Problem: qsort() comparison functions should be transitive
Solution: Do not subtract values, but rather use explicit comparisons
Improve qsort() comparison functions
There has been a recent report on qsort() causing out-of-bounds read &
write in glibc for non transitive comparison functions
https://www.qualys.com/2024/01/30/qsort.txt
Even so the bug is in glibc's implementation of the qsort() algorithm,
it's bad style to just use substraction for the comparison functions,
which may cause overflow issues and as hinted at in OpenBSD's manual
page for qsort(): "It is almost always an error to use subtraction to
compute the return value of the comparison function."
So check the qsort() comparison functions and change them to be safe.
closes: #13980
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: TextChanged not triggered for :norm! commands
(machakann, after v9.0.2031)
Solution: Only reset curbuf->b_last_changedtick if TextChangedI
was triggered in insert mode (and not blocked)
Note: for unknown reasons, the test fails on Windows (but seems to work
fine when running interactively)
fixes: #13967closes: #13984
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Restore lastused_tabpage too early in do_arg_all() function it
will change later in the function.
Solution: Restore lastused_tabpage a bit later, when being done with
tabpages (glepnir)
closes: #13992
Signed-off-by: glepnir <glephunter@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Problem when scrolling using slow touchpads scroll event
Solution: better ways to determine if a smooth scroll has ended (when
available) (lilydjwg)
related: #13997
Signed-off-by: lilydjwg <lilydjwg@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: X11 scroll size changes after accessing clipboard
(Ernie Rael)
Solution: use GDK_SCROLL_MASK for X11 and GDK_SMOOTH_SCROLL_MASK for
Wayland (lilydjwg)
because GDK_SCROLL_SMOOTH events don't work well for x11 (see comments).
fixes#13994closes: #13997
Signed-off-by: lilydjwg <lilydjwg@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Visual hl wrong when it ends before multibyte 'showbreak'.
(lacygoil)
Solution: Use vcol_sbr instead of adding n_extra.
(zeertzjq)
fixes: #11272closes: #13996
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Redrawing can be improved when inserting/deleting lines with 'number'.
Solution: Only redraw the number column of lines below changed lines.
Add a test as this wasn't previously tested.
(zeertzjq)
closes: #13985
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Redrawing can be improved when deleting lines with 'cursorline'.
Solution: Use smarter invalidation and adjustment. Remove unnecessary
UPD_VALID as it is already set at the top of the loop. Make
the test for #4862 fail without the fix.
(zeertzjq)
closes: #13986
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: X11 mouse-scrolling stutters
(Ron Aaron, after 9.1.0064)
Solution: Handle GDK_SCROLL_SMOOTH fractional distance events
(lilydjwg)
I don't know why, but with GDK_SMOOTH_SCROLL_MASK we get wheel events as
GDK_SCROLL_SMOOTH. What's more, one wheel scroll is counted as 1.5
distance in Wayland but 1.0 in X11.
I failed to find any docs on gtk.org about this.
fixes: #13987closes: #13991
Signed-off-by: lilydjwg <lilydjwg@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: unexpected error for modifying final list using += operator
(Ernie Rael)
Solution: Allow List value modification of a final variable using +=
operator
(Yegappan Lakshmanan)
fixes: #13745fixes: #13959closes: #13962
Signed-off-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: LineNrAbove and LineNrBelow background wrong on wrapped lines.
Solution: Update number column also for wrapped part of a line.
(zeertzjq)
closes: #13974
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: GTK3: using wrong style for pre-edit area
Solution: remove the widget name, adjust css
(lilydjwg)
closes: #13972
Signed-off-by: lilydjwg <lilydjwg@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: The call to redraw_for_cursorline() in nv_mousescroll() is
unnecessary because redraw_for_cursorline() only sets redraw
type to UPD_VALID, and all code paths in do_mousescroll()
already set redraw type to at least UPD_VALID.
