Commit Graph

319 Commits

Author SHA1 Message Date
MichaHoffmann
affef0aab8 add func_attr_printf in :
log.c
message.c
strings.c

fixed some printf warnings in:

src/nvim/undo.c
src/nvim/eval.c
src/nvim/eval/encode.c
src/nvim/eval/typval.c
src/nvim/ex_getln.c
src/nvim/fileio.c
src/nvim/lua/executor.c
src/nvim/main.c
src/nvim/regexp_nfa.c
src/nvim/shada.c
src/nvim/spellfile.c
src/nvim/tui/terminfo.c
src/nvim/garray.h
2018-09-24 21:28:04 +02:00
Justin M. Keyes
a02d22cca8 IO: shada should respect 'fsync' option
shada_write_file() is called on exit (:quit and friends), this can be
very slow.

Note: AFAICT Vim (do_viminfo()) does not appear to fsync() viminfo.
2018-04-21 12:50:28 +02:00
ZyX
bdf5f57989 shada: Fix conversion warnings 2018-04-01 23:37:23 +03:00
ZyX
dd1b493f75 shada: Fix some memory leaks and completely ignore numbered mark names
Problems:
- In two places in shada_read_when_writing() memory just was not freed. Both 
  places were verified to cause test failures.
- Numbered marks got assigned incorrect (off-by-one compared to position in the 
  array) numbers in replace_numbered_mark.
- It was possible to have non-continuously populated array of numbered marks 
  which messed up code for merging them.

(Note about tests: marks with additional data are always compared different when 
merging, that caused some confusion regarding why test did not work the way 
I expected.)
2018-04-01 21:29:47 +03:00
ZyX
f5373e2cdc shada: Add functions to format ShaDa entries for debugging purposes
To be used in debugging printfs.
2018-04-01 21:23:43 +03:00
ZyX
200898546e shada: When storing numeric marks reset the numbers
Attempt to fix observed crash. Crash currently not reproduced.
2018-04-01 20:05:19 +03:00
ZyX
3df11cfbca Revert "shada: In place of ignoring cursor position with lnum 0 save with 1"
This reverts commit aa728798b4.
2018-03-27 03:03:02 +03:00
ZyX
aa728798b4 shada: In place of ignoring cursor position with lnum 0 save with 1 2018-03-27 01:35:21 +03:00
ZyX
30e7fb2e32 shada: Also filter out invalid cursor position when writing '0' mark
Based on https://github.com/neovim/neovim/pull/5908#issuecomment-375909903, but 
with adjusted condition as line number or column less then zero should not 
appear at all based on what I know.
2018-03-27 01:33:59 +03:00
ZyX
7941aaa3bf macros: Rename LAST_ARRAY_ENTRY to ARRAY_LAST_ENTRY 2018-03-27 01:33:59 +03:00
ZyX
0bfc91be96 shada: Make ignore_buf also inline 2018-03-27 01:33:59 +03:00
ZyX
607e0d2202 shada: Save numbered marks
Problems so far:

- Marks in the current instance are not adjusted.
- Duplicates are not removed (not that it works in Vim either now, not at 
  8.0.134 at least).
2018-03-26 00:38:29 +03:00
Justin M. Keyes
7ae4144208 refactor/rename: path_try_shorten_fname() 2018-03-24 14:17:40 +01:00
Daniel Hahler
f588113191 shada: fix typo (#8066) 2018-02-25 20:14:32 +01:00
ZyX
9ea1752d60 *: Provide list length when allocating lists 2018-01-14 01:33:16 +03:00
ZyX
6b45dbca04 mark: Make sure that jumplist item will not have zero lnum
Fixes #7169
2017-12-24 00:32:43 +03:00
ZyX
f4132fb38b *: Fix linter errors 2017-12-11 10:19:20 +03:00
ZyX
ac4bbf55f6 *: Hide list implementation in other files as well 2017-12-10 22:04:43 +03:00
ZyX
63f72ac27c shada: Fix linter error 2017-07-04 17:47:45 +03:00
ZyX
91b9ad7d82 shada: Make sure that code does not attempt to read too long items
Fixes #6957
2017-07-04 15:41:59 +03:00
AdnoC
2f2eeb19ba shada: Add default value for is_unnamed 2017-05-31 13:31:06 -04:00
AdnoC
0c3dea5c4d style/lint 2017-05-31 13:31:01 -04:00
AdnoC
336412e1db shada/linting: Moved some code out of shada_write.
shada_write was too long (over 500 lines) and caused a linting error.
Register initialization was moved to its own function in order to save lines.
2017-05-31 13:19:09 -04:00
AdnoC
a00b03d03f shada: Set the unnamed register to the previous unnamed register on startup 2017-05-31 13:18:59 -04:00
ZyX
b2265a0977 shada: Fix V581: adjacent branches with same condition 2017-05-20 04:29:54 +03:00
James McCoy
5ec72aadbf *: Use __attribute__((fallthrough)) where comments aren't supported
Although GCC now detects possibly unintentional fall through, there
rules around which the comments are detected are rather strict.  In
cases where a comment isn't detected, upstream [recommends] using their
fallthrough attribute.

