markdown
Apart from the exceptions discussed above, the rest of the documentation now supports the use of Markdown. As such the use of Doxygen special commands for HTML tags is unnecessary.
There are a few exceptions and extensions to the Markdown [standard](http://daringfireball.net/projects/markdown/) that are documented in the Doxygen [documentation](http://www.stack.nl/~dimitri/doxygen/manual/markdown.html).
### \`Backticks\`
As is standard in Markdown and 'Github Flavoured Markdown' (GFM), backticks can be used to denote inline technical terms in the documentation, `like so`. In the documentation this will set the font to the monospaced 'console' typeface and will cause the enclosed term to stand out.
However, fenced code blocks using 4 spaces or 3 backticks (\`\`\`) should be avoided as Doxygen will interpret these as \\code blocks and try to apply standard syntax colouring, which doesn't work so well for fish examples. Use `\fish..\endfish` blocks instead.
### Lists
Standard Markdown list rules apply, but as Doxygen will collapse white space on output, combined with the use of long lines, it's a good idea to include an extra new line between long list items to assist future editing.
## Special cases
The following can be used in \\fish blocks to render some fish scenarios. These are mostly used in the tutorial when an interactive situation needs to be displayed.
### Custom formatting tags
```html
<u>: <u>These words are underlined.</u>
<s>: auto<s>suggestion</s>.
<m>: <m>Matched</m> items, such as tab completions.
<sm>: Matched items <sm>searched</sm> for, like grep results.
<bs>: Render the contents with a preceding backslash. Useful when presenting output.
<eror>: <eror>This would be shown as an error. (Note eror, not error).</eror>
<asis>: <asis>This text will not be parsed for fish markup.</asis>
<outp>: <outp>This would be rendered as command/script output.</outp>
{{ and }}: Required when wanting curly braces in regular expression example.
```
### Prompts and cursors
```html
>_: Display a basic prompt.
~>_: Display a prompt with a the home directory as the current working directory.
___ (3 underscores): Display a cursor.
```
### Keyboard shortcuts: @key{} and @cursor_key{}
Graphical keyboard shortcuts can be defined using the following special commands. These allow for the different text requirements across the html and man pages. The HTML uses CSS to create a keyboard style, whereas the man page would display the key as text.
-`@key{lable}`
Displays a key with a purely textual lable, such as: 'Tab', 'Page Up', 'Page Down', 'Home', 'End', 'F1', 'F19' and so on.
-`@key{modifier,lable}`
Displays a keystroke requiring the use of a 'modifier' key, such as 'Control-A', 'Shift-X', 'Alt-Tab' etc.
-`@key{modifier,entity,lable}`
Displays a keystroke using a graphical entity, such as an arrow symbol for cursor key based shortcuts.
-`@cursor_key{entity,lable}`
A special case for cursor keys, when no modifier is needed. i.e. `@cursor_key{↑,up}` for the up arrow key.
Some useful Unicode/HTML5 entities:
- Up arrow: `↑`
- Down arrow: `↓`
- Left arrow: `←`
- Right arrow `→`
- Shift: `⇧`
- Tab: `⇥`
- Mac option: `⌥`
- Mac command: `⌘`
## Notes
### Doxygen
Tested on:
- Ubuntu 14.04 with Doxygen 1.8.8, built from [GitHub source](https://github.com/doxygen/doxygen.git).
- CentOS 6.5 with Doxygen 1.8.8, built from [GitHub source](https://github.com/doxygen/doxygen.git).
- Mac OS X 10.9 with Homebrew install Doxygen 1.8.7 and 1.8.8.
Graphviz was also installed in all the above testing.
Doxygen 1.8.6 and lower do not have the \\htmlonly[block] directive which fixes a multitude of problems in the rendering of the docs. In Doxygen 1.8.7 the list of understood HTML entities was greatly increased. I tested earlier versions and many little issues returned.
As fish ships with pre-built documentation, I don't see this as an issue.
### Updated Configure/Makefile
- Tested on Ubuntu 14.04, CentOS 6.5 and Mac OS X 10.9.
- Makefile has GNU/BSD sed/grep detection.
### HTML output
- The output HTML is HTML5 compliant, but should quickly and elegantly degrade on older browsers without losing basic structure.
- The CSS avoids the use or browser specific extenstions (i.e. -webkit, -moz etc), using the W3C HTML5 standard instead.
- It's been tested in Chrome 37.0 and Firefox 32.0 on Mac OS X 10.9 (+Safari 7), Windows 8.1 (+Internet Explorer 11) and Ubuntu Desktop 14.04.
- My assumption is basically that if someone cares enough to want to install fish, they'll be keeping a browser current.
### Man page output
- Tested on Ubuntu 14.04, CentOS 6.5 and Mac OS X 10.9.
- Output is substantially cleaner.
- Tested in cat, less, more and most pagers using the following fish script:
```
function manTest --description 'Test manpage' --argument page
set -l pager
for i in $argv
switch $i
case "-l"
set pager -P '/usr/bin/less -is'
case "-m"
set pager -P '/usr/bin/more -s'
case "-c"
set pager -P '/bin/cat'
end
end
man $pager ~/Projects/OpenSource/fish-shell/share/man/man1/$page.1
end
# Assumes 'most' is the default system pager.
# NOT PORTABLE! Paths would be need to be updated on other systems.
```
#### Author: Mark Griffiths [@GitHub](https://github.com/MarkGriffiths)
`abbr` manipulates the list of abbreviations that fish will expand.
Abbreviations are user-defined character sequences or words that are replaced with longer phrases after they are entered. For example, a frequently-run command such as `git checkout` can be abbreviated to `gco`. After entering `gco` and pressing @key{Space} or @key{Enter}, the full text `git checkout` will appear in the command line.
Abbreviations are stored in a variable named `fish_user_abbreviations`. This is automatically created as a universal variable the first time an abbreviation is created. If you want your abbreviations to be private to a particular fish session you can put the following in your *~/.config/fish/config.fish* file before you define your first abbrevation:
\fish
if status --is-interactive
set -g fish_user_abbreviations
abbr --add first 'echo my first abbreviation'
abbr --add second 'echo my second abbreviation'
# etcetera
end
\endfish
You can create abbreviations directly on the command line and they will be saved automatically and made visible to other fish sessions if `fish_user_abbreviations` is a universal variable. If you keep the variable as universal, `abbr --add` statements in <a href="tutorial.html#tut_startup">config.fish</a> will do nothing but slow down startup slightly.