Solution: Remove call to redraw_for_cursorline() in nv_mousescroll().
(zeertzjq)
closes: #13979
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: insert completion not correct when adding new leader
Solution: Reset compl_curr_match to compl_shown_match
(glepnir)
closes: #13957
Co-authored-by: glepnir <glephunter@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: did_set_breakat() should be in optionstr.c as 'breakat' is a
string option.
Solution: Move did_set_breakat() to optionstr.c.
(zeertzjq)
closes: #13958
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Looping over modifier_keys_table[] unnecessarily with only
MOD_MASK_ALT or MOD_MASK_CMD, as modifier_keys_table[] only
contains MOD_MASK_SHIFT and MOD_MASK_CTRL, and the loop won't
do anything.
Solution: Remove MOD_MASK_ALT and MOD_MASK_CMD from the condition.
(zeertzjq)
closes: #13963
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Not able to build without FEAT_DIFF
(John Marriott, after 9.1.0071)
Solution: Adjust #ifdefs
(Yegappan Lakshmanan)
closes: #13964
Signed-off-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Need a diff() Vim script function
Solution: Add the diff() Vim script function using the
xdiff internal diff library, add support for
"unified" and "indices" mode.
(Yegappan Lakshmanan)
fixes: #4241closes: #12321
Signed-off-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: CI: testsuite not run on M1 Mac
Solution: Make it run on gh runners for M1, disable failing tests for
now, until we figure the problem with the failings tests out
(rhysd)
closes: #13943
Signed-off-by: rhysd <lin90162@yahoo.co.jp>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: ScreenLines may not be correctly initialized, causing hang
(Olaf Seibert, after 9.0.0220)
Solution: always initialize ScreneLines when allocating a screen
(Olaf Seibert)
ScreenLines and related structures could be left uninitialized
causing a screen update to run into an infinite loop when using latin1
encoding.
Partly caused because by patch 9.0.0220, which makes mb_ptr2len return
zero for NUL
related: #12671closes: #13946
Signed-off-by: Olaf Seibert <rhialto@falu.nl>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Visual highlighting can still be improved
Solution: Update Visual highlighting for 8 color terminals,
use uniform grey highlighting for dark and light bg
(Maxim Kim)
Update terminal Visual
1. Use `ctermbg=Grey ctermfg=Black` for both dark and light
This uniforms Visual highlighting between default dark and light colors
And should work for vim usually detecting light background for terminals
with black/dark background colors.
Previously used `ctermfg=White` leaks `cterm=bold` if available colors
are less than 16.
2. Use `term=reverse cterm=reverse ctermbg=NONE ctermfg=NONE`
for terminals reporting less than 8 colors available
If the terminal has less than 8 colors, grey just doesn't work right
closes: #13940
Signed-off-by: Maxim Kim <habamax@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: gcc still complains about use of uninitialized var
((Tony Mechelynck, after 9.1.0064/9.1.0066))
Solution: This time init the correct variable
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: gcc complains about use of uninitialized var
(Tony Mechelynck, after 9.1.0064)
Solution: initialize button to silence gcc
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Segfault with CompleteChanged autocommand
(markonm )
Solution: Test match->cp_prev for being NULL before accessing it
closes: #13929
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: No Wayland support
Solution: Add Wayland UI support
(lilydjwg)
closes: #9639
Signed-off-by: lilydjwg <lilydjwg@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: GTK code can be improved
Solution: Improve GTK code for initial Wayland support
(lilydjwg)
related: #9639
Signed-off-by: lilydjwg <lilydjwg@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Internal error when :luado/perldo/pydo etc delete lines
Solution: Test that the line is still valid line number
(zeertzjq)
closes: #13931
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: UX of visual highlighting can be improved
Solution: Improve readibility of visual highlighting,
by setting better foreground and background
colors
The default visual highlighting currently is nice in that it overlays
the actual syntax highlighting by using a separate distinct background
color.