[recommends]: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77817#c11
2017-05-12 17:07:25 -04:00
ZyX
c2f3e361c5 *: Add comment to all C files 2017-04-19 19:11:50 +03:00
ZyX
7701014b65 *: Remove useless asserts 2017-04-08 20:22:46 +03:00
ZyX
e814695223 eval/*code,shada: Drop support for converting UTF-8 from/to p_enc
Not needed any longer since p_enc is always utf-8.
2017-04-08 19:20:41 +03:00
ZyX
b3587a456b shada: Initialize vimconv_T 2017-04-08 19:00:05 +03:00
ZyX
e586047a53 eval/decode,shada: Do not forget to clean up converters 2017-04-08 19:00:04 +03:00
ZyX
b10880dadc eval: Make writefile() able to disable fsync() 2017-04-02 22:11:35 +03:00
ZyX
2ad4fba46d eval: Move copy_tv to eval/typval 2017-03-29 10:08:06 +03:00
ZyX
28dafe3ff0 eval,*: Move get_tv_string to typval.c
Function was renamed and changed to return `const char *`.
2017-03-29 10:08:05 +03:00
ZyX
e18a578308 *: Move some dictionary functions to typval.h and use char*
Also fixes buffer reusage in setmatches() and complete().
2017-03-29 10:07:42 +03:00
ZyX
fb146e80aa eval: Split eval.c into smaller files 2017-03-29 10:05:06 +03:00
ZyX
5ac5c7971c coverity/133898: Fix sizeof argument 2017-01-23 02:16:16 +03:00
James McCoy
492f2cfeff vim-patch:7.4.1925
Problem:    Viminfo does not merge file marks properly.
Solution:   Use a timestamp.  Add the :clearjumps command.

2d35899721
2017-01-09 20:23:41 -05:00
ZyX
10d9c6d813 shada: Mark ALWAYS_INLINE function as inline 2017-01-06 01:47:07 +03:00
ZyX
ac50971f87 shada: Move setpcmark()/cleanup_jumplist() out of the cycle 2017-01-04 16:55:28 +03:00
ZyX
ec975a74ad shada: Fix offset saving 2017-01-04 16:42:43 +03:00
ZyX
40cedfd0f5 shada: Silence linter 2017-01-04 16:37:31 +03:00
ZyX
81ef9f55ae shada: Save current cursor position before saving jumps 2017-01-04 16:34:06 +03:00
James McCoy
966be42a2b shada: Move '%' handling from shada_write to its own function
shada_get_buflist finds all the relevant buffers and generates the
kSDItemBufferList ShadaEntry.
2016-12-14 06:56:48 -05:00
James McCoy
7eb91c781b shada: Respect the optional buffer count for shada-%
Closes #5759
2016-12-12 14:35:09 -05:00
James McCoy
ca292c9768 Avoid serializing NULL string through msgpack
Attempting to serialize a NULL string through msgpack results in
msgpack_sbuffer_write attempting to memcpy from a NULL pointer, which is
undefined behavior.
2016-11-23 20:30:00 -05:00
Justin M. Keyes
df072c3b2b refactor: eliminate misc2.c
move `call_shell` to misc1.c
Move some fns to state.c
Move some fns to option.c
Move some fns to memline.c
Move `vim_chdir*` fns to file_search.c
Move some fns to new module, bytes.c
Move some fns to fileio.c
2016-09-13 16:20:09 +02:00
ZyX
6127eaef05 shada: Fix non-writeable ShaDa directory handling
Before this change,

    nvim -i /etc/shada

segfaults on exit if the file does not exist and user does not have
permissions to create the file at /etc/shada.

Closes #5296
Reported in #5277
https://github.com/neovim/neovim/issues/5277#issuecomment-243937255
2016-09-04 21:45:48 -04:00
ZyX
5fc72882cc option: Also escape commas in options other then &runtimepath 2016-07-10 02:04:27 +03:00
ZyX
75bd39d49c *: Satisfy linter (newest type casts rule) 2016-06-24 00:07:56 +03:00