Not for _everything_ because that causes too many options to be
generated (which is an issue for git as it is), but for modified, staged
and added files - which is where it is most useful.
Fixes#901 as far as I'm concerned.
git has options that can appear before commands, but not all of
them, and some of them need an argument. This means
`__fish_seen_subcommand_from` will give too many false-positives, while
`[ (count $cmd) -eq 2 ]` will give too many false-negatives.
Instead go through all arguments and check if they are in that list of
options that can be before a command and skip the argument for them, if
any.
This is to the benefit of systems with ancient GNU sed, which does not
recognize "-E", but only "-r".
Fixes#2305 - even if it doesn't replace all `sed -E` invocations in the
codebase, the others are unlikely to occur on CentOS and other similarly
crusty systems.
`__fish_apm_using_command` was incorrectly taking lists of commands, new function added to support multiple a command having synonyms.
Simplify switch statement
Also remove superfluous function.
Allow for multiple completions after a command
Useful for removing packages, will complete for more than one.
Code improvements
`sort -u | uniq` is completely redundant, calling grep for every
status-pair is unnecessary, `contains` doesn't take the word "in" as
special.
None of these are critical and there's basically no performance benefit
since this function is utterly dominated by hg calls.
This doesn't add anything except slowing the function down by about
33%. Checking for a branch is just as good and that is displayed in the
prompt anyway.
According to the newer code below:
xdg_data_home = os.getenv('XDG_DATA_HOME', '~/.local/share')
the actual default path is ~/.local/share/fish/generated_completions/
This is used in at least 4 places, all of which have a bug in that they
print "options" as a valid repo. It seems better to fix it once,
especially given that there are tons of AUR helpers and pacman wrappers,
all of which might need this info.
net_tools, which provides `ifconfig` and `netstat`, among other things,
has last been updated in 2013. This means `ifconfig` on linux is
basically dead.
Instead of ifconfig, use `ip` (from iproute2), which is much more powerful and
provides a much more annoying commandline syntax.
Instead of netstat, just look at /sys/class/net.
fish_user_key_bindings is the user's, and they should know if they want
vi-ish bindings or emacs-ish (or nano-ish). If they want to define
multiple, they can also do that (e.g. via checking what
$fish_key_bindings is set to).
Fixes#2254
CC @kballard
This doesn't work with fish_config.
For terlar and pythonista, remove unnecessary color setting.
For informative+git and pythonista, move variable setup into fish_prompt
Fixes#1141
See #1925: This allows users to disable the cnf-logic which can be quite
slow on small hardware (like a raspberry pi).
Squashed commit of the following:
commit 742a59e30d8db24b6bb5067d4204d4b5cc01c1c3
Author: Fabian Homborg <FHomborg@gmail.com>
Date: Sun Aug 30 18:23:41 2015 +0200
Erase startup cnf-handler early
Simplifies the code a bit - in particular it removes the special-casing
from the startup handler.
commit 638a97e7f31f302b65e044c93c638c03a69e31f5
Author: Fabian Homborg <FHomborg@gmail.com>
Date: Mon Aug 24 20:14:46 2015 +0200
Make overriding cnf-handler work
Do this by renaming the __fish_command_not_found_handler used during
startup to __fish_startup_command_not_found_handler. That allows us to
check if __fish_command_not_found_handler has been defined and skip the
setup of the normal one.
Now disabling cnf-handling can be done via defining an empty
__fish_command_not_found_handler in config.fish
This adds a special colorscheme and prompt function guaranteed to work
on a VT and activates them automatically if $TERM = "linux".
set_color is overridden to only allow the 8 colors VTs have (under the
assumption those are always the same) and the color variables are
shadowed with global ones so they don't pollute our nice capable terms.
This is already done by fish before calling the completion.
It breaks completion with combiners (#2025) and also with wrappers.
(This does not include git because that's better solved in #2145)
There are two main problems in the existing Fossil autocompletion that this
patch solves:
* Because Fossil lacks an alias system similar to those in Hg and Git,
wrapper scripts are common, and aliasing them to `fossil` is also fairly
common. The lack of the `command fossil` pattern in the completions script
meant that the actual fossil command might not be called, but rather the
alias. This problem has been fixed by introducing a __fish_fossil command,
similar to the __fish_hg and __fish_git commands in those completion shells,
that does this, and converting all explicit fossil calls in the completion
script to use __fish_fossil instead
* Because there's now a centralized location for calling Fossil, I also moved
all of the repetitive stderr redirects that function.
