Problem: Incsearch highlighting is attempted halfway a mapping.
Solution: Only do incsearch highlighting if keys were typed or there is no
more typeahead.
ccb148ac63
Problem: Get E12 in a job callback when searching for tags. (Andy Stewart)
Solution: Use the sandbox only for executing a command, not for searching.
(closesvim/vim#8511)
547f94f330
N/A patches for version.c:
vim-patch:8.2.3164: MS-Windows: reported version lacks patchlevel
Problem: MS-Windows: reported version lacks patchlevel, causing some update
tools to update too often. (Klaus Frank)
Solution: Add the patchlevel to the version. (Christian Brabandt)
0894e0d808
vim-patch:8.2.3192: build failure with small version
Problem: Build failure with small version (Tony Mechelynck).
Solution: Remove stray #ifdef.
11d7e62f1d
vim-patch:8.2.3208: dynamic library load error does not mention why it failed
Problem: Dynamic library load error does not mention why it failed.
Solution: Add the error message. (Martin Tournoij, closesvim/vim#8621)
1a3e5747b7
vim-patch:8.2.3214: MS-Windows: passing /D does not set the install location
Problem: MS-Windows: passing /D does not set the install location.
Solution: Adjust how the installer uses $VIM. Update the documentation.
(Christian Brabandt, closesvim/vim#8605)
7d60384a00
vim-patch:8.2.3231: build failure with small features
Problem: Build failure with small features.
Solution: Adjust #ifdef.
9088784972
vim-patch:8.2.3243: MS-Windows: "edit with multiple Vim" choice is less useful
Problem: MS-Windows: the "edit with multiple Vim" choice is not that
useful.
Solution: Change it to "Edit with multiple tabs". (Michael Soyka,
closesvim/vim#8645)
83cd0156e0
vim-patch:8.2.3247: using uninitialized memory when checking for crypt method
Problem: Using uninitialized memory when checking for crypt method.
Solution: Check the header length before using the salt and seed.
77ab4e28a2
vim-patch:8.2.3250: MS-Windows: cannot build with libsodium
Problem: MS-Windows: cannot build with libsodium.
Solution: Change FEAT_SODIUM into HAVE_SODIUM. (Christian Brabandt,
closesvim/vim#8668, closesvim/vim#8663)
1790be6cb6
vim-patch:8.2.3253: channel test fails randomly
Problem: Channel test fails randomly.
Solution: Add a sleep after sending the "echoerr" command. (Michael Soyka)
890ee4e2be
vim-patch:8.2.3260: build failure with small features
Problem: Build failure with small features.
Solution: Add #ifdef.
335c8c7b20
Problem: Coverity complains about free_wininfo() use.
Solution: Add a condition that "wip2" is not equal to "wip". (Neovim vim/vim#14996)
b5098060f4
This fix came from https://github.com/neovim/neovim/pull/14996.
This commit adds only a comment to be in sync with Vim.
N/A patches for version.c:
vim-patch:8.2.3063: crash when switching 'cryptmethod' to xchaha20 with undo file
Problem: Crash when switching 'cryptmethod' to xchaha20 with an existing
undo file. (Martin Tournoij)
Solution: Disable reading undo file when decoding can't be done inplace.
(issue vim/vim#8467)
65aee0b714
vim-patch:8.2.3101: missing function prototype for vim_round()
Problem: Missing function prototype for vim_round().
Solution: Add the prototype.
67b17a6fc6
vim-patch:8.2.3119: compiler warning for unused argument
Problem: Compiler warning for unused argument.
Solution: Add UNUSED.
6a9e5c69cf
vim-patch:8.2.3120: crypt with sodium test fails on MS-Windows
Problem: Crypt with sodium test fails on MS-Windows.
Solution: Make the tests pass. (closesvim/vim#8428)
db86472770
vim-patch:8.2.3131: MS-Windows: ipv6 channel test is very flaky in the GUI
Problem: MS-Windows: ipv6 channel test is very flaky in the GUI.
Solution: Skip the test.
981217c11f
vim-patch:8.2.3140: MS-Windows: ipv6 channel test is very flaky also without GUI
Problem: MS-Windows: ipv6 channel test is very flaky also without the GUI.
Solution: Skip the test also without the GUI.
482d2f37a5
vim-patch:8.2.3157: crypt test may fail on MS-Windows
Problem: Crypt test may fail on MS-Windows.
Solution: Ignore "[unix]" in the file message. (Christian Brabandt,
closesvim/vim#8561)
16e26a3116
vim-patch:8.2.3218: when using xchaha20 crypt undo file is not removed
Problem: When using xchaha20 crypt undo file is not removed.
Solution: Reset 'undofile' and delete the file. (Christian Brabandt,
closesvim/vim#8630, closesvim/vim#8467)
8a4c812ede
vim-patch:8.2.3245: the crypt key may appear in a swap partition
Problem: The crypt key may appear in a swap partition.
Solution: When using xchaha20 use sodium_mlock(). (Christian Brabandt,
closesvim/vim#8657)
131530a54d
Problem: win_gettype() does not recognize a quickfix window.
Solution: Add "quickfix" and "loclist". (Yegappan Lakshmanan, closesvim/vim#8676)
28d8421bfb
Problem: Location list window may open a wrong file.