\subsection abbr-options Options
The following parameters are available:
- `-a WORD PHRASE` or `--add WORD PHRASE` Adds a new abbreviation, causing WORD to be expanded to PHRASE.
- `-r WORD NEW_WORD` or `--rename WORD NEW_WORD` Renames an abbreviation, from WORD to NEW_WORD.
- `-s` or `--show` Show all abbreviated words and their expanded phrases in a manner suitable for export and import.
- `-l` or `--list` Lists all abbreviated words.
- `-e WORD` or `--erase WORD` Erase the abbreviation WORD.
Note: fish version 2.1 supported `-a WORD=PHRASE`. This syntax is now deprecated but will still be converted.
\subsection abbr-example Examples
\fish
abbr -a gco git checkout
\endfish
Add a new abbreviation where `gco` will be replaced with `git checkout`.
\fish
abbr -r gco gch
\endfish
Renames an existing abbreviation from `gco` to `gch`.
\fish
abbr -e gco
\endfish
Erase the `gco` abbreviation.
\fish
ssh another_host abbr -s | source
\endfish
Import the abbreviations defined on another_host over SSH.
`alias` is a simple wrapper for the `function` builtin, which creates a function wrapping a command. It has similar syntax to POSIX shell `alias`. For other uses, it is recommended to define a <a href='#function'>function</a>.
`fish` marks functions that have been created by `alias` by including the command used to create them in the function description. You can list `alias`-created functions by running `alias` without arguments. They must be erased using `functions -e`.
- `NAME` is the name of the alias
- `DEFINITION` is the actual command to execute. The string `$argv` will be appended.
You cannot create an alias to a function with the same name. Note that spaces need to be escaped in the call to `alias` just like at the command line, _even inside quoted parts_.
\subsection alias-example Example
The following code will create `rmi`, which runs `rm` with additional arguments on every invocation.
\fish
alias rmi="rm -i"
# This is equivalent to entering the following function:
function rmi --wraps rm --description 'alias rmi=rm -i'
rm -i $argv
end
# This needs to have the spaces escaped or "Chrome.app..." will be seen as an argument to "/Applications/Google":
alias chrome='/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome banana'
\section and and - conditionally execute a command
\subsection and-synopsis Synopsis
\fish{synopsis}
COMMAND1; and COMMAND2
\endfish
\subsection and-description Description
`and` is used to execute a command if the current exit status (as set by the previous command) is 0.
`and` statements may be used as part of the condition in an <a href="#if">`if`</a> or <a href="#while">`while`</a> block. See the documentation for <a href="#if">`if`</a> and <a href="#while">`while`</a> for examples.
`and` does not change the current exit status. The exit status of the last foreground command to exit can always be accessed using the <a href="index.html#variables-status">$status</a> variable.
\subsection and-example Example
The following code runs the `make` command to build a program. If the build succeeds, `make`'s exit status is 0, and the program is installed. If either step fails, the exit status is 1, and `make clean` is run, which removes the files created by the build process.
The block is unconditionally executed. `begin; ...; end` is equivalent to `if true; ...; end`.
`begin` is used to group a number of commands into a block. This allows the introduction of a new variable scope, redirection of the input or output of a set of commands as a group, or to specify precedence when using the conditional commands like `and`.
`begin` does not change the current exit status.
\subsection begin-example Example
The following code sets a number of variables inside of a block scope. Since the variables are set inside the block and have local scope, they will be automatically deleted when the block ends.
\fish
begin
set -l PIRATE Yarrr
...
end
echo $PIRATE
# This will not output anything, since the PIRATE variable
# went out of scope at the end of the block
\endfish
In the following code, all output is redirected to the file out.html.
`bg` sends <a href="index.html#syntax-job-control">jobs</a> to the background, resuming them if they are stopped. A background job is executed simultaneously with fish, and does not have access to the keyboard. If no job is specified, the last job to be used is put in the background. If PID is specified, the jobs with the specified process group IDs are put in the background.
The PID of the desired process is usually found by using <a href="index.html#expand-process">process expansion</a>.
When at least one of the arguments isn't a valid job specifier (i.e. PID),
`bg` will print an error without backgrounding anything.
When all arguments are valid job specifiers, bg will background all matching jobs that exist.
\subsection bg-example Example
`bg %1` will put the job with job ID 1 in the background.
`bg 123 456 789` will background 123, 456 and 789.
If only 123 and 789 exist, it will still background them and print an error about 456.
`bg 123 banana` or `bg banana 123` will complain that "banana" is not a valid job specifier.
`bind` adds a binding for the specified key sequence to the specified command.
SEQUENCE is the character sequence to bind to. These should be written as <a href="index.html#escapes">fish escape sequences</a>. For example, because pressing the Alt key and another character sends that character prefixed with an escape character, Alt-based key bindings can be written using the `\e` escape. For example, @key{Alt,w} can be written as `\ew`. The control character can be written in much the same way using the `\c` escape, for example @key{Control,X} (^X) can be written as `\cx`. Note that Alt-based key bindings are case sensitive and Control-based key bindings are not. This is a constraint of text-based terminals, not `fish`.
The default key binding can be set by specifying a `SEQUENCE` of the empty string (that is, ```''``` ). It will be used whenever no other binding matches. For most key bindings, it makes sense to use the `self-insert` function (i.e. ```bind '' self-insert```) as the default keybinding. This will insert any keystrokes not specifically bound to into the editor. Non- printable characters are ignored by the editor, so this will not result in control sequences being printable.
If the `-k` switch is used, the name of the key (such as 'down', 'up' or 'backspace') is used instead of a sequence. The names used are the same as the corresponding curses variables, but without the 'key_' prefix. (See `terminfo(5)` for more information, or use `bind --key-names` for a list of all available named keys.)
`COMMAND` can be any fish command, but it can also be one of a set of special input functions. These include functions for moving the cursor, operating on the kill-ring, performing tab completion, etc. Use `bind --function-names` for a complete list of these input functions.
When `COMMAND` is a shellscript command, it is a good practice to put the actual code into a <a href="#function">function</a> and simply bind to the function name. This way it becomes significantly easier to test the function while editing, and the result is usually more readable as well.
If such a script produces output, the script needs to finish by calling `commandline -f repaint` in order to tell fish that a repaint is in order.
When multiple `COMMAND`s are provided, they are all run in the specified order when the key is pressed.
