Re: "git stash pop" is doing an unwanted "git add" when there are conflicts.

[Alan Mackenzie <acm@muc.de>]

Hello, Jeff.

On Tue, Dec 22, 2015 at 04:30:33AM -0500, Jeff King wrote:
> On Tue, Dec 22, 2015 at 09:17:38AM +0100, Dennis Kaarsemaker wrote:

> > On ma, 2015-12-21 at 14:29 +0000, Alan Mackenzie wrote:
> > > Hello, git project.

> > > Last night, whilst clearing out a stale "stash stack", I did "git stash
> > > pop".  There were conflicts in two files.

> > > However, all the popped files became staged.  This doesn't normally happen.
> > > It was intensely irritating, and required me to do "git reset HEAD" on
> > > each of the files, none of which I wanted to commit.

> > > I searched the git-stash man page for this scenario, but found nothing
> > > about it.

> > > Surely staging all the files is a bug?

> > That depends. A stash is two commits: one for all changes that were in
> > the index when you ran 'git stash save' and one for all changes not yet
> > in the index. When you pop the stash, these then get restored as staged
> > resp. unstaged changes. So if your changes are now all staged, I'd
> > wager that they were staged when you ran git stash save.

> No, I think there's something else going on. Try this:
  
>     git init repo &&
>     cd repo &&
    
>     echo base >one &&
>     echo base >two &&
>     git add . &&
>     git commit -m base &&
    
>     echo stash >one &&
>     echo stash >two &&
>     git stash &&
    
>     echo "==> No conflicts, nothing staged"
>     git stash apply &&
    
>     git reset --hard &&
>     echo changes >two &&
>     git commit -am changes &&
    
>     echo "==> Conflict stages non-conflicting file 'one'"
>     ! git stash apply &&
>     git status

Thanks for creating a reproducible test case for me!

> It seems to be a side effect of merge-recursive to stage the results,
> and in the no-conflict path we explicitly reset the index. For the
> conflicting case, it's trickier, because we would want to retain the
> unmerged entries.

> So I agree it's kind of weird, but the conflicting case is inherently
> going to touch the index, and you'd generally have to `git add` to mark
> the resolutions (but if you really want to just touch the working tree,
> you'd need to `git reset`).

>From the point of view of a user, this is suboptimal.  git stash is an
abstraction: the preservation of uncomitted changes for later.  Staging
previously unstaged changes with git stash pop severely damages this
abstraction.

Are there any prospects of this getting fixed?

> -Peff

-- 
Alan Mackenzie (Nuremberg, Germany).