bug-gnu-emacs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#59158: [PATCH] vc-git: Ensure git doesn't try to colorize stashes


From: Stefan Kangas
Subject: bug#59158: [PATCH] vc-git: Ensure git doesn't try to colorize stashes
Date: Fri, 11 Nov 2022 21:00:00 -0800

[Please use "Reply to all" so that the discussion is recorded in the bug
tracker.]

Morgan Smith <Morgan.J.Smith@outlook.com> writes:

> Stefan Kangas <stefankangas@gmail.com> writes:
>
>> How can I reproduce this issue, starting from emacs -Q?
>
> I'm using emacs from commit d04433b962
>
>
> A little background first:
>
> So when in Emacs using eshell, git won't give us colored output.  In
> "emacs -Q" we can navigate using eshell to a git repo and type "git
> show HEAD" and we will see no color.  This might prompt us to run "git
> config --global color.ui always".  Now we get all the color with no
> issues.
>
>
> The steps:
>
> Go to a git repo with a stash
>
> Run "git config --global color.ui always"
>
> Run M-x vc-dir (C-x C-v d)
>
> Goto stash and hit RET to open the stash
>
> There are now spurious color control characters everywhere

I still can't reproduce this, using this recipe:

0. emacs -Q
1. C-x d ~/wip/emacs RET
2. C-x C-v d

   I'm now looking at a buffer *vc-dir* that looks like this:

       VC backend : Git
       Working dir: ~/wip/emacs/
       Branch     : master
       Remote     : skangas@git.sv.gnu.org:/srv/git/emacs.git
       Stash      : Hide all stashes (11)
                    {0}: On master: assign
                    {1}: On feature/tree-sitter: wip
                    {2}: On master: tree-sitter
       [...snip...]

3. With point on "{0}:", type RET

Now I get a buffer *vc-git-stash* which is using diff-mode, but I don't
see any spurious control characters.

[time passes]

Okay, I see it now, but I had to do this too:

    git config --global color.diff always

But I cannot reproduce it if I use this instead:

    git config --global color.diff auto

However, as Emacs can't process those control characters in any case, I
think we might as well apply the patch you suggest.  This will save
users from shooting themselves in the foot.

Does anyone see any reason not to?

> Thanks,
>
> Morgan





reply via email to

[Prev in Thread] Current Thread [Next in Thread]