If no `SEQUENCE` is provided, all bindings (or just the bindings in the specified `MODE`) are printed. If `SEQUENCE` is provided without `COMMAND`, just the binding matching that sequence is printed.
Key bindings are not saved between sessions by default. **Bare `bind` statements in <a href="index.html#initialization">config.fish</a> won't have any effect because it is sourced before the default keybindings are setup.** To save custom keybindings, put the `bind` statements into a function called `fish_user_key_bindings`, which will be <a href="tutorial.html#tut_autoload">autoloaded</a>.
Key bindings may use "modes", which mimics Vi's modal input behavior. The default mode is "default", and every bind applies to a single mode. The mode can be viewed/changed with the `$fish_bind_mode` variable.
The following parameters are available:
- `-k` or `--key` Specify a key name, such as 'left' or 'backspace' instead of a character sequence
- `-K` or `--key-names` Display a list of available key names. Specifying `-a` or `--all` includes keys that don't have a known mapping
- `-f` or `--function-names` Display a list of available input functions
- `-L` or `--list-modes` Display a list of defined bind modes
- `-M MODE` or `--mode MODE` Specify a bind mode that the bind is used in. Defaults to "default"
- `-m NEW_MODE` or `--sets-mode NEW_MODE` Change the current mode to `NEW_MODE` after this binding is executed
- `-e` or `--erase` Erase the binding with the given sequence and mode instead of defining a new one. Multiple sequences can be specified with this flag. Specifying `-a` or `--all` with `-M` or `--mode` erases all binds in the given mode regardless of sequence. Specifying `-a` or `--all` without `-M` or `--mode` erases all binds in all modes regardless of sequence.
- `-a` or `--all` See `--erase` and `--key-names`
The following special input functions are available:
- `accept-autosuggestion`, accept the current autosuggestion completely
- `backward-char`, moves one character to the left
- `backward-bigword`, move one whitespace-delimited word to the left
- `backward-delete-char`, deletes one character of input to the left of the cursor
- `backward-kill-bigword`, move the whitespace-delimited word to the left of the cursor to the killring
- `backward-kill-line`, move everything from the beginning of the line to the cursor to the killring
- `backward-kill-path-component`, move one path component to the left of the cursor (everything from the last "/" or whitespace exclusive) to the killring
- `backward-kill-word`, move the word to the left of the cursor to the killring
- `backward-word`, move one word to the left
- `beginning-of-buffer`, moves to the beginning of the buffer, i.e. the start of the first line
- `beginning-of-history`, move to the beginning of the history
- `beginning-of-line`, move to the beginning of the line
- `begin-selection`, start selecting text
- `capitalize-word`, make the current word begin with a capital letter
- `complete`, guess the remainder of the current token
- `complete-and-search`, invoke the searchable pager on completion options (for convenience, this also moves backwards in the completion pager)
- `delete-char`, delete one character to the right of the cursor
- `downcase-word`, make the current word lowercase
- `end-of-buffer`, moves to the end of the buffer, i.e. the end of the first line
- `end-of-history`, move to the end of the history
- `end-of-line`, move to the end of the line
- `end-selection`, end selecting text
- `forward-bigword`, move one whitespace-delimited word to the right
- `forward-char`, move one character to the right
- `forward-word`, move one word to the right
- `history-search-backward`, search the history for the previous match
- `history-search-forward`, search the history for the next match
- `kill-bigword`, move the next whitespace-delimited word to the killring
- `kill-line`, move everything from the cursor to the end of the line to the killring
- `kill-selection`, move the selected text to the killring
- `kill-whole-line`, move the line to the killring
- `kill-word`, move the next word to the killring
- `suppress-autosuggestion`, remove the current autosuggestion
- `swap-selection-start-stop`, go to the other end of the highlighted text without changing the selection
- `transpose-chars`, transpose two characters to the left of the cursor
- `transpose-words`, transpose two words to the left of the cursor
- `upcase-word`, make the current word uppercase
- `yank`, insert the latest entry of the killring into the buffer
- `yank-pop`, rotate to the previous entry of the killring
\subsection bind-example Examples
\fish
bind <asis>\\cd</asis> 'exit'
\endfish
Causes `fish` to exit when @key{Control,D} is pressed.
\fish
bind -k ppage history-search-backward
\endfish
Performs a history search when the @key{Page Up} key is pressed.
\fish
set -g fish_key_bindings fish_vi_key_bindings
bind -M insert \\cc kill-whole-line force-repaint
\endfish
Turns on Vi key bindings and rebinds @key{Control,C} to clear the input line.
\subsection special-case-escape Special Case: The escape Character
The escape key can be used standalone, for example, to switch from insertion mode to normal mode when using Vi keybindings. Escape may also be used as a "meta" key, to indicate the start of an escape sequence, such as function or arrow keys. Custom bindings can also be defined that begin with an escape character.
fish waits for a period after receiving the escape character, to determine whether it is standalone or part of an escape sequence. While waiting, additional key presses make the escape key behave as a meta key. If no other key presses come in, it is handled as a standalone escape. The waiting period is set to 300 milliseconds (0.3 seconds) in the default key bindings and 10 milliseconds in the vi key bindings. It can be configured by setting the `fish_escape_delay_ms` variable to a value between 10 and 5000 ms. It is recommended that this be a universal variable that you set once from an interactive session.
Note: fish 2.2.0 and earlier used a default of 10 milliseconds, and provided no way to configure it. That effectively made it impossible to use escape as a meta key.
\section block block - temporarily block delivery of events
\subsection block-synopsis Synopsis
\fish{synopsis}
block [OPTIONS...]
\endfish
\subsection block-description Description
`block` prevents events triggered by `fish` or the <a href="commands.html#emit">`emit`</a> command from being delivered and acted upon while the block is in place.
In functions, `block` can be useful while performing work that should not be interrupted by the shell.
The block can be removed. Any events which triggered while the block was in place will then be delivered.
Event blocks should not be confused with code blocks, which are created with `begin`, `if`, `while` or `for`
The following parameters are available:
- `-l` or `--local` Release the block automatically at the end of the current innermost code block scope
- `-g` or `--global` Never automatically release the lock
- `-e` or `--erase` Release global block
\subsection block-example Example
\fish
# Create a function that listens for events
function --on-event foo foo; echo 'foo fired'; end
# Block the delivery of events
block -g
emit foo
# No output will be produced
block -e
# 'foo fired' will now be printed
\endfish
\subsection notes Notes
Note that events are only received from the current fish process as there is no way to send events from one fish process to another.
