On 10/5/2014 10:58 PM, Alberto Scotto wrote:