Interactive fish with output redirected ("fish >/dev/null")
is not a common use case but it is valid.
Perhaps surprisingly, "fish >some-file" *does* print terminal escape codes
(colors, cursor movement etc.) even if terminal output is not a TTY.
This is typically harmless, and potentially useful for debugging.
We also send blocking terminal queries but those are not harmless in this case.
Since no terminal will receive the queries, we hang; indefinitely as of today,
but we should give up after a timeout and print an error. Either way that
seems needlessly surprising. Suppress queries in this case.
In future, we should probably do something similar if stdin is not a terminal;
though in that case we're even less likely to be interactive (only "-i"
I think).
This documents an invariant established by 532abaddae (Invalidate stale
autosuggestions eagerly, 2024-12-25). It was initially broken but fixed in
ba4ead6ead (Stop saving autosuggestions that we can't restore, 2025-01-17).
After nix updated to 0.30, all functions related to file descriptor accepts impl AsFd, e.g., BorrowedFd. This PR is a minimal update. It tries to use impl AsFd as long as possible, but uses BorrowedFd in some places. Yes it introduces unsafe, but doesn't introduce new unsafe code.
We can parse two different things via Ast:
1. A regular job list
2. A freestanding argument list, as used in `complete --arguments ...`
This second case is specific to one use.
Prior to this commit, we parsed the Ast and then "forgot" what we parsed,
storing a &dyn Node. Then we had to cast it to the right type, and assert,
and etc.
Make Ast generic over the Node type it parsed, and default the Node type to
JobList. This simplifies call sites.
The input queue doesn't want to be blocked, so let's decide later what to
do with a mouse click. This fixes the potential problem where a mouse click
is ignored if it is received while we're waiting for the terminal to repond
to a query.
These code paths have a tiny bit of logic in common, share that.
Maybe this should be nested in ImplicitEvent, and maybe Eof and CheckExit
shouldn't be..
This object is initialized once just before we start reading from the terminal.
Once seems to be the appropriate type for this. This gets rid of an awkward
enum variant.
We never need to access this from other threads, so a Mutex is overkill.
Leave behind stale variable names like "wait_guard" to be cleaned up by the
next commit.
Since TestInputEventQueuer is used concurrently in tests,
give it its own private object, to avoid borrowing conflicts.
Same for fish_key_reader; this fixes the issue that fish_key_reader potentially
reads keyboard input before a query is finished.
Whenever config.fish runs (interactive) builtin read, we push and pop a
top-level, before the main shell's reader.
The terminal state outlives all readers, so its scope should reflect that
to avoid redundant initialization. Move it into the parser.
This is also used by a following commit that wants to access the query state
from a builtin. This should work even if no reader is active.
Note that Mutex doesn't really make sense here - the next commit will fix that.
We only query for the cursor position if either
1. if the terminal sends an XTGETTCAP response indicating it supports scroll
forward (terminfo code indn)
2. if the terminal sends a mouse click event
In practice, the terminals that do either are also capable of reporting the
cursor position, so let's require that for now.
We could definitely make this optional, and degrade gracefully -- that's
what I originally intended.
But maybe we don't need it? Keeps things simpler. In future, we can add a
timeout, and print an error, to help debugging terminals.
Old versions of ConHost and Putty can't parse DCS sequences.
For this reason, we briefly switch to the alternate screen buffer while sending
DCS-format (e.g. XTGETTCAP) queries. For extra paranoia, we wrapped this
procedure in a synchronized update. This doesn't seem to be needed; neither
ConHost nor Putty show glitches when the synchronized update is omitted.
As of today, every terminal that implements XTGETTCAP also implements
synchronized updates but that might change.
Let's remove it, to reduce surprise for users and terminal developers.
As a bonus, this also fixes a glitch on Terminal.app which fails to parse
the synchronized-update query (`printf '\x1b[?2026$p'`) and echoes the "p"
(bug report ID FB17141059). Else we could work around this with another
alternate screen buffer.
Unfortunately, this change surfaces two issues with GNU screen. For one,
they don't allow apps to use the alternate screen features (the user may
allow it with "altscreen on"). Second, screen unconditionally echoes
the payload of DCS commands. A possible fix has been suggested at
https://lists.gnu.org/archive/html/screen-devel/2025-04/msg00010.html
I think this combination of behaviors is unique among terminals. I'm sure
there are more terminals that don't parse DCS commands yet, but I think almost
all terminals implement alternate screen buffer. Probably only terminal
multiplexers are prone to this issue. AFAICT very few multiplexers exists,
so we can work around those until they are fixed.
Disable XTGETTCAP queries for GNU screen specifically. Unfortunately screen
does not implement XTVERSION, so I don't know how to reliably identify
it. Instead, check STY and some commonly-used values TERM values.