\section break break - stop the current inner loop
\subsection break-synopsis Synopsis
\fish{synopsis}
LOOP_CONSTRUCT; [COMMANDS...] break; [COMMANDS...] end
\endfish
\subsection break-description Description
`break` halts a currently running loop, such as a <a href="#for">for</a> loop or a <a href="#while">while</a> loop. It is usually added inside of a conditional block such as an <a href="#if">if</a> statement or a <a href="#switch">switch</a> statement.
There are no parameters for `break`.
\subsection break-example Example
The following code searches all .c files for "smurf", and halts at the first occurrence.
\section case case - conditionally execute a block of commands
\subsection case-synopsis Synopsis
\fish{synopsis}
switch VALUE; [case [WILDCARD...]; [COMMANDS...]; ...] end
\endfish
\subsection case-description Description
`switch` performs one of several blocks of commands, depending on whether a specified value equals one of several wildcarded values. `case` is used together with the `switch` statement in order to determine which block should be executed.
Each `case` command is given one or more parameters. The first `case` command with a parameter that matches the string specified in the switch command will be evaluated. `case` parameters may contain wildcards. These need to be escaped or quoted in order to avoid regular wildcard expansion using filenames.
Note that fish does not fall through on case statements. Only the first matching case is executed.
Note that command substitutions in a case statement will be evaluated even if its body is not taken. All substitutions, including command substitutions, must be performed before the value can be compared against the parameter.
\subsection case-example Example
If the variable \$animal contains the name of an animal, the following
code would attempt to classify it:
\fish
switch $animal
case cat
echo evil
case wolf dog human moose dolphin whale
echo mammal
case duck goose albatross
echo bird
case shark trout stingray
echo fish
# Note that the next case has a wildcard which is quoted
case '*'
echo I have no idea what a $animal is
end
\endfish
If the above code was run with `$animal` set to `whale`, the output
If `DIRECTORY` is supplied, it will become the new directory. If no parameter is given, the contents of the `HOME` environment variable will be used.
If `DIRECTORY` is a relative path, the paths found in the `CDPATH` environment variable array will be tried as prefixes for the specified path.
Note that the shell will attempt to change directory without requiring `cd` if the name of a directory is provided (starting with `.`, `/` or `~`, or ending with `/`).
Fish also ships a wrapper function around the builtin `cd` that understands `cd -` as changing to the previous directory. See also <a href="commands.html#prevd">`prevd`</a>. This wrapper function maintains a history of the 25 most recently visited directories in the `$dirprev` and `$dirnext` global variables.
\subsection cd-example Examples
\fish
cd
# changes the working directory to your home directory.
cd /usr/src/fish-shell
# changes the working directory to /usr/src/fish-shell
`command` forces the shell to execute the program `COMMANDNAME` and ignore any functions or builtins with the same name.
The following options are available:
- `-s` or `--search` returns the name of the disk file that would be executed, or nothing if no file with the specified name could be found in the `$PATH`.
With the `-s` option, `command` treats every argument as a separate command to look up and sets the exit status to 0 if any of the specified commands were found, or 1 if no commands could be found. Additionally passing a `-q` or `--quiet` option prevents any paths from being printed, like the `type -q`, for testing only the exit status.
For basic compatibility with POSIX `command`, the `-v` flag is recognized as an alias for `-s`.
\subsection command-example Examples
`command ls` causes fish to execute the `ls` program, even if an `ls` function exists.
`command -s ls` returns the path to the `ls` program.
\section commandline commandline - set or get the current command line buffer
\subsection commandline-synopsis Synopsis
\fish{synopsis}
commandline [OPTIONS] [CMD]
\endfish
\subsection commandline-description Description
`commandline` can be used to set or get the current contents of the command line buffer.
With no parameters, `commandline` returns the current value of the command line.
With `CMD` specified, the command line buffer is erased and replaced with the contents of `CMD`.
The following options are available:
- `-C` or `--cursor` set or get the current cursor position, not the contents of the buffer. If no argument is given, the current cursor position is printed, otherwise the argument is interpreted as the new cursor position.
- `-f` or `--function` inject readline functions into the reader. This option cannot be combined with any other option. It will cause any additional arguments to be interpreted as readline functions, and these functions will be injected into the reader, so that they will be returned to the reader before any additional actual key presses are read.
The following options change the way `commandline` updates the command line buffer:
- `-a` or `--append` do not remove the current commandline, append the specified string at the end of it
- `-i` or `--insert` do not remove the current commandline, insert the specified string at the current cursor position
- `-r` or `--replace` remove the current commandline and replace it with the specified string (default)
The following options change what part of the commandline is printed or updated:
- `-b` or `--current-buffer` select the entire buffer (default)
- `-j` or `--current-job` select the current job
- `-p` or `--current-process` select the current process
- `-t` or `--current-token` select the current token.
The following options change the way `commandline` prints the current commandline buffer:
- `-c` or `--cut-at-cursor` only print selection up until the current cursor position
- `-o` or `--tokenize` tokenize the selection and print one string-type token per line
If `commandline` is called during a call to complete a given string using `complete -C STRING`, `commandline` will consider the specified string to be the current contents of the command line.
The following options output metadata about the commandline state:
- `-L` or `--line` print the line that the cursor is on, with the topmost line starting at 1
- `-S` or `--search-mode` evaluates to true if the commandline is performing a history search
- `-P` or `--paging-mode` evaluates to true if the commandline is showing pager contents, such as tab completions
\subsection commandline-example Example
`commandline -j $history[3]` replaces the job under the cursor with the third item from the command line history.
If the commandline contains
\fish
>_ echo $fl___ounder >&2 | less; and echo $catfish
`fish` ships with a large number of builtin commands, shellscript functions and external commands. These are all described below.
Almost all fish commands respond to the `-h` or `--help` options to display their relevant help, also accessible using the `help` and `man` commands, like so:
\fish
echo -h
echo --help
# Prints help to the terminal window
man echo
# Displays the man page in the system pager
# (normally 'less', 'more' or 'most').
help echo
# Open a web browser to show the relevant documentation
For an introduction to specifying completions, see <a
href='index.html#completion-own'>Writing your own completions</a> in
the fish manual.
- `COMMAND` is the name of the command for which to add a completion.
- `SHORT_OPTION` is a one character option for the command.
- `LONG_OPTION` is a multi character option for the command.
- `OPTION_ARGUMENTS` is parameter containing a space-separated list of possible option-arguments, which may contain command substitutions.
