github-mirror/fish-shell
2
0
Fork 0
mirror of https://github.com/fish-shell/fish-shell.git synced 2024-11-30 14:24:35 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity
10,282 Commits 36 Branches 80 Tags 117 MiB
2ce827343e
Commit Graph

2183 Commits

Author SHA1 Message Date
ridiculousfish
2ce827343e Universal variables to become a latch variable 2019-05-04 20:58:35 -07:00
ridiculousfish
b0a695e9fa Migrate the global exec_count into the parser 2019-05-04 20:58:35 -07:00
ridiculousfish
fe68287cb0 Make miscellaneous variables thread-safe 2019-05-04 20:58:35 -07:00
ridiculousfish
6e143bf50f Eliminate history_collection_t 
This was a pretty useless type.
 2019-05-04 20:58:35 -07:00
ridiculousfish
3d63a68dd0 Make builtin_random thread safe 2019-05-04 20:58:35 -07:00
ridiculousfish
f66e010949 Turn a lot of common.h variables into getter functions 
Improves thread safety.
 2019-05-04 20:58:35 -07:00
ridiculousfish
9dc1fd50c9 Introduce global_safety.h 
This is a set of types that enable characterizing the proper way to access
global variables.
 2019-05-04 20:53:32 -07:00
ridiculousfish
ec522e5978 Migrate s_main_thread_request_queue to owning_lock 2019-05-04 20:42:56 -07:00
ridiculousfish
1e57424011 Thread a parser into function_exists 
Since this may autoload, it needs a parser with which to autoload.
 2019-05-04 20:20:52 -07:00
ridiculousfish
bffacd2fbf Thread a parser into expansion 
Expansion may perform command substitution, which needs to know the parser
to use.
 2019-05-04 19:30:00 -07:00
ridiculousfish
4ce485525e Correct duplicate completion detection 
Stop assuming that completions with the same hash are equal.
 2019-05-04 18:35:22 -07:00
ridiculousfish
923a7ca0f0 Thread the parser into complete() 
Eliminates uses of principal_parser
 2019-05-04 18:17:18 -07:00
ridiculousfish
63bdc949ab Make completion request flags an enum_set 2019-05-04 17:55:57 -07:00
ridiculousfish
acd33cbabb Remove an unnecessary fetch of the principal parser 2019-05-04 16:56:38 -07:00
ridiculousfish
fbd4ff027a Initialize a field that was previously uninitialized 
Caught by UBSan
 2019-05-04 16:47:27 -07:00
ridiculousfish
c0c7b0f86f Make tsan detection gcc compatible 2019-05-04 16:13:55 -07:00
ridiculousfish
ec45f31ad1 Make debug_level an atomic 
Fixes a tsan warning
 2019-05-04 15:28:44 -07:00
ridiculousfish
0dd9f64bd9 Make topic monitor compatible with tsan 
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.
 2019-05-04 13:06:06 -07:00
Fabian Homborg
e91eb85949 argparse: Fix -- regression 
I was a tad overzealous there with not appending the remaining
arguments.

Weird that we didn't test it, though.

Fixes #5857.
 2019-05-03 16:20:40 +02:00
ridiculousfish
649d3ac101 Simplify reporting of invalid config paths 
Do this at a well defined point, instead of randomly the first time they're
queried.
 2019-05-01 17:51:51 -07:00
ridiculousfish
72e43a514b Correct the warning for invalid directories 
This was inadvertently broken.
 2019-05-01 17:47:50 -07:00
ridiculousfish
55e3270ac4 Remove erase_list from process_clean_after_marking 
We don't need to maintain an erase_list in this function any more.
Simply remove jobs that are completed.
 2019-05-01 16:32:14 -07:00
ridiculousfish
3dfaa192da Put back process and job exit events 
These were removed in f8b2e818ed under a
belief that they were unused. But they are documented and supported.
 2019-05-01 16:32:14 -07:00
ridiculousfish
43d668bdc8 Continue to refactor internal loop of process_clean_after_marking 
Factor our logic around when to print a message.
 2019-05-01 16:32:14 -07:00
ridiculousfish
b5d3fadf44 Factor out the individual process handling in process_clean_after_marking 
Helps break up this monolith.
 2019-05-01 16:32:14 -07:00
ridiculousfish
b8170ec1ce Clarify return value of job_reap and process_clean_after_marking 2019-05-01 16:32:14 -07:00
ridiculousfish
9700800ecf Factor disowned job removal into its own function 
This helps break up process_clean_after_marking.
 2019-05-01 16:31:21 -07:00
ridiculousfish
c05e72749a Rename PENDING_REMOVAL to DISOWN_REQUESTED 
A commend implied that PENDING_REMOVAL was broader than it was. In practice
only disown() sets this flag. Rename the flag for clarity.
 2019-05-01 15:37:53 -07:00
David Adam
665ae3787a Switch to runtime check for /proc/self/stat 
Removes a compile-time check that may have affected cross-compilation.

