aboutsummaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAge
...
* Move shorten_fname{,1,s} and shorten_filenames from fileio.cJohn Schmidt2014-04-01
|
* Move mch_expandpath from os_unix.cJohn Schmidt2014-04-01
|
* Move pathcmp from misc2.cJohn Schmidt2014-04-01
|
* Move same_directory from misc2.cJohn Schmidt2014-04-01
|
* Move after_pathsep from misc2.cJohn Schmidt2014-04-01
|
* Move fix_fname from buffer.cJohn Schmidt2014-04-01
|
* Move functions from window.cJohn Schmidt2014-04-01
|
* Move simplify_filename from tag.cJohn Schmidt2014-04-01
|
* remove HAVE_{ISWUPPER,ISWLOWER,TOWUPPER,TOWLOWER}Julian Orth2014-03-31
|
* remove HAVE_STDARG_HJulian Orth2014-03-31
|
* remove HAVE_INTTYPESJulian Orth2014-03-31
|
* remove HAVE_MATH_HJulian Orth2014-03-31
|
* remove HAVE_SETJMP_HJulian Orth2014-03-31
|
* Remove feature HAVE_QSORToni-link2014-03-31
| | | | qsort conforms to C99, so we don't need our own version.
* alloc: xmalloc and xrealloc can't return nullNicolas Hillegeer2014-03-31
| | | | | With this annotation/function attribute, we can inform gcc that these functions never return null, which can save a lot of redundant null checks.
* func-attr: add some extra handy function attrsNicolas Hillegeer2014-03-31
| | | | | | | | | | | The attributes in question are: - nonnull: specify whether a function argument cannot/may not be null - returns_nonnull: specify whether a function will not return a null pointer (example: xmalloc can't return null, so it should be annotated as such). Only available from gcc 4.9 onwards. Currently these attributes are only supported by gcc.
* func-attr: clang 3.5 doesn't support alloc_sizeNicolas Hillegeer2014-03-31
| | | | | | | The attribute was removed in commit c047507 in the clang repository as it was never properly implemented anyway. This fixes compiling with clang 3.5. Fixes issue #429
* Rename ga_init2() to ga_init()John Schmidt2014-03-31
|
* Inline ga_init()John Schmidt2014-03-31
|
* Fix shell.{c,h} comments to match the style guideThiago de Arruda2014-03-31
|
* Add doxygen comments to the functions in shell.cThiago de Arruda2014-03-31
|
* Rewrite `show_shell_mess` as a `ShellOpts` flagThiago de Arruda2014-03-31
|
* Update `call_shell` signature to use `ShellOpts`Thiago de Arruda2014-03-31
|
* Refactor SHELL_* defines into enum typedefThiago de Arruda2014-03-31
| | | | | | The SHELL_* defines are the bitflags that can be passed to `mch_call_shell`. The enum is defined in 'os/shell.h', where all shell-related functions will eventually be defined.
* Refactor/cleanup argument parsing functionsThiago de Arruda2014-03-31
|
* Extract `shell_build_argv` from `mch_call_shell`Thiago de Arruda2014-03-31
|
* Make `extra_shell_arg` a `mch_call_shell` parameterThiago de Arruda2014-03-31
|
* Extract `shell_count_argc` from `mch_call_shell`Thiago de Arruda2014-03-31
|
* Extract `shell_skip_word` from `mch_call_shell`Thiago de Arruda2014-03-31
|
* Split code for counting and parsing argumentsThiago de Arruda2014-03-31
|
* Add doxygen-style comments to functions from os/fs.c.Thomas Wienecke2014-03-31
|
* xrealloc(): similar to xmalloc()Felipe Oliveira Carvalho2014-03-31
| | | | | Replaced all calls to realloc by xrealloc. All `== NULL` tests can be removed and the code within `!= NULL` tests can be unwrapped.