- `DESCRIPTION` is a description of what the option and/or option arguments do.
- `-c COMMAND` or `--command COMMAND` specifies that `COMMAND` is the name of the command.
- `-p COMMAND` or `--path COMMAND` specifies that `COMMAND` is the absolute path of the program (optionally containing wildcards).
- `-e` or `--erase` deletes the specified completion.
- `-s SHORT_OPTION` or `--short-option=SHORT_OPTION` adds a short option to the completions list.
- `-l LONG_OPTION` or `--long-option=LONG_OPTION` adds a GNU style long option to the completions list.
- `-o LONG_OPTION` or `--old-option=LONG_OPTION` adds an old style long option to the completions list (See below for details).
- `-a OPTION_ARGUMENTS` or `--arguments=OPTION_ARGUMENTS` adds the specified option arguments to the completions list.
- `-f` or `--no-files` specifies that the options specified by this completion may not be followed by a filename.
- `-r` or `--require-parameter` specifies that the options specified by this completion always must have an option argument, i.e. may not be followed by another option.
- `-x` or `--exclusive` implies both `-r` and `-f`.
- `-w WRAPPED_COMMAND` or `--wraps=WRAPPED_COMMAND` causes the specified command to inherit completions from the wrapped command (See below for details).
- `-n` or `--condition` specifies a shell command that must return 0 if the completion is to be used. This makes it possible to specify completions that should only be used in some cases.
- `-CSTRING` or `--do-complete=STRING` makes complete try to find all possible completions for the specified string.
- `-C` or `--do-complete` with no argument makes complete try to find all possible completions for the current command line buffer. If the shell is not in interactive mode, an error is returned.
- `-A` and `--authoritative` no longer do anything and are silently ignored.
- `-u` and `--unauthoritative` no longer do anything and are silently ignored.
Command specific tab-completions in `fish` are based on the notion of options and arguments. An option is a parameter which begins with a hyphen, such as '`-h`', '`-help`' or '`--help`'. Arguments are parameters that do not begin with a hyphen. Fish recognizes three styles of options, the same styles as the GNU version of the getopt library. These styles are:
- Short options, like '`-a`'. Short options are a single character long, are preceded by a single hyphen and may be grouped together (like '`-la`', which is equivalent to '`-l -a`'). Option arguments may be specified in the following parameter ('`-w 32`') or by appending the option with the value ('`-w32`').
- Old style long options, like '`-Wall`'. Old style long options can be more than one character long, are preceded by a single hyphen and may not be grouped together. Option arguments are specified in the following parameter ('`-ao null`').
- GNU style long options, like '`--colors`'. GNU style long options can be more than one character long, are preceded by two hyphens, and may not be grouped together. Option arguments may be specified in the following parameter ('`--quoting-style shell`') or by appending the option with a '`=`' and the value ('`--quoting-style=shell`'). GNU style long options may be abbreviated so long as the abbreviation is unique ('`--h`') is equivalent to '`--help`' if help is the only long option beginning with an 'h').
The options for specifying command name and command path may be used multiple times to define the same completions for multiple commands.
The options for specifying command switches and wrapped commands may be used multiple times to define multiple completions for the command(s) in a single call.
Invoking `complete` multiple times for the same command adds the new definitions on top of any existing completions defined for the command.
When `-a` or `--arguments` is specified in conjunction with long, short, or old style options, the specified arguments are only used as completions when attempting to complete an argument for any of the specified options. If `-a` or `--arguments` is specified without any long, short, or old style options, the specified arguments are used when completing any argument to the command (except when completing an option argument that was specified with `-r` or `--require-parameter`).
Command substitutions found in `OPTION_ARGUMENTS` are not expected to return a space-separated list of arguments. Instead they must return a newline-separated list of arguments, and each argument may optionally have a tab character followed by the argument description. Any description provided in this way overrides a description given with `-d` or `--description`.
The `-w` or `--wraps` options causes the specified command to inherit completions from another command. The inheriting command is said to "wrap" the inherited command. The wrapping command may have its own completions in addition to inherited ones. A command may wrap multiple commands, and wrapping is transitive: if A wraps B, and B wraps C, then A automatically inherits all of C's completions. Wrapping can be removed using the `-e` or `--erase` options. Note that wrapping only works for completions specified with `-c` or `--command` and are ignored when specifying completions with `-p` or `--path`.
When erasing completions, it is possible to either erase all completions for a specific command by specifying `complete -c COMMAND -e`, or by specifying a specific completion option to delete by specifying either a long, short or old style option.
\subsection complete-example Example
The short style option `-o` for the `gcc` command requires that a file follows it. This can be done using writing:
\fish
complete -c gcc -s o -r
\endfish
The short style option `-d` for the `grep` command requires that one of the strings '`read`', '`skip`' or '`recurse`' is used. This can be specified writing:
\fish
complete -c grep -s d -x -a "read skip recurse"
\endfish
The `su` command takes any username as an argument. Usernames are given as the first colon-separated field in the file /etc/passwd. This can be specified as:
\fish
complete -x -c su -d "Username" -a "(cat /etc/passwd | cut -d : -f 1)"
\endfish
The `rpm` command has several different modes. If the `-e` or `--erase` flag has been specified, `rpm` should delete one or more packages, in which case several switches related to deleting packages are valid, like the `nodeps` switch.
\section contains contains - test if a word is present in a list
\subsection contains-synopsis Synopsis
\fish{synopsis}
contains [OPTIONS] KEY [VALUES...]
\endfish
\subsection contains-description Description
`contains` tests whether the set `VALUES` contains the string `KEY`. If so, `contains` exits with status 0; if not, it exits with status 1.
The following options are available:
- `-i` or `--index` print the word index
Note that, like GNU tools, `contains` interprets all arguments starting with a `-` as options to contains, until it reaches an argument that is `--` (two dashes). See the examples below.
\subsection contains-example Example
\fish
for i in ~/bin /usr/local/bin
if not contains $i $PATH
set PATH $PATH $i
end
end
\endfish
The above code tests if `~/bin` and `/usr/local/bin` are in the path and adds them if not.
\fish
function hasargs
if contains -- -q $argv
echo '$argv contains a -q option'
end
end
\endfish
The above code checks for `-q` in the argument list, using the `--` argument to demarcate options to `contains` from the key to search for.