Work on #1067.
 2019-04-30 16:23:28 +08:00
Fabian Homborg
43929ced90 src/builtin_argparse: Work around wgetopt crash 
If on the last argument, and it was an unrecognized option, we can't
call `wgetopt_long()` again, or it'll crash.
 2019-04-29 16:55:55 +02:00
Fabian Homborg
8c9359fdd4 src/builtin_argparse: Add --ignore-unknown flag 
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.
 2019-04-29 15:57:56 +02:00
ridiculousfish
d8ac051f89 Move selection_direction_t to pager.h and make it a class enum 2019-04-28 14:06:03 -07:00
ridiculousfish
2507162f80 Revert "Add a test for autoload_t" 
This reverts commit 51c62d6cc6.

Back out the test while I attempt to fix it
 2019-04-27 20:14:06 -07:00
ridiculousfish
51c62d6cc6 Add a test for autoload_t 2019-04-27 16:16:48 -07:00
ridiculousfish
4ff50eba41 Remove autoload_t, rename autoloader_t to autoload_t 
Now that there are no more clients of autoload_t, delete it and
rename autoloader_t to autoload_t. Also clean up the headers.
 2019-04-27 15:47:08 -07:00
ridiculousfish
68a28106b2 Reimplement completion autoloading via autoloader_t 
This switches the completion autoloading machinery to autoloader_t.
 2019-04-27 15:37:24 -07:00
ridiculousfish
960266fe24 Reimplement the function store and autoloading 
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.
 2019-04-27 15:30:11 -07:00
ridiculousfish
3950dab9ff Add 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.
 2019-04-27 15:26:18 -07:00
ridiculousfish
b7ad6b5bdc Add autoload_file_cache_t 
This will provide the "backing store" for autoloading.
 2019-04-27 15:09:24 -07:00
ridiculousfish
6ec7c50ace Stop removing functions and completions in autoload 
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.
 2019-04-27 14:49:05 -07:00
ridiculousfish
7b44b5ef15 Titlecase LRU template parameters 2019-04-27 12:07:16 -07:00
ridiculousfish
f297543ca1 Make owning_lock's template parameter titlecase instead of uppercase 2019-04-27 12:04:36 -07:00
Fabian Homborg
af0e08e9f1 argparse: Use the current function name by default 
This makes the `--name` option usually unnecessary.

See #5835.
 2019-04-27 15:55:49 +02:00
ridiculousfish
9bc5d60eaf Clean up enum_set.h header 
Include a missing array header, and switch to idiomatic include guards.
 2019-04-26 16:08:19 -07:00
Fabian Homborg
009ecfd7e6 src/tokenizer: Add ":@" to the list of non-path-component chars 
This makes kill-path-component stop there.

Fixes #5841.
 2019-04-26 19:16:21 +02:00
David Gowers
38cadc9d4f Variables as commands are in fact supported, eval docs should not claim otherwise. (#5819) 
Provide an example that somewhat justifies eval's existence in light of this change.

Also correct similar misinformation found in a comment.
 2019-04-26 15:30:13 +02:00
Fabian Homborg
22ce8c23c6 builtin_complete: Allow complete -C something 
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_.
 2019-04-26 15:02:29 +02:00
ridiculousfish
cd86c0ee88 Remove the COMPLETE_SEP define 
It was unused.
 2019-04-25 14:23:37 -07:00
ridiculousfish
96bc8a14ca Promote completion_mode_t to a real type 
Eliminate big #defines like NO_COMMON.
 2019-04-25 14:21:06 -07:00
ridiculousfish
d962668aa0 Remove PATH and COMMAND defines 
Also clean up a bit of builtin_complete
 2019-04-25 13:26:43 -07:00