aboutsummaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAge
...
| * shada: Only check errno if read/write returned -1ZyX2015-10-08
| | | | | | | | | | | | | | According to the manual (POSIX) this is the only case when errno is set by these functions. This is needed because some functions (e.g. buflist_new) leave errno set to non-zero value under some conditions (e.g. when opening non-existing files).
| * shada: Put pointer to the close function into reader/writer structureZyX2015-10-08
| |
| * shada: Handle all possible msgpack_unpacker_next returnsZyX2015-10-08
| |
| * viminfo: Remove old codeZyX2015-10-08
| |
| * shada: Add support for merging everything like described in the docZyX2015-10-08
| |
| * shada: Do not copy history entries when writingZyX2015-10-08
| |
| * shada: Do not save search pattern/sub string when it is olderZyX2015-10-08
| |
| * shada: Use os_strerror in place of strerror in some placesZyX2015-10-08
| |
| * shada: Use os_mkdir_recurse to create directory for ShaDa fileZyX2015-10-08
| |
| * shada: Refactor history merger to use statically sized 2linked listZyX2015-10-08
| | | | | | | | | | | | | | Also fixes the error: when writing shada files existing history items may be freed. Warning: valgrind reports some memory leaks.
| * shada: Refactor history merging support to separate functionsZyX2015-10-08
| |
| * shada: Preserve existing file permissionsZyX2015-10-08
| |
| * shada: Populate v:oldfilesZyX2015-10-08
| |
| * shada: Save v:hlsearch valueZyX2015-10-08
| |
| * main: Silence -Wunused-result in main.c for shada_read_fileZyX2015-10-08
| |
| * shada: Add more size_t castsZyX2015-10-08
| | | | | | | | As usual, GCC is overparanoid with -Wconversion.
| * Deprecate &viminfo and :[rw]v, add &shada and :[rw]shZyX2015-10-08
| |
| * Replace references to viminfo in various placesZyX2015-10-08
| |
| * shada: Use “write then rename” variant by default for writing shadaZyX2015-10-08
| |
| * shada: Test that history string does not contain zero byteZyX2015-10-08
| | | | | | | | | | | | | | This test can safely be ignored for other kinds of strings, but history string contains separator after the last NUL byte and string containing NUL byte inside thus means that separator will be placed at one position, but seeked at another.
| * shada: Add support for converting strings that are read or writtenZyX2015-10-08
| |
| * shada: Refactor file reading/writing to use os_openZyX2015-10-08
| |
| * shada,functests: Refactor shada items skippingZyX2015-10-08
| |
| * shada: Reassign some entry types to WantInfo and WantMarksZyX2015-10-08
| | | | | | | | | | kShaDaWantMarks now loads only local file change list and file-local marks. kShaDaWantInfo now loads global marks and jump list.
| * shada: Add support for dumping and restoring buffer listZyX2015-10-08
| |
| * shada: Use pre-populated set of buffers on removable mediaZyX2015-10-08
| | | | | | | | | | Should be faster then constantly running shada_removable for each buffer many times.
| * shada: Do not record removable buffers and marksZyX2015-10-08
| |
| * shada: Add support for dumping/restoring bufs changes and win jumpsZyX2015-10-08
| |
| * shada: Initial support for merging historyZyX2015-10-08
| | | | | | | | Currently only merges history when reading ShaDa file. No tests yet.
| * shada: Skip reading some items if neededZyX2015-10-08
| | | | | | | | Avoids unneeded malloc()s/free()s and lots of code for parsing.
| * main: Read marks at startupZyX2015-10-08
| | | | | | | | | | I do not know why marks were not read originally, but without this change marks are not going to be read at startup when Vim starts without arguments.
| * shada: Use glibc functions to convert from big endian, add a fallbackZyX2015-10-08
| |
| * viminfo: First version of ShaDa file dumpingZyX2015-10-08
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | What works: 1. ShaDa file dumping: header, registers, jump list, history, search patterns, substitute strings, variables. 2. ShaDa file reading: registers, global marks, variables. Most was not tested. TODO: 1. Merging. 2. Reading history, local marks, jump and buffer lists. 3. Documentation update. 4. Converting some data from &encoding. 5. Safer variant of dumping viminfo (dump to temporary file then rename). 6. Removing old viminfo code (currently masked with `#if 0` in a ShaDa file for reference).
| * mark: Fix valgrind error in mark.cZyX2015-10-08
| | | | | | | | Caused by using memcpy for assigning one structure to another.
| * os/fs: Define os_strerror as an alias to uv_strerrorZyX2015-10-08
| | | | | | | | It is not needed to know that os/* uses libuv.
* | tui: Flush input on timer_cbThiago de Arruda2015-10-12
|/ | | | | | This is required to have <esc> working correctly when the `ttimeout`/`ttimeoutlen` options are set. Ref: https://github.com/neovim/neovim/issues/2093#issuecomment-146506480
* api: represent api type String as msgpack type STR. closes #1250Björn Linse2015-10-08
|
* tui: fix_terminfo: Handle "tmux" like "screen". #3422Daniel Hahler2015-10-04
| | | | Closes #3399
* menu: support :emenu invoked by cmenu map.bambu2015-10-03
|
* ex_docmd: Fix terminal mode check condition for ex_normalThiago de Arruda2015-10-01
| | | | | | | | | Using the `curbuf->terminal` condition alone is wrong since it does not necessarily mean nvim is in terminal mode. It needs to be used with `State & TERM_FOCUS` because the current buffer could have changed without `terminal_enter` exiting. Close #3403
* tui: Fix queueing of large chunks of inputThiago de Arruda2015-10-01
| | | | | Take the return value of input_enqueue into consideration, waiting for the main thread to consume input. Close #1714 #3377.
* rbuffer: Enhance rbuffer_reset to work with filled RBuffersThiago de Arruda2015-10-01
|
* os/input: Don't advance past incomplete sequences in input_enqueueThiago de Arruda2015-10-01
| | | | | This allows callers to incrementally process buffers that are filled by incomplete chunks more easily.
* tui: Don't use 'pastetoggle' for automatic pastingThiago de Arruda2015-10-01
| | | | | | | Add a new special key that can be used by UIs to toggle the 'paste' option and use it in the TUI instead of the user's 'pastetoggle' value. Close #2843 #2092
* shell: Ensure silent bang mappings won't cut outputThiago de Arruda2015-10-01
| | | | | | | Setting `msg_didout` after `call_shell` is enough as it will cause `hit_return_msg()` to print on next line. Close #3269
* Merge pull request #3410 from oni-link/fix.spell_suggestJustin M. Keyes2015-10-01
|\ | | | | spell.c: Fix memory leak
| * spell.c: Fix memory leakoni-link2015-10-01
| |
* | Merge pull request #3407 from oni-link/fix.ga_concatJustin M. Keyes2015-10-01
|\ \ | |/ |/| garray.c: Prevent ga_concat() using memcpy(NULL,...)
| * garray.c: Prevent ga_concat() using memcpy(NULL,...)oni-link2015-09-30
| | | | | | | | | | | | Calling ga_grow(gap, 0) does not reallocate memory for garray gap. Because of this, gap->ga_data can be NULL after such a call, if gap does not have memory allocated.
* | eval/string(): Share code between msgpackdump() and string()+:echoZyX2015-09-30
| |
generated by cgit (git 2.34.1) at 2025-07-19 03:15:42 +0000