\section continue continue - skip the remainder of the current iteration of the current inner loop
\subsection continue-synopsis Synopsis
\fish{synopsis}
LOOP_CONSTRUCT; [COMMANDS...;] continue; [COMMANDS...;] end
\endfish
\subsection continue-description Description
`continue` skips the remainder of the current iteration of the current inner loop, such as a <a href="#for">for</a> loop or a <a href="#while">while</a> loop. It is usually added inside of a conditional block such as an <a href="#if">if</a> statement or a <a href="#switch">switch</a> statement.
\subsection continue-example Example
The following code removes all tmp files that do not contain the word smurf.
\section count count - count the number of elements of an array
\subsection count-synopsis Synopsis
\fish{synopsis}
count $VARIABLE
\endfish
\subsection count-description Description
`count` prints the number of arguments that were passed to it. This is usually used to find out how many elements an environment variable array contains.
`count` does not accept any options, including `-h` or `--help`.
`count` exits with a non-zero exit status if no arguments were passed to it, and with zero if at least one argument was passed.
\subsection count-example Example
\fish
count $PATH
# Returns the number of directories in the users PATH variable.
count *.txt
# Returns the number of files in the current working directory ending with the suffix '.txt'.
`dirh` prints the current directory history. The current position in the history is highlighted using the color defined in the `fish_color_history_current` environment variable.
`dirh` does not accept any parameters.
Note that the `cd` command limits directory history to the 25 most recently visited directories. The history is stored in the `$dirprev` and `$dirnext` variables.
\section disown disown - remove a process from the list of jobs
\subsection disown-synopsis Synopsis
\fish{synopsis}
disown [ PID ... ]
\endfish
\subsection disown-description Description
`disown` removes the specified <a href="index.html#syntax-job-control">job</a> from the list of jobs. The job itself continues to exist, but fish does not keep track of it any longer.
Jobs in the list of jobs are sent a hang-up signal when fish terminates, which usually causes the job to terminate; `disown` allows these processes to continue regardless.
If no process is specified, the most recently-used job is removed (like `bg` and `fg`). If one or more `PID`s are specified, jobs with the specified process IDs are removed from the job list. Invalid jobs are ignored and a warning is printed.
If a job is stopped, it is sent a signal to continue running, and a warning is printed. It is not possible to use the `bg` builtin to continue a job once it has been disowned.
The PID of the desired process is usually found by using <a href="index.html#expand-process">process expansion</a>, which can specify jobs or search by process name.
`disown` returns 0 if all specified jobs were disowned successfully, and 1 if any problems were encountered.
\subsection disown-example Example
`firefox &; disown` will start the Firefox web browser in the background and remove it from the job list, meaning it will not be closed when the fish process is closed.
`disown (jobs -p)` removes all jobs from the job list without terminating them.
\section else else - execute command if a condition is not met
\subsection else-synopsis Synopsis
\fish{synopsis}
if CONDITION; COMMANDS_TRUE...; [else; COMMANDS_FALSE...;] end
\endfish
\subsection else-description Description
`if` will execute the command `CONDITION`. If the condition's exit status is 0, the commands `COMMANDS_TRUE` will execute. If it is not 0 and `else` is given, `COMMANDS_FALSE` will be executed.
\subsection else-example Example
The following code tests whether a file `foo.txt` exists as a regular file.
`emit` emits, or fires, an event. Events are delivered to, or caught by, special functions called event handlers. The arguments are passed to the event handlers as function arguments.
\subsection emit-example Example
The following code first defines an event handler for the generic event named 'test_event', and then emits an event of that type.
\fish
function event_test --on-event test_event
echo event test: $argv
end
emit test_event something
\subsection notes Notes
Note that events are only sent to the current fish process as there is no way to send events from one fish process to another.
\section eval eval - evaluate the specified commands
\subsection eval-synopsis Synopsis
\fish{synopsis}
eval [COMMANDS...]
\endfish
\subsection eval-description Description
`eval` evaluates the specified parameters as a command. If more than one parameter is specified, all parameters will be joined using a space character as a separator.
\subsection eval-example Example
The following code will call the ls command. Note that `fish` does not support the use of shell variables as direct commands; `eval` can be used to work around this.
\section exec exec - execute command in current process
\subsection exec-synopsis Synopsis
\fish{synopsis}
exec COMMAND [OPTIONS...]
\endfish
\subsection exec-description Description
`exec` replaces the currently running shell with a new command. On successful completion, `exec` never returns. `exec` cannot be used inside a pipeline.
\subsection exec-example Example
`exec emacs` starts up the emacs text editor, and exits `fish`. When emacs exits, the session will terminate.
`exit` causes fish to exit. If `STATUS` is supplied, it will be converted to an integer and used as the exit code. Otherwise, the exit code will be that of the last command executed.
If exit is called while sourcing a file (using the <a href="#source">source</a> builtin) the rest of the file will be skipped, but the shell itself will not exit.
\section faq-envvar How do I set or clear an environment variable?
Use the <a href="commands.html#set">`set`</a> command:
\fish{cli-dark}
set -x key value
set -e key
\endfish
\section faq-login-cmd How do I run a command every login? What's fish's equivalent to .bashrc?
Edit the file `~/.config/fish/config.fish`, creating it if it does not exist (Note the leading period).
<hr>
\section faq-prompt How do I set my prompt?
The prompt is the output of the `fish_prompt` function. Put it in `~/.config/fish/functions/fish_prompt.fish`. For example, a simple prompt is:
\fish{cli-dark}
function fish_prompt
set_color $fish_color_cwd
echo -n (prompt_pwd)
set_color normal
echo -n ' > '
end
\endfish
You can also use the Web configuration tool, <a href="commands.html#fish_config">`fish_config`</a>, to preview and choose from a gallery of sample prompts.
\section faq-cmd-history How do I run a command from history?
Type some part of the command, and then hit the @cursor_key{↑,up} or @cursor_key{↓,down} arrow keys to navigate through history matches.
<hr>
\section faq-subcommand How do I run a subcommand? The backtick doesn't work!
`fish` uses parentheses for subcommands. For example:
\fish{cli-dark}
for i in (ls)
echo $i
end
\endfish
\section faq-exit-status How do I get the exit status of a command?
Use the `$status` variable. This replaces the `$?` variable used in some other shells.
\fish{cli-dark}
somecommand
if test $status -eq 7
echo "That's my lucky number!"
end
\endfish
If you are just interested in success or failure, you can run the command directly as the if-condition:
\fish{cli-dark}
if somecommand
echo "Command succeeded"
else
echo "Command failed"
end
\endfish
See the documentation for <a href="commands.html#test">`test`</a> and <a href="commands.html#if">`if`</a> for more information.
