Git completion hanging on "git add -f"

Git completion hanging on "git add -f"

[Julien Nicoulaud]

[-- Attachment #1: Type: text/plain, Size: 479 bytes --]

Hi all,

I have my whole /home versionned in a git repository, with a .gitignore at
the root containing '*'. This allows me to put under version control some
selected dotfiles.
If I do a "git add -f <TAB>", the completion totally hangs, here is the
output of zprof: http://paste2.org/p/1219089 (I did a <CTRL>-<C> after a
while).
It seems that the completion recurses under the whole repository, instead of
just displaying the first level files and directories.

Regards,
Julien

Re: Git completion hanging on "git add -f"

[Frank Terbeck]

Julien Nicoulaud wrote:
[...]
> It seems that the completion recurses under the whole repository, instead of
> just displaying the first level files and directories.

Did you try if Nikolai's updated _git from
<http://www.zsh.org/mla/workers/2011/msg00088.html> improves the
situation?

Regards, Frank

-- 
In protocol design, perfection has been reached not when there is
nothing left to add, but when there is nothing left to take away.
                                                  -- RFC 1925

Re: Git completion hanging on "git add -f"

[Julien Nicoulaud]

[-- Attachment #1: Type: text/plain, Size: 959 bytes --]

Yes, I tried it but it's the same (same zprof output too).

I tried to unplug the completer provided by _git for this context with:
zstyle ':completion:*:*:git:ignored-other-files' _path_files
but there must be something I missed with zstyle syntax, because it does not
work (the part that matters is here:
https://github.com/now/zsh/blob/git/Completion/Unix/Command/_git#L59).

2011/1/30 Frank Terbeck <ft@bewatermyfriend.org>

> Julien Nicoulaud wrote:
> [...]
> > It seems that the completion recurses under the whole repository, instead
> of
> > just displaying the first level files and directories.
>
> Did you try if Nikolai's updated _git from
> <http://www.zsh.org/mla/workers/2011/msg00088.html> improves the
> situation?
>
> Regards, Frank
>
> --
> In protocol design, perfection has been reached not when there is
> nothing left to add, but when there is nothing left to take away.
>                                                  -- RFC 1925
>