This results in more robust and cleaner code.
This was too simplistic, among other things it completed things that
looked like key ids but weren't, didn't turn "\x3a" back into
colons (which made the argument invalid)....
gpg is weird.
Might fix#2150
Bit one: Make all the fossil command invocations throw away stderr so we don't
get annoying messages when not in a repository.
Also:
- Move checkout into alphabetical order.
- Fix ls to complete against tags for -r option, not no option.
- Add missing option to delete command.
- Make commit complete against modified files.
- Make add only complete against extra files.
- Remove now ununused function to list extra & modified files.
- Add -f option in a number of places where it seemd appropriate.
__fish_complete_mime used in that way is a no-op on current fish anyway,
and emacs is by no means useful for just text files (it can also view
PDFs, images, ...).
Otherwise this completion currently only offers options, not arguments.
For most these are pretty much incompletable (lisp code, for example),
and for others it's just not all that useful.
Signed-off-by: David Adam <zanchey@ucc.gu.uwa.edu.au>
This used to be a function because we didn't have complete -w
Use that and it becomes a bit simpler.
This also simplifies the code in a few other ways (like removing a
useless-use-of-cat)
and adds comments about a few edgecases.
New sample prompt from Acidhub (github.com/acidhub)
This prompt show user|path (full), and a small symbol to
show last command status.
If in a git repository, it's show after the path several
symbols to indicate the branch status and the branch name.
Very handy to me so far.
Signed-off-by: Fernando Manfredi <contact@acidhub.click>
changed `function __trap_handler_EXIT --on-exit %self` to `function __trap_handler_EXIT --on-process-exit %self`
I'm guessing the on-exit syntax was from an older version? Trapping EXIT with that syntax caused errors.
The following behaviour is added:
- an empty pushd exchanges the top two directories in the stack;
- pushd +<n> rotates the stack so that the n-th directory (counting from the left of the list shown by dirs, starting with zero) is at the top;
- pushd -<n> rotates the stack so that the nth directory (counting from the right of the list shown by dirs, starting with zero) is at the top.
1. When run with no arguments, make abbr do the equivalent
of `abbr --show`
2. Enable "implicit add", e.g. `abbr gco git checkout`
3. Teach `abbr --show` to not use quotes for simple cases
4. Teach abbr to output -- when the abbreviation has
leading dashes
Add some basic tests to abbr too.
Add a new function fish_mode_prompt which (if it is defined) has its output
prepended to the left prompt. Rather than replacing the prompt wholesale, make
fish_vi_mode enable this function by setting a variable __fish_vi_mode. This
enables vi mode to interoperate nicely with custom prompts. Users who want
to change how the mode is reported can either redefine this function or
erase it entirely. Fixes#1988.
- Add four new functions: forward-bigword, backward-bigword,
kill-bigword, backward-kill-bigword
- Add new enum move_word_style_whitespace and related state machine
method
- Change vi key bindings to operate on bigwords: B, gE, W, E, dW, diW,
daW, dE, dB, dgE, cW, ciW, caW, cE, cB, cgE, yW, yiW, yaW, yE, yB,
ygE
Notification is sent using an OSC 777 escape sequence as described at
http://known.phyks.me/2014/local-notifications-for-weechat-and-urxvt.
The specific notification is crafted to match that emitted by bash
when running under Fedora 22 with the "vte-profile" RPM installed.
See the code for "__vte_prompt_command" starting at
http://pkgs.fedoraproject.org/cgit/vte291.git/tree/vte291-command-notify.patch#n307
to see exactly what bash produces. My approach is, however, a bit
more paranoid about control characters embedded in commands.
Gnome-terminal 3.16 responds to this escape sequence by posting a
desktop notification if the containing terminal window does not have
focus. This lets the user know that a long-running background command
has completed. Job notification is promoted as a Fedora 22 feature
(http://fedoramagazine.org/terminal-job-notifications-in-fedora-22-workstation/),
so it would be good for fish users to be benefit from it.
Conversely, anyone who does not want this feature can use "functions
--erase __notify_vte_command_completed" to turn it off.
With the fix for #365, fish_command_not_found event handlers
receive the command and all of its arguments. But commands
like /usr/lib/command-not-found expect only the command name.
So when invoking an external command, just pass the command
name, not all of the arguments.
In 73f344f41b, we allowed autoloaded functions to be deleted.