Use the <a href="commands.html#fish_update_completions">`fish_update_completions`</a> command.
<hr>
\section faq-single-env How do I set an environment variable for just one command?
<i>`SOME_VAR=1 command` produces an error: `Unknown command "SOME_VAR=1"`.</i>
Use the `env` command.
`env SOME_VAR=1 command`
You can also declare a local variable in a block:
\fish{cli-dark}
begin
set -lx SOME_VAR 1
command
end
\endfish
\section faq-customize-colors How do I customize my syntax highlighting colors?
Use the web configuration tool, <a href="commands.html#fish_config">`fish_config`</a>, or alter the <a href="index.html#variables-color">`fish_color` family of environment variables</a>.
<hr>
\section faq-update-manpage-completions How do I update man page completions?
Use the <a href="commands.html#fish_update_completions">`fish_update_completions`</a> command.
<hr>
\section faq-cwd-symlink Why does cd, $PWD and and various fish commands always resolve symlinked directories to their canonical path?
<i>For example if `~/images` is a symlink to `~/Documents/Images`, if I write '`cd images`', my prompt will say `~/Documents/Images`, not `~/images`.</i>
Because it is impossible to consistently keep symlinked directories unresolved. It is indeed possible to do this partially, and many other shells do so. But it was felt there are enough serious corner cases that this is a bad idea. Most such issues have to do with how '..' is handled, and are variations of the following example:
Writing `cd images; ls ..` given the above directory structure would list the contents of `~/Documents`, not of `~`, even though using `cd ..` changes the current directory to `~`, and the prompt, the `pwd` builtin and many other directory information sources suggest that the current directory is `~/images` and its parent is `~`. This issue is not possible to fix without either making every single command into a builtin, breaking Unix semantics or implementing kludges in every single command. This issue can also be seen when doing IO redirection.
Another related issue is that many programs that operate on recursive directory trees, like the find command, silently ignore symlinked directories. For example, ```find $PWD -name '*.txt'``` silently fails in shells that don't resolve symlinked paths.
<hr>
\section faq-cd-implicit I accidentally entered a directory path and fish changed directory. What happened?
If fish is unable to locate a command with a given name, and it starts with '`.`', '`/`' or '`~`', fish will test if a directory of that name exists. If it does, it is implicitly assumed that you want to change working directory. For example, the fastest way to switch to your home directory is to simply press `~` and enter.
<hr>
\section faq-open The open command doesn't work.
The `open` command uses the MIME type database and the `.desktop` files used by Gnome and KDE to identify filetypes and default actions. If at least one of these environments is installed, but the open command is not working, this probably means that the relevant files are installed in a non-standard location. Consider <a href="index.html#more-help">asking for more help</a>.
<hr>
\section faq-default How do I make fish my default shell?
If you installed fish manually (e.g. by compiling it, not by using a package manager), you first need to add fish to the list of shells by executing the following command (assuming you installed fish in /usr/local):
\fish{cli-dark}
echo /usr/local/bin/fish | sudo tee -a /etc/shells
\endfish
If you installed a prepackaged version of fish, the package manager should have already done this for you.
In order to change your default shell, type:
\fish{cli-dark}
chsh -s /usr/local/bin/fish
\endfish
You may need to adjust the above path to e.g. `/usr/bin/fish`. Use the command `which fish` if you are unsure of where fish is installed.
Unfortunately, there is no way to make the changes take effect at once. You will need to log out and back in again.
<hr>
\section faq-titlebar I'm seeing weird output before each prompt when using screen. What's wrong?
Quick answer:
Run the following command in fish:
\fish{cli-dark}
function fish_title; end; funcsave fish_title
\endfish
Problem solved!
The long answer:
Fish is trying to set the titlebar message of your terminal. While screen itself supports this feature, your terminal does not. Unfortunately, when the underlying terminal doesn't support setting the titlebar, screen simply passes through the escape codes and text to the underlying terminal instead of ignoring them. It is impossible to detect and resolve this problem from inside fish since fish has no way of knowing what the underlying terminal type is. For now, the only way to fix this is to unset the titlebar message, as suggested above.
Note that fish has a default titlebar message, which will be used if the fish_title function is undefined. So simply unsetting the fish_title function will not work.
<hr>
\section faq-greeting How do I change the greeting message?
Change the value of the variable `fish_greeting` or create a `fish_greeting` function. For example, to remove the greeting use:
\fish{cli-dark}
set fish_greeting
\endfish
<hr>
\section faq-history Why doesn't history substitution ("!$" etc.) work?
Because history substitution is an awkward interface that was invented before interactive line editing was even possible. Fish drops it in favor of perfecting the interactive history recall interface. Switching requires a small change of habits: if you want to modify an old line/word, first recall it, then edit. E.g. don't type "sudo !!" - first press Up, then Home, then type "sudo ".
Fish history recall is very simple yet effective:
- As in any modern shell, the Up arrow, @cursor_key{↑,Up} recalls whole lines, starting from the last line executed. A single press replaces "!!", later presses replace "!-3" and the like.
- If the line you want is far back in the history, type any part of the line and then press Up one or more times. This will constrain the recall to lines that include this text, and you will get to the line you want much faster. This replaces "!vi", "!?bar.c" and the like.
- @key{Alt,↑,Up} recalls individual arguments, starting from the last argument in the last line executed. A single press replaces "!$", later presses replace "!!:4" and the like.
- If the argument you want is far back in history (e.g. 2 lines back - that's a lot of words!), type any part of it and then press @key{Alt,↑,Up}. This will show only arguments containing that part and you will get what you want much faster. Try it out, this is very convenient!
- If you want to reuse several arguments from the same line ("!!:3*" and the like), consider recalling the whole line and removing what you don't need (@key{Alt,D} and @key{Alt,Backspace} are your friends).
See <a href='index.html#editor'>documentation</a> for more details about line editing in fish.
<hr>
\section faq-cd-minus How to make `-` a shortcut for `cd -`?
In fish versions prior to 2.5.0 it was possible to create a function named `-` that would do `cd -`. Changes in the 2.5.0 release included several bug fixes that enforce the rule that a bare hyphen is not a valid function (or variable) name. However, you can achieve the same effect via an abbreviation:
\fish{cli-dark}
abbr -a -- - 'cd -'
\endfish
<hr>
\section faq-uninstalling Uninstalling fish
Should you wish to uninstall fish, first ensure fish is not set as your shell. Run `chsh -s /bin/bash` if you are not sure.