* Document xmalloc() and deprecate lalloc()Felipe Oliveira Carvalho2014-03-31
|
* Avoid a seg fault when exiting after OOM errorFelipe Oliveira Carvalho2014-03-31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | I'm debugging OOM behavior using http://www.nongnu.org/failmalloc/ on Linux. gdb environment: ``` set env LD_PRELOAD=libfailmalloc.so set env FAILMALLOC_SPACE=0xF00000 ``` SEGV was happening like this: ``` Starting program: /home/felipe/code/neovim/build/bin/nvim [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Vim: Error: Out of memory. Program received signal SIGSEGV, Segmentation fault. 0x00000000004d3719 in getout (exitval=1) at /home/felipe/code/neovim/src/main.c:836 836 if (*p_viminfo != NUL) (gdb) ``` After the fix it works as expected: ``` (gdb) set environment LD_PRELOAD=libfailmalloc.so (gdb) set environment FAILMALLOC_SPACE=0xF00000 (gdb) r Starting program: /home/felipe/code/neovim/build/bin/nvim [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Vim: Error: Out of memory. Vim: Finished. [Inferior 1 (process 21505) exited with code 01] (gdb) ```
* xmalloc() that succeeds or gracefully abortsFelipe Oliveira Carvalho2014-03-31
|
* Re-integrate FEAT_SIGNS code, close #383Justin M. Keyes2014-03-29
| | | | | - omit FEAT_NETBEANS_INTG and FEAT_SIGN_ICONS - omit FEAT_GUI blocks
* use function attributes on allocation functionsNicolas Hillegeer2014-03-29
| | | | | This should allow gcc/clang/icc to generate better code and clarifies function intent.
* add func_attr.h to be able to use func attribsNicolas Hillegeer2014-03-29
| | | | | Currently enables these for every gnu-like compilers, i.e.: gcc, clang and icc.
* Move remove_duplicates to garray.cJohn Schmidt2014-03-29
|
* Extract path.c from misc1.cJohn Schmidt2014-03-29
|
* Address clint warnings and other style issues.Thomas Wienecke2014-03-27
| | | | | | | * Add const. * Fix conditions (move && from end to start of line). * Use int32_t instead of long. * Use //-style comments.
* Rename mch_* functions to os_* in os module.Thomas Wienecke2014-03-27
|
* Fix prepare watcher teardown in the event loopThiago de Arruda2014-03-27
|
* Re-integrate FEAT_FILTERPIPE codeThiago de Arruda2014-03-27
| | | | | | | | | | | | | This feature was accidentally removed when doing the initial import from vim. It makes vim use pipes instead of temporary files for filtering buffers through shell commands. I found that this was missing when looking for references of SHELL_READ/SHELL_WRITE outside mch_call_shell`. When `mch_call_shell` is reimplemented on top of libuv process management facilities, pipes will always be used for communication with child processes so it makes sense to enable the feature permanently.
* Make `EventType` enum private to input.cThiago de Arruda2014-03-27
| | | | | This enum doesn't need to be public since `event_poll` is only interested in user input(but other events may be handled by libuv callbacks).
* Cleanup of indent.c to make it comply with style guide.John2014-03-27
| | | | | | | | | | | | - Uncrustify indent.c - Fix comments - Fix whitespace - Format get_lisp_indent (needs to be refactored completely) - Fix endline comments - Change to postincrement - Fix clint.py errors - Move static function last - Fix blank endlines
* Reimplement input functions on top of libuvThiago de Arruda2014-03-26
| | | | | | | | | | | | | | | | The functions `mch_inchar`, `mch_breakcheck`, `mch_char_avail` were reimplemented on top of libuv. Here's how it works: - When Neovim needs to wait for characters, it will transfer control to libuv event loop. - When the libuv event loop gets user input, it will transfer control back to Neovim - Neovim uses the `input_read` function to get the actual data read by libuv. With this scheme its possible to keep Neovim single-threaded while enjoying the benefits provided by libuv. This commit leaves SIGWINCH broken for now
* Use realloc instead of vim_reallocFelipe Oliveira Carvalho2014-03-26
|
* HAVE_AVAIL_MEM is always undefined. Remove tests and dead codeFelipe Oliveira Carvalho2014-03-26
| | | | | Defining HAVE_AVAIL_MEM doesn't even build. The code tries to call `mch_avail_mem` which is not defined.
* Remove MEM_PROFILE related codeFelipe Oliveira Carvalho2014-03-26
| | | | | | | | | Code around `#ifdef MEM_PROFILE` was used to profile vim's memory comsumption. It's very likely broken as new code is using malloc() and free() directly. In this day and age, valgrind can solve in a much reliable way what this code was trying to do.
generated by cgit (git 2.34.1) at 2025-06-26 17:00:18 +0000