ridiculousfish
02ddc20c87
Correct signatures of main_thread_request_t's deleted functions
2017-01-23 10:58:38 -08:00
ridiculousfish
520e567390
Eliminate the old-style iothread_perform_on_main
...
All clients now use the std::function form (i.e. lambdas)
2017-01-23 10:46:42 -08:00
ridiculousfish
144e7b0616
Adopt lamba-style perform_on_main in the tests
2017-01-23 10:45:48 -08:00
ridiculousfish
fe8e99138d
Adopt lambda-style perform_on_main in expand.cpp
...
Allows eliminating find_job_data_t
2017-01-23 10:44:36 -08:00
ridiculousfish
e0abfee370
Adopt lambda-style perform_on_main in complete.cpp
...
Allows eliminating complete_load_no_reload
2017-01-23 10:43:34 -08:00
ridiculousfish
66a0f18410
Another fix for Linux build
2017-01-23 10:39:53 -08:00
ridiculousfish
3181bdcb9b
Attempt to fix the Linux build
2017-01-23 10:38:55 -08:00
ridiculousfish
e1adc3a6b2
Make iothread's perform_on_main use std::function
...
This will allow clients to use lambdas instead of having to
define an out-of-line function
2017-01-23 10:37:16 -08:00
ridiculousfish
7a76efa629
Use set_cloexec instead of fcntl directly in iothread.cpp
2017-01-23 09:59:56 -08:00
ridiculousfish
216f7d912a
Clean up some of the memory management in iothread.cpp
...
Store requests directly on the queue, instead of via a heap allocation
2017-01-23 09:56:02 -08:00
ridiculousfish
d373f1fc1d
Lower-snake-case for MainThreadRequest_t and SpawnRequest_t
...
Matches style of rest of the project
2017-01-23 09:34:30 -08:00
ridiculousfish
36a0f745cd
Fix some bogus error messages in builtin_commandline
2017-01-23 09:32:50 -08:00
ridiculousfish
ab189a75ab
Switch a job's process list from a linked list to a vector of pointers
...
Clarifies and simplifies the memory management around process handling.
2017-01-23 09:28:34 -08:00
ridiculousfish
f4476100f2
Remove comment about job_iterator_t being used from signal handlers
...
It is no longer used from signal handlers, and has not been for a while
2017-01-22 00:59:50 -08:00
ridiculousfish
6bdab62358
Make io_buffer_t::create return a shared_ptr
...
Eliminates some manual memory management
2017-01-22 00:44:04 -08:00
ridiculousfish
009a677e0d
Use the STL's make_unique if available
...
Fixes a build error with g++ 6.1
Fixes #3759
2017-01-22 00:32:08 -08:00
ridiculousfish
2e47817adf
Second attempt to fix Xcode build for PCRE2 changes
...
Update osx/shared_headers/pcre2.h
Fixes #3755
2017-01-22 00:11:40 -08:00
Kurtis Rader
176a291ed2
deal with multiline commands which have flags
...
Fixes #3758
2017-01-21 20:23:06 -08:00
ridiculousfish
1875222e93
Use one invocation of $INSTALL instead of looping
...
Speed up our Makefile by using $INSTALL once per type of file,
rather than passing them individually.
2017-01-21 17:39:00 -08:00
ridiculousfish
439f233ccc
Make fire_event_callback take a std::function instead of function pointer
2017-01-21 17:15:45 -08:00
ridiculousfish
a91dad35db
Use unique_ptr instead of new for history tests
2017-01-21 17:14:56 -08:00
ridiculousfish
6f745762bb
Make universal_notifier_t use unique_ptr instead of raw pointers
2017-01-21 16:56:45 -08:00
ridiculousfish
f0065cda13
Clean up event_t handling
...
Use shared_ptr instead of the silly killme list
2017-01-21 16:48:07 -08:00
ridiculousfish
8a0d4854e8
Replace auto_ptr with unique_ptr
2017-01-21 16:10:42 -08:00
ridiculousfish
b3fff2d779
Switch to using unique_ptr for builtin_test
...
Removes a lot of terrifying manual memory management
2017-01-21 16:08:53 -08:00
ridiculousfish
754b0e9b91
Use unique_ptr in string_replace()
...
Eliminates some manual calls to delete
2017-01-21 15:47:12 -08:00
ridiculousfish
3139ad0d4d
Use unique_ptr in builtin_commandline_scoped_transient_t
...
Allows removing a manual call to delete
2017-01-21 15:45:38 -08:00
ridiculousfish
3272505891
Eliminate the "Clean up the block stack" logic in eval_block_node
...
The block stack is now sound, and no longer needs this ancient
cleanup logic, which tried to account for cases where blocks
were pushed but never popped.
2017-01-21 15:42:35 -08:00
ridiculousfish
0991e398bb
Clean up parser_t's block stack
...