Next, do the following (assuming fish was installed to /usr/local):
\section faq-reserved-chars Unicode private-use characters reserved by fish
Fish reserves the <a href="http://www.unicode.org/faq/private_use.html">Unicode private-use character range</a> from U+F600 thru U+F73F for internal use. Any attempt to feed characters in that range to fish will result in them being replaced by the Unicode "replacement character" U+FFFD. This includes both interactive input as well as any file read by fish (but not programs run by fish).
<hr>
\section faq-third-party Where can I find extra tools for fish?
The fish user community extends fish in unique and useful ways via scripts that aren't always appropriate for bundling with the fish package. Typically because they solve a niche problem unlikely to appeal to a broad audience. You can find those extensions, including prompts, themes and useful functions, in various third-party repositories. These include:
This is not an exhaustive list and the fish project has no opinion regarding the merits of the repositories listed above or the scripts found therein. We mention these only because you may find within them a solution to a need you have such as supporting the `&&` and `||` operators or improved integration with other tools that you use.
`fg` brings the specified <a href="index.html#syntax-job-control">job</a> to the foreground, resuming it if it is stopped. While a foreground job is executed, fish is suspended. If no job is specified, the last job to be used is put in the foreground. If PID is specified, the job with the specified group ID is put in the foreground.
The PID of the desired process is usually found by using <a href="index.html#expand-process">process expansion</a>. Fish is capable of expanding far more than just the numeric PID, including referencing itself and finding PIDs by name.
\subsection fg-example Example
`fg %1` will put the job with job ID 1 in the foreground.
\section fish fish - the friendly interactive shell
\subsection fish-synopsis Synopsis
\fish{synopsis}
fish [OPTIONS] [-c command] [FILE [ARGUMENTS...]]
\endfish
\subsection fish-description Description
`fish` is a command-line shell written mainly with interactive use in mind. The full manual is available <a href='index.html'>in HTML</a> by using the <a href='#help'>help</a> command from inside fish.
The following options are available:
- `-c` or `--command=COMMANDS` evaluate the specified commands instead of reading from the commandline
- `-d` or `--debug-level=DEBUG_LEVEL` specify the verbosity level of fish. A higher number means higher verbosity. The default level is 1.
- `-i` or `--interactive` specify that fish is to run in interactive mode
- `-l` or `--login` specify that fish is to run as a login shell
- `-n` or `--no-execute` do not execute any commands, only perform syntax checking
- `-p` or `--profile=PROFILE_FILE` when fish exits, output timing information on all executed commands to the specified file
- `-v` or `--version` display version and exit
- `-D` or `--debug-stack-frames=DEBUG_LEVEL` specify how many stack frames to display when debug messages are written. The default is zero. A value of 3 or 4 is usually sufficient to gain insight into how a given debug call was reached but you can specify a value up to 128.
The fish exit status is generally the exit status of the last foreground command. If fish is exiting because of a parse error, the exit status is 127.
\section fish_config fish_config - start the web-based configuration interface
\subsection fish_config-description Description
`fish_config` starts the web-based configuration interface.
The web interface allows you to view your functions, variables and history, and to make changes to your prompt and color configuration.
`fish_config` starts a local web server and then opens a web browser window; when you have finished, close the browser window and then press the Enter key to terminate the configuration session.
`fish_config` optionally accepts name of the initial configuration tab. For e.g. `fish_config history` will start configuration interface with history tab.
If the `BROWSER` environment variable is set, it will be used as the name of the web browser to open instead of the system default.
\subsection fish_config-example Example
`fish_config` opens a new web browser window and allows you to configure certain fish settings.
\section fish_indent fish_indent - indenter and prettifier
\subsection fish_indent-synopsis Synopsis
\fish{synopsis}
fish_indent [OPTIONS]
\endfish
\subsection fish_indent-description Description
`fish_indent` is used to indent a piece of fish code. `fish_indent` reads commands from standard input and outputs them to standard output or a specified file.
The following options are available:
- `-w` or `--write` indents a specified file and immediately writes to that file.
- `-i` or `--no-indent` do not indent commands; only reformat to one job per line.
- `-v` or `--version` displays the current fish version and then exits.
- `--ansi` colorizes the output using ANSI escape sequences, appropriate for the current $TERM, using the colors defined in the environment (such as `$fish_color_command`).
- `--html` outputs HTML, which supports syntax highlighting if the appropriate CSS is defined. The CSS class names are the same as the variable names, such as `fish_color_command`.
- `-d` or `--debug-level=DEBUG_LEVEL` enables debug output and specifies a verbosity level (like `fish -d`). Defaults to 0.
- `-D` or `--debug-stack-frames=DEBUG_LEVEL` specify how many stack frames to display when debug messages are written. The default is zero. A value of 3 or 4 is usually sufficient to gain insight into how a given debug call was reached but you can specify a value up to 128.
- `--dump-parse-tree` dumps information about the parsed statements to stderr. This is likely to be of interest only to people working on the fish source code.
`fish_key_reader` is used to study input received from the terminal and can help with key binds. The program is interactive and works on standard input. Individual characters themselves and their hexadecimal values are displayed.
The tool will write an example `bind` command matching the character sequence captured to stdout. If the character sequence matches a special key name (see `bind --key-names`), both `bind CHARS ...` and `bind -k KEYNAME ...` usage will be shown. Additional details about the characters received, such as the delay between chars, are written to stderr.
The following options are available:
- `-c` or `--continuous` begins a session where multiple key sequences can be inspected. By default the program exits after capturing a single key sequence.
- `-d` or `--debug-level=DEBUG_LEVEL` enables debug output and specifies a verbosity level (like `fish -d`). Defaults to 0.
- `-D` or `--debug-stack-frames=DEBUG_LEVEL` specify how many stack frames to display when debug messages are written. The default is zero. A value of 3 or 4 is usually sufficient to gain insight into how a given debug call was reached but you can specify a value up to 128.
The delay in milliseconds since the previous character was received is included in the diagnostic information written to stderr. This information may be useful to determine the optimal `fish_escape_delay_ms` setting or learn the amount of lag introduced by tools like `ssh`, `mosh` or `tmux`.
`fish_key_reader` intentionally disables handling of many signals. To terminate `fish_key_reader` in `--continuous` mode do:
- press `Ctrl-C` twice, or
- press `Ctrl-D` twice, or
- type `exit`, or
- type `quit`
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