However, this can cause hard to read text, because the contrast
between the actual syntax element and the background color is way too
low. That is an issue, that has been bothering colorschemes authors for
quite some time so much, that they are defining the Visual highlighting
group to use a separate foreground and background color, so that the
syntax highlighting vanishes, but the text remains readable (ref:
vim/colorschemes#250)
So this is an attempt to perform the same fix for the default Visual
highlighting and just use a default foreground and background color
instead of using reverse.
I also removed the hard-coded changes to the Visual highlighting in
init_highlight. It's not quite clear to me, why those were there and not
added directly to the highlighting_init_<dark|light> struct.
closes: #13663
related: vim/colorschemes#250
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Recorded register cannot be translated using keytrans() when
it involves character search (iddqd505)
Solution: Record a K_IGNORE instead of a K_NOP (zeertzjq)
related: #13916closes: #13925
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: No event is triggered before creating a window.
(Sergey Vlasov)
Solution: Add the WinNewPre event (Sergey Vlasov)
fixes: #10635closes: #12761
Signed-off-by: Sergey Vlasov <sergey@vlasov.me>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Cannot map Super Keys in GTK UI
(Casey Tucker)
Solution: Enable Super Key mappings in GTK using <D-Key>
(Casey Tucker)
As a developer who works in both Mac and Linux using the same keyboard,
it can be frustrating having to remember different key combinations or
having to rely on system utilities to remap keys.
This change allows `<D-z>` `<D-x>` `<D-c>` `<D-v>` etc. to be recognized
by the `map` commands, along with the `<D-S-...>` shifted variants.
```vimrc
if has('gui_gtk')
nnoremap <D-z> u
nnoremap <D-S-Z> <C-r>
vnoremap <D-x> "+d
vnoremap <D-c> "+y
cnoremap <D-v> <C-R>+
inoremap <D-v> <C-o>"+gP
nnoremap <D-v> "+P
vnoremap <D-v> "-d"+P
nnoremap <D-s> :w<CR>
inoremap <D-s> <C-o>:w<CR>
nnoremap <D-w> :q<CR>
nnoremap <D-q> :qa<CR>
nnoremap <D-t> :tabe<CR>
nnoremap <D-S-T> :vs#<CR><C-w>T
nnoremap <D-a> ggVG
vnoremap <D-a> <ESC>ggVG
inoremap <D-a> <ESC>ggVG
nnoremap <D-f> /
nnoremap <D-g> n
nnoremap <D-S-G> N
vnoremap <D-x> "+x
endif
```
closes: #12698
Signed-off-by: Casey Tucker <dctucker@hotmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: MS-Windows: Key event test still fails
Solution: Skip testing CTRL-C, since it causes an Interrupt
which causes a test failure (after 9.1.0053)
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Incorrect number of trailing spaces inserted for multibyte
characters when pasting a blockwise register in blockwise visual
mode (VanaIgr)
Solution: Skip over trailing UTF-8 bytes when computing the number of trailing
spaces (VanaIgr)
When pasting in blockwise visual mode, and the register type is <CTRL-V>, Vim
aligns the text after the replaced area by inserting spaces after pasted
lines that are shorter than the longest line. When a shorter line contains
multibyte characters, each trailing UTF-8 byte's width is counted in addition
to the width of the character itself. Each trailing byte counts as being 4
cells wide (since it would be displayed as <xx>).
closes: #13909
Signed-off-by: VanaIgr <vanaigranov@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: formatting long lines is slow
(kawaii-Code)
Solution: optimize gq (internal_format) for long
lines (kawaii-Code)
Implemented two workarounds that significantly reduce
the amount of pointless calls. Ideally the algorithm
would be rewritten not to be n^2, but it's too complicated
with too many corner cases.
closes: #13914
Signed-off-by: kawaii-Code <nia.personal.0@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Patch 9.1.0041 causes regressions for users
(Gleb Fotengauer-Malinovskiy)
Solution: Revert 9.1.0041 and restore old behaviour
Revert "patch 9.1.0041: xxd -i may generate incorrect C statements"
This reverts commit 7062be1312.
