bug#80837: treesit-forward-comment off-by-one issue: overshoots by one character

> On Aug 18, 2026, at 11:53 AM, Juri Linkov <[email protected]> wrote:
> 
>>> I'm very sorry to reopen it, but we still have the same problem
>>> in the emacs-31 release branch.  For example, try in c-ts-mode to
>>> uncomment a comment, and it leaves the trailing */ comment endings.
>> 
>> I'm not sure this is important enough to block the release, but we can
>> at least discuss whether we think there is a possible small,
>> non-invasive for the emacs-31 branch.  Yuan, sorry to go back and forth
>> so many times here, but do you think there is a way to do it less
>> invasively?
> 
> Leaving emacs-31 unfixed makes c-ts-mode somewhat unusable
> since uncommenting is a frequent operation.  If a small fix
> for treesit-forward-comment is still unfeasible, then at least
> we need to disable the 'comment' thing for c-ts-mode
> in the release branch:
> 
> diff --git a/lisp/progmodes/c-ts-mode.el b/lisp/progmodes/c-ts-mode.el
> index 5f064716a89..3efbee6f8a7 100644
> --- a/lisp/progmodes/c-ts-mode.el
> +++ b/lisp/progmodes/c-ts-mode.el
> @@ -1275,8 +1275,7 @@ c-ts-mode--thing-settings
>                     "goto_statement"
>                     "case_statement")))
>     (text ,(regexp-opt '("comment"
> -                         "raw_string_literal")))
> -    (comment "comment"))
> +                         "raw_string_literal"))))
>   "`treesit-thing-settings' for both C and C++.")
> 
> ;;; Support for FOR_EACH_* macros
> 

Actually this makes a lot of sense. I don’t think we are missing anything by not using tree-sitter for the comment thing, forward/backward-comment already works perfectly fine without it. We can also add a comment to explain why we leave the comment thing out for c-ts-mode in the release version.

Yuan