runtime(comment): consider &tabstop in lines after whitespace indent

The count `strlen()` in

```vim
line = printf(indent_start .. substitute(cms, '%s\@!', '%%', 'g'),
                       strpart(getline(lnum), strlen(indent_start)))
```

is too large if the block of lines to be operated on contains different
whitespace indenting (tab vs. spaces).

Considering using `2gcc` on the first line with 4 spaces as indenting
and on the next line using a single tab character (with &tabstop value
of 8):

Using `strlen(indent_start) = 4` for an initial indent of 4 spaces is
correct for the first line, but wrong for the next line and will
therefore wrongly comment out the tab-indented line (and possibly
deleting some content).

The new check is still too simple because it assumes that as soon as
there's a tab the whole indent is made of tabs; it's a start of entering
the mixed tab and whitespace indent rabbit hole.

fixes: #15797
closes: #15805

Signed-off-by: Konfekt <Konfekt@users.noreply.github.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
This commit is contained in:
Konfekt
2024-10-05 16:17:04 +02:00
committed by Christian Brabandt
parent 40c6babc17
commit 7b27fc49a8

View File

@ -1,7 +1,7 @@
vim9script
# Maintainer: Maxim Kim <habamax@gmail.com>
# Last Update: 2024-09-30
# Last Update: 2024 Oct 05
#
# Toggle comments
# Usage:
@ -58,9 +58,13 @@ export def Toggle(...args: list<string>): string
# handle % with substitute
line = printf(substitute(cms, '%s\@!', '%%', 'g'), getline(lnum))
else
# handle % with substitute
line = getline(lnum)
var indent_start_len = strlen(indent_start)
# handle % with substitute,
# consider different whitespace indenting
line = printf(indent_start .. substitute(cms, '%s\@!', '%%', 'g'),
strpart(getline(lnum), strlen(indent_start)))
strpart(line, (line[0 : strlen(indent_start_len) - 1] =~ '\t' ?
indent_start_len / &tabstop : indent_start_len)))
endif
else
line = substitute(getline(lnum), $'^\s*\zs{cms_l[0]} \?\| \?{cms_l[1]}$', '', 'g')