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Win32 Keyboard handling is sub-optimal and has
many issues with international keyboards
(risa2000) (after v8.2.4807)
Solution: soft-rollback the change, allow the user to select
a suitable key-translation strategy
(Anton Sharonov)
Read about the details in the help:
:h w32-experimental-keycode-trans-strategy
fixes: #10615fixes: #12595closes: #12752
Signed-off-by: Anton Sharonov <anton.sharonov@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: E95 is possible if a buffer called "[Command Line]" already
exists when opening the cmdwin. This can also happen if the
cmdwin's buffer could not be deleted when closing.
Solution: Un-name the cmdwin buffer, and give it a special name instead,
similar to what's done for quickfix buffers and for unnamed
prompt and scratch buffers. As a result, BufFilePre/Post are
no longer fired when opening the cmdwin. Add a "command" key
to the dictionary returned by getbufinfo() to differentiate
the cmdwin buffer instead. (Sean Dewar)
NOTE: This is technically a breaking change... maybe this needs a different
solution? (Or maybe this issue can be ignored...)
A GitHub search reveals some plugins expect the old behaviour. However, many of
those plugins also do not seem to account for the string being translated, so
they are subtly broken anyway (not withstanding the fact that you can call any
old buffer "[Command Line]" too...)
closes: #12819
Signed-off-by: Sean Dewar <seandewar@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Autocmds triggered from opening the cmdwin (in win_split and
do_ecmd) can cause issues such as E199, as the current checks
are insufficient.
Solution: Commands executed from the cmdwin apply to the old curwin/buf,
so they should be kept in a "suspended" state; abort if
they've changed. Also abort if cmdwin/buf was tampered with,
and check that curwin is correct. Try to clean up the cmdwin
buffer (only if hidden and non-current to simplify things; the
same approach is used when closing cmdwin normally), and add a
beep. (Sean Dewar)
It'd be nice to also check that curwin was *really* created by win_split, as
autocommands can change curwin before it returns (so it can't be assumed to be
that of the split); for now, this means that the cmdwin may not be the botwin in
that case, which is probably OK.
closes: #12819
Signed-off-by: Sean Dewar <seandewar@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Things that temporarily change/restore curwin/buf (e.g:
win_execute, some autocmds) may break assumptions that
curwin/buf is the cmdwin when "cmdwin_type != 0", causing
issues.
Solution: Expose the cmdwin's real win/buf and check that instead. Also
try to ensure these variables are NULL if "cmdwin_type == 0",
allowing them to be used directly in most cases without
checking cmdwin_type. (Sean Dewar)
Alternatively, we could ban win_execute in the cmdwin and audit all places that
temporarily change/restore curwin/buf, but I didn't notice any problems arising
from allowing this (standard cmdwin restrictions still apply, so things that may
actually break the cmdwin are still forbidden).
closes: #12819
Signed-off-by: Sean Dewar <seandewar@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: :drop does not re-use empty buffer
(Rocco Mao)
Solution: Make :drop re-use an empty buffer
(Rocco Mao)
fixes: #13851closes: #13881
Signed-off-by: Rocco Mao <dapeng.mao@qq.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: --remote-silent applies the wildignore option
to each argument, which may result in "E479: No match"
(hebaronson)
Solution: temporarily reset 'wildignore' setting when building
the :drop command
closes: #13835
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: po Makefiles can be improved
Solution: Improve the style of the Makefiles, update
Makefile variables, update documentation
(RestorerZ)
closes: #13858closes: #13857
Signed-off-by: RestorerZ <restorer@mail2k.ru>
Signed-off-by: Christian Brabandt <cb@256bit.org>