This has false negatives in some edge cases.
But the worst thing that happens is that "+q696e646e" will be echoed once
at startup, which is easy to ignore, or work around with something like
function workaround --on-event fish_prompt
commandline -f repaint
functions --erase workaround
end
which I don't think we should apply by default, because it can mask other
issues.
We should give screen more time to respond. I guess I'll open an issue so
we don't forget. In doubt, we can always go back to the previous approach
(but implement it in fish script).
Alternative workaround: instead of the alternative screen buffer, we could
try something like clr_eol/clr_bol to erase the spuriously echoed text. I
tried to do this in various ways but (surprisingly) failed.
This workaround was added in commit d66700a0e4 (Color work, 2012-02-11).
I don't understand why it would be necessary, it seems redundant. Bravely
remove it. Would be interesting to know which terminal caused the motivating
problem. Terminal.app and others seem to be unaffected by this change.
Our highlighting module allows different highlight roles for foreground
and background.
Other users of our color-parsing logic have no need for this.
Historically we have passed an "is_background" bool and only parsed what we're
interested in. This sort of makes sense because it's exactly what we need,
but it meant that the special behavior spread quite far.
There is no real need for spreading it; a function with the behavior "parse
a text face but honor is_background" is strictly more complex than "parse a
text face". Additionally, any performance optimization is only relevant if
the user specifies faces that won't be used, which is a very unusual case.
Let's isolate this logic in the highlighting module.
The parent commit got rid of the is_background parameter that determined
the meaning of the return value (fg/bg); since we always return both now,
give them names.
A following commit wants to add other underline styles. At most one underline
style can be active at one time. This can't be modelled well by a flat
bitset, so let's use a composable type here.
Our "Color" type also includes text styling (bold, italics, ...). This doesn't
seem like the perfect way to model this, because the background text styles
are always unused (this is noticeable with some fish_color_* variables).
Introduce a type that represents attributes to apply to text, and nothing
else. Prefer to pass this instead of two Color values with redundant text
style is redundant.
RgbColor can also be a named color, so this name is surprising to readers
who don't have the implicit assumption that named colors are the default
and that RGB support is something novel.
This function hid a bug! It converted two Nodes to `*const ()` which discards
the vtable pointer, using only the data pointer; but two nodes can and do have
the same data pointer (e.g. if one node is the first item in another).
Add the (statically dispatched) is_same_node function with a warning comment,
and adopt that instead.
Now, the right prompt will not be executed only if it is undefined `fish_right_prompt`.
This allows the use of `read -p 'echo left' -R 'echo right'`.
Also changes condition for use of `DEFAULT_PROMPT` for consistency.
This gets the reader out of asting the source and is needed for
autoloader to get it to read a source string directly
Also add an "eval_file_wstr" method as a convenience to run a wstr as if it is a *file*, with a block and stuff
Our use of the terminfo database in /usr/share/terminfo/$TERM is both
1. a way for users to configure app behavior in their terminal (by
setting TERM, copying around and modifying terminfo files)
2. a way for terminal emulator developers to advertise support for
backwards-incompatible features that are not otherwise easily observable.
To 1: this is not ideal (it's very easy to break things). There's not many
things that realistically need configuration; let's use shell variables
instead.
To 2: in practice, feature-probing via terminfo is often wrong. There's not
many backwards-incompatible features that need this; for the ones that do
we can still use terminfo capabilities but query the terminal via XTGETTCAP
directly, skipping the file (which may not exist on the same system as
the terminal).
---
Get rid of terminfo. If anyone finds a $TERM where we need different behavior,
we can hardcode that into fish.
* Allow to override this with `fish_features=no-ignore-terminfo fish`
Not sure if we should document this, since it's supposed to be removed soon,
and if someone needs this (which we don't expect), we'd like to know.
* This is supported on a best-effort basis; it doesn't match the previous
behavior exactly. For simplicity of implementation, it will not change
the fact that we now:
* use parm_left_cursor (CSI Ps D) instead of cursor_left (CSI D) if
terminfo claims the former is supported
* no longer support eat_newline_glitch, which seems no longer present
on today's ConEmu and ConHost
* Tested as described in https://github.com/fish-shell/fish-shell/pull/11345#discussion_r2030121580
* add `man fish-terminal-compatibility` to state our assumptions.
This could help terminal emulator developers.
* assume `parm_up_cursor` is supported if the terminal supports XTGETTCAP
* Extract all control sequences to src/terminal_command.rs.
* Remove the "\x1b(B" prefix from EXIT_ATTRIBUTE_MODE. I doubt it's really
needed.