For some reason, funcsave immediately deletes the function it
creates. This previously did very little, since the function would
immediately be re-autoloaded, but with the fix for 73f344f41b
the function gets tombstoned. So the effect is that funcsave
makes the function disappear! This simply removes the erase call,
which dates back to fish 1.x.
It seems that `ul` can't handle the escape sequences for bold text that `nroff` generates on my system. Fixed by either removing `| ul`, or adding `-c` to the `nroff` command.
Needs testing for old (OSX?) versions of nroff.
Unfortunately, list-unit-files doesn't understand --state=loaded
This needs a new function to explicitly use list-units
This reverts commit 9f521b7694.
Support for space-delimited abbreviations was added to the expansion
parser in fbade198; this commit extends that support to the user-facing
tools, and documents the space-separated behaviour. Equals-delimited
abbreviations are expected to be removed before the next release.
Work on #731.
There is no CTRL-C handler for the default mode in the vi bindings. This makes it difficult to say "never mind" and start a new command line like you can do in bash's vi mode.
There were CTRL-C handlers for insert and visual modes that go back to default mode, but nothing happens in default mode. I copy-pasted the CTRL-C handler from the default key bindings file.
(Ideally, the behaviour of git could be implemented: pipe the input
through a pager iff the length is > window size and in interactive
mode).
Closes#1076.
The terminal width magic that __fish_print_help learned doesn't help
when builtin_print_help runs it in a subshell. Instead, add an
undocumented --tty-width flag to __fish_print_help that's used to pass
the terminal width.
As a result of this rewrite, the output now:
* Expands to fit the terminal width, like `man` does
* Preprocesses the manpage with `tbl` just in case, since `man` does
this, even though I doubt any fish manpages use `tbl` formatting.
* Handle bold/underline with the `ul` command as it was designed for
instead of trying to fake it with `sed`.
* Compresses blank lines as `man` does with the default `less -is`
pager.
The usage is still the same, but it's a lot more robust, and also no
longer assumes $fish_user_abbreviations must be a universal variable.
This also fixes the unexpected error output when calling `abbr -a` with
no existing abbreviations.
Calling `abbr -a` with an abbreviation that already exists now silently
overwrites the abbreviation, just like `function` and `bind` do, instead
of complaining.
This fixes the issue with nonexistant directories (some Linux
distributions put these for local modules), and also fixes the
issue of dot meaning any character instead of simply dot.
# The first commit's message is:
Simplify default fish_prompt
No need for the set_color caching now that it's a builtin.
Also simplify the 3 classic prompts in fish_config's sample_prompts set.
Remove comment that AFAICT is not true anymore.
Ensure someone setting __fish_active_key_bindings as a universal
variable doesn't screw up the initial keybinding load.
All opam subcommands and descriptions are covered, along with
all the flags that are common to all commands. However, only
`opam config` has complete subsubcommand coverage.
Apparently, in zsh, Meta+H can be used to display the manpage for
the current command. This commit adds this zsh feature to fish shell.
The F1 keybinding is left, although it's now secondary according to
fish help, as some terminal emulators don't let the user press F1 key.
Use the new `read -z` flag to complete git aliases better. This approach
won't break if an alias contains a newline.
Also fix stash completion, which was broken on BSD sed.
Setting a non-existant path component to PATH logs an error to stderr.
This is not appropriate for non-interactive temporary modifications,
like the one done by the `sudo` completion helper function.
In the base config.fish, fish_function_path and fish_complete_path have
$__fish_datadir/{functions,completions} added to them if not already
present. For some reason they were replacing the final path component
instead of being added on to the end.
The new --wraps functionality was breaking aliases of the form
`alias foo='bar baz'`. That is, aliases where the body is multiple
words. Extract the first word of the body and use that instead.
Use better errors for aliases with no name or no body.
Add the --wraps option to 'complete' and 'function'. This allows a
command to (recursively) inherit the completions of a wrapped command.
Fixes#393.
When evaluating a completion, we inspect the entire "wrap chain" for a
command, i.e. we follow the sequence of wrapping until we either hit a
loop (which we silently ignore) or the end of the chain. We then
evaluate completions as if the wrapping command were substituted with
the wrapped command. Currently this only works for commands, i.e.
'complete --command gco --wraps git\ checkout' won't work (that would
seem to encroaching on abbreviations anyways). It might be useful to
show an error message for that case.
The commandline builtin reflects the commandline with the wrapped
command substituted in, so e.g. git completions (which inspect the
command line) will just work. This sort of command line munging is
also performed by 'complete -C' so it's not totally without precedent.