Currently the block stack is just a vector of pointers.
Clients must manually use new() to allocate a block, and then
transfer ownership to the stack (so must NOT delete it).
Give the parser itself responsibility for allocating blocks too,
so that it takes over both allocation and deletion. Use unique_ptr
to make deletion less error-prone.
2017-01-21 15:35:35 -08:00
ridiculousfish
ac8b27fcb1
Implement and use make_unique
...
Allows avoiding some explicit calls to new(), which can look suspicious
2017-01-21 15:02:41 -08:00
ridiculousfish
5b108efde4
Use unique_ptr in builtin_string
...
Avoids manual calls to delete
2017-01-21 14:54:01 -08:00
ridiculousfish
f2884343b3
Use unique_ptr in a parser's execution_context list
...
Avoids requiring manual calls to delete
2017-01-21 14:53:52 -08:00
ridiculousfish
16bc7b48b5
Make profile_items use unique_ptr instead of raw pointers
2017-01-21 14:33:17 -08:00
ridiculousfish
9efa897d0d
Early steps towards rationalizing SIGINT handling
...
Previously we would try to walk all the blocks (from within the
signal handler!) and mark them as skipped. Stop doing that, it's
wildly unsafe.
Also rationalize how the skip flag is set per block. Remove places
that shouldn't set it (e.g. break and continue shouldn't set skip
on the loop block).
2017-01-21 14:15:03 -08:00
ridiculousfish
d8a6c0a91b
Add a SIGALRM handler that does nothing
...
This will be part of a future signal torture-test, to ensure
we are handling EINTR correctly
2017-01-21 14:07:54 -08:00
ridiculousfish
7e3db843cd
Remove FUNCTION_DEF_BLOCK and FAKE_BLOCK
...
These are old-parser block types that are no longer used.
2017-01-21 13:57:05 -08:00
ridiculousfish
bb65b82c56
Mark some signal-related variables as 'volatile sig_atomic_t'
2017-01-21 13:33:46 -08:00
ridiculousfish
bb686a2236
Enable some test_illegal_command_exit_code tests
...
A comment suggests these cause bad_alloc, but this
doesn't seem to happen. If it does happen, we want to hit
it so we can track it down!
2017-01-21 12:56:22 -08:00
ridiculousfish
fc803c75a6
Remove some errant newlines in fish_tests.cpp
2017-01-21 12:55:01 -08:00
ridiculousfish
812e977e62
Clean up /tmp/fish_chunked_read_test.txt after tests
2017-01-21 12:47:05 -08:00
ridiculousfish
5a5a6a6968
Correctly handle multiple chunks from read_in_chunks
...
read_in_chunks does not clear the intermediate string 'str'
between iterations, so every chunk has every other chunk prepended
to it.
A secondary issue is that it calls str2wcstring() on an intermediate
chunk, which may split multi-byte sequences. This needs to be deferred
to the end.
Test added. Fixes #3756
2017-01-21 12:43:20 -08:00
ridiculousfish
f7c133da00
Use long instead of int in read_in_chunks()
...
Fixes warnings about narrowing conversions
2017-01-21 11:53:49 -08:00
ridiculousfish
4c56c89afc
Use wcscmp instead of comparing against a string literal
...
Should fix OpenSUSE build
2017-01-21 11:51:06 -08:00
ridiculousfish
9361470070
Update Xcode build in response to PCRE2 upgrade
...
Replace the osx/pcre2/config.h with a new config.h
from the PCRE2 upgrade
Fixes #3755
2017-01-21 11:48:09 -08:00
Fabian Homborg
ab3149257b
Make test errors redirectable
...
This can't use `fwprintf`, since that goes directly to actual stderr.
It needs to use the passed stream.
2017-01-21 13:11:54 +01:00
Kurtis Rader
2e38cf2a4b
implement means to learn about a functions source
...
This implements a way to use the `functions` command to perform
introspection to learn about the characteristics of a function. Such as
where it came from.
Fixes #3295
2017-01-20 21:48:41 -08:00
Kurtis Rader
2be1288cac
handling when stty reports zero for termsize
...
If the kernel reports a size of zero for the rows or columns (i.e., what
`stty -a` reports) fall back to the `COLUMNS` and `LINES` variables. If
the resulting values are not reasonable fallback to using 80x24.
Fixes #3740
2017-01-20 15:34:29 -08:00
Clément Martinez
082d4b5a7e
Add light completions
2017-01-21 00:33:58 +01:00
Fabian Homborg
c111a65622
Stack completions: Fix dead link
...
This was just reshuffled a bit. Now it uses the main page (which then
redirects to the docs anyway), which should be a bit more stable.
2017-01-20 21:46:59 +01:00
Fabian Homborg
e89c66a224
git completions: git push uses set-upstream
...
This was a misunderstanding in 96a28df
2017-01-20 15:55:39 +01:00