Solution: Also update the qf_ptr field. (Wei-Chung Wen, closesvim/vim#8565,
closesvim/vim#8566)
1557b16dad
Problem: Location list only has the start position.
Solution: Make it possible to add an end position. (Shane-XB-Qian,
closesvim/vim#8393)
6864efa596
N/A patches for version.c:
vim-patch:8.2.3002: Vim doesn't abort on a fatal Tcl error
Problem: Vim doesn't abort on a fatal Tcl error.
Solution: Change emsg() to iemsg(). (Dominique Pellé, closesvim/vim#8383)
affd0bc626
vim-patch:8.2.3030: Coverity reports a memory leak
Problem: Coverity reports a memory leak.
Solution: Fix the leak and a few typos. (Dominique Pellé, closesvim/vim#8418)
cb54bc6562
Patch v8.2.3022 is mostly N/A but cannot be included here
because of new feature check for "has()".
vim-patch:8.2.3032: build problems with MSVC, other crypt issues with libsodium
Problem: Build problems with MSVC, other crypt issues with libsodium.
Solution: Adjust MSVC makefile. Disable swap file only when 'key' is set.
Adjust error message used when key is wrong. Fix Coverity issues.
(Christian Brabandt, closesvim/vim#8420, closesvim/vim#8411)
226b28b961
vim-patch:8.2.3044: Amiga MorphOS and AROS: process ID is not valid
Problem: Amiga MorphOS and AROS: process ID is not valid.
Solution: Use FindTask to return something which is unique to all processes.
(Ola Söder, closesvim/vim#8444)
3a62b14077
vim-patch:8.2.3046: Amiga MorphOS: Term mode is set using DOS packets
Problem: Amiga MorphOS: Term mode is set using DOS packets.
Solution: Use the same way of setting term mdoe on all next gen Amiga-like
systems. (Ola Söder, closesvim/vim#8445)
b420ac9d20
Problem: Using getchar() in Vim9 script is problematic.
Solution: Add getcharstr(). (closesvim/vim#8343)
3a7503c34c
Cherry-pick Test_getchar() changes from patch v8.1.2304
to sync with upstream.
Port f_getcharstr() to src/nvim/eval/funcs.c, not src/nvim/getchar.c.
Patch v8.1.2042 is not ported yet.
Remove syncolor.vim in favor of defining the default highlight groups
directly in `init_highlight`. This approach provides a number of
advantages:
1. The highlights are always defined, regardless of whether or not the
syntax regex engine is enabled.
2. Redundant sourcing of syntax files is eliminated (syncolor.vim was
often sourced multiple times based on how the user's colorscheme file
was written).
3. The syntax highlighting regex engine and the highlight groups
themselves are more fully decoupled.
4. Removal of the confusing `:syntax on` / `:syntax enable` dichotomy
(they now both do the same thing).
This approach also correctly solves a number of bugs related to
highlighting (#15176, #12573, #15205).
This fixes an issue introduced in #14771 (fix: source syncolors.vim
before startup scripts) that affected highlights for users who set
'background' to light in their startup script. Because syncolor.vim
checks for the value of &background, it was always setting up the 'dark'
background colors, which looked wrong for users using light backgrounds.
The primary benefit of #14771 is that it decoupled highlighting from the
syntax engine. This is useful for e.g. treesitter, which still makes use
of highlights even if the syntax engine is disabled. For this reason, it
is still worthwhile to source syncolor.vim separately from synload.vim,
which #14771 accomplishes. However, we should still source syncolor.vim
after the user startup scripts, to ensure that we are respecting the
options the user sets.
Another corollary benefit is that this reduces some redundancy in
highlight definitions, since we now only source syncolors.vim if the
user did not already enable a colorscheme.
lua is used as part of implementation for more core features. As an
example, every user keypress will invoke a lua function to check for
keypress handlers (regardless if they are registered or not). Thus not
starting lua until it is first used doesn't make much sense anymore.
nlua_enter was also needed due to the earlier stateful &rtp
translation, which by now have been made stateless.
Problem: When 'hidden' is set session creates extra buffers.
Solution: Move :badd commands to the end. (Jason Franklin)
d39e275b57
Adjust some tests in ex_cmds/mksession_spec.lua:
- 'restores same :terminal buf in splits': Buffers aren't always :badded
in the same order as they're :edited, :balted, etc, so the order of
buffers in the buffer list may change slightly now that :badd happens
afterwards.
- 'restores buffers with tab-local CWD': This is explained in a comment.
This fixes an issue (#12573) where colorscheme files are sourced twice
upon startup. This occurs when the startup script calls `:colorscheme`,
which sets the `g:colors_name` global variable. When syntax highlighting
is enabled in `syn_maybe_enable()` the `syntax.vim` script is sourced
which in turn sources `synload.vim`. This script checks to see if
`g:colors_name` is set and, if so, runs
exe "colors " . colors_name
This is done to ensure that highlight groups are defined before enabling
the syntax highlighting engine.
Instead, source syncolors.vim before the startup scripts which sets up
default highlights and only load the full syntax engine after
the startup scripts or when the user runs `:syntax on`. Add a guard
variable `did_syncolor` to prevent syncolor.vim from being sourced
twice and remove the line mentioned above from synload.vim so that
the colorscheme file is not re-sourced when the syntax engine is loaded.