'alias will also now mark its generated function as wrapping the
'target.
- Require all requests to use a session path.
- Use a redirect file to avoid exposing the '/start' URL on the
command line, as it contains the cookie value.
Fix for CVE-2014-2914.
Closes#1438.
Currently fish doesn't recognize toor as special. However, it's likely
that on BSD systems, fish shell will be used on toor, not on root (toor
is an intentionally existing account to use more advanced shell on, like
shell).
Use `functions -q` instead of searching the `functiosn -na` list for the
provided word. This may result in an automatically-loaded function being
sourced, but that happens anyway with the default output.
This change means the results of `test -q foo` can be relied upon to
indicate whether `foo` can actually be invoked. Previosly, if `foo` was
the name of an automatically-loaded function file but did not actually
define a function `foo`, and there was no execuable `foo`, then `type -q
foo` would lie and say `foo` can be invoked when it can't.
The --quiet flag is useful when only the exit status matters.
Fix the documentation for the -t flag to no longer claim that `type` can
print "keyword", as it never does that.
Stop printing a blank line for functions/builtins when the -p flag has
been passed. It's just not useful.
Track whether -a and -f have been supplied separately. That way both
`type -a -f command` and `type -f -a command` behaves correctly, as does
`type -a -f foo` where there are multiple executables named `foo` in the
$PATH.
Stop using getopt to parse flags. It's far more expensive than
necessary, and results in long flags not being parsed on OS X. This also
allows args starting with - after the options list to be properly
interpreted as a value to test.
Print the error message to stderr as is appropriate.
Use the new `command -p` functionality when the -a flag has not been
provided (`command` does not have any equivalent to the -a flag),
instead of using `which`. This is faster and also avoids any possible
disagreement between `which` and what fish thinks is valid.
Stop testing every path to see if it's executable, that test has already
been done by `which` or `command -p`.
The end result is `type -P ls` is roughly 250% faster, according to
profiling, on my OS X machine.
* use $XDG_CACHE_HOME for __fish_print_packages completion caches
* when starting fishd, redirect fishd output to /dev/null, not a
predictable path
Fix for CVE-2014-3219.
Closes#1440.
When you chroot in Debian, bash shows the chroot environment in the prompt:
```bash
...
if [ -z "${debian_chroot:-}" ] && [ -r /etc/debian_chroot ]; then
debian_chroot=$(cat /etc/debian_chroot)
fi
PS1='${debian_chroot:+($debian_chroot)}\u@\h:\w\$ '
...
```
This is the effect:
```
(chroot_env) user@host:~#
```
It is useful when chrooting, since usually the hostname remains the same and thus you can't distinguish where you are.
Fix for CVE-2014-2906.
Closes a race condition in funced which would allow execution of
arbitrary code; closes a race condition in psub which would allow
alternation of the data stream.
Note that `psub -f` does not work (#1040); a fix should be committed
separately for ease of maintenance.
Closes#1437
The if statement checking the output of hg bookmarks uses two conditions
joined by the or keyword. However, only the first part was being used.
Wrapping the two statements with begin and end properly combines them.
At some point the non-verbose, non-informative variant of the prompt
(e.g. the variant that looks like the bash prompt) was modified to try
and show the behind/ahead counts the same way the informative prompt
does. Besides being wrong, it also didn't work because behind/ahead
weren't defined.
Before this change, fish config used 0 as its address. However, this
isn't a good idea from security point of view, as web service can be
accessed from everywhere, and do anything on the account it was ran on.
This also deals with firewalls which block the access to 0 even from
the host machine itself. It possibly might fix#673, but I'm not sure.
Previously, fish's command_not_found handler would be installed in
__fish_config_interactive. Errors that occured early in startup (e.g. in
config.fish) or in non-interactive mode would therefore not be reported.
With this change, fish now exposes its default cnf handler as
__fish_default_command_not_found_handler . config.fish then installs a
cnfh that invokes the default. When fish goes interactive, the initial
cnfh is overwritten with a fancier one, that may in turn fall back to
invoking the default.
commit d81ae2665f
Author: Max Gonzih <gonzih@gmail.com>
Date: Sun Feb 2 16:22:18 2014 +0300
Check for command-not-found command on suse
commit 004b794c82
Author: Max Gonzih <gonzih@gmail.com>
Date: Sun Feb 2 14:04:41 2014 +0300
Fix cnf handler for Suse and Fedora
fixes#1208