* assume it's generally okay to output 256 colors
Things have improved since commit 3669805627 (Improve compatibility with
0-16 color terminals., 2016-07-21).
Apparently almost every actively developed terminal supports it, including
Terminal.app and GNU screen.
* That is, we default `fish_term256` to true and keep it only as a way to
opt out of the the full 256 palette (e.g. switching to the 16-color
palette).
* `TERM=xterm-16color` has the same opt-out effect.
* `TERM` is generally ignored but add back basic compatiblity by turning
off color for "ansi-m", "linux-m" and "xterm-mono"; these are probably
not set accidentally.
* Since `TERM` is (mostly) ignored, we don't need the magic "xterm" in
tests. Unset it instead.
* Note that our pexpect tests used a dumb terminal because:
1. it makes fish do a full redraw of the commandline everytime, making it
easier to write assertions.
2. it disables all control sequences for colors, etc, which we usually
don't want to test explicitly.
I don't think TERM=dumb has any other use, so it would be better
to print escape sequences unconditionally, and strip them in
the test driver (leaving this for later, since it's a bit more involved).
Closes#11344Closes#11345
This part of the code could use some love; when we happen to clear the
selected text, we should end the selection.
But that's not how it works today. This is fine for Vi mode, because Vi
mode never deletes in visual mode.
Let's fix the crash for now.
Fixes#11367
With
bind ctrl-r 'sleep 1' history-pager
typing ctrl-r,escape crashes fish in the history pager completion callback,
because the history pager has already been closed.
Prior to 55fd43d86c (Port reader, 2023-12-22), the completion callback
would not crash open a pager -- which causes weird races with the
user input.
Apparently this crash as been triggered by running "playwright",
and -- while that's running typing ctrl-r ligh escape.
Those key strokes were received while the kitty keyboard protocol
was active, possibly a race.
Fixes#11355
Consider command line modifications triggered from fish script via abbreviation
expansion:
function my-abbr-func
commandline -r ""
echo expanded
end
abbr -a foo --function my-abbr-func
Prior to commit 8386088b3d (Update commandline state changes eagerly as well,
2024-04-11), we'd silently ignore the command line modification.
This is because the abbreviation machinery runs something similar to
if my-abbr-func
commandline -rt expanded
end
except without running "apply_commandline_state_changes()" after
"my-abbr-func", so the «commandline -r ""» update is lost.
Commit 8386088b3d applies the commandline change immediately in the abbrevation
function callback, invalidating abbrevation-expansion state.
The abbreviation design does not tell us what should happen here. Let's ignore
commandline modifications for now. This mostly matches historical behavior.
Unlike historical behavior we also ignore modifications if the callback fails:
function my-abbr-func
commandline -r ""
false
end
Remove the resulting dead code in editable_line.
See #11324
Sometimes, the dirname of a completion is much longer than the basename.
When the dirname is the same for many completions, a long common dirname
prefix makes it hard to see the interesting differences in the basenames.
Fix this by collapsing the dirname prefix to "…/".
In future, we should find a generic way to collapse completions that don't
contain slashes.
Closes#8618Closes#11250
In C++ it's easy to make an RAII-type object like "increment a counter for
the duration of this function." Such an object might accept a pointer or
reference, increment the value, and then restore it in its destructor. We
do this all the time - for example to mark a region of code as
non-interactive, etc.
Rust makes this more awkward, because now the reference is tracked by the
borrow checker: it "owns" the object for the duration of the function. This
leads to approaches like "zelf" where the object that marks the parser as
non-interactive itself becomes the new parser, but we can't call it "self"
and it's just yucky.
In this commit we introduce a notion of the "scoped data" of the Parser,
factored out of the library data. This is data which is typically set in a
scoped fashion: whether we are a subshell, are interactive, emit fish_trace
debugging info, etc. Crucially we set this as Rc: this allow the scope
itself to share data with the Parser and we can get rid of lots of "zelf"s.
Introduce a new function `Parser::push_scope` which creates a new scope and
allows modifying these variables associated with the scope. This ends up as
a nice simplification.
This add two commands history-last-token-search-backward and
history-last-token-search-forward which behaves like bash's yank-last-arg. So
similar to history-token-search-* but only considers the last argument for
each command.
Closes#10756Closes#11258
Commit 29dc307111 (Insert some completions with quotes instead of backslashes,
2024-04-13) breaks some workflows. Given
touch '[test] file1'
touch '[test] file2'
ls tes<Tab>
we insert completions quoted, which is inconvenient when using globs.
This implicit quoting feature is somewhat minor. But quotes look nicer,
so let's try to keep them. Either way, users can ask for it by starting a
token with «"».
Use quoting only when we insert unique completions.
Closes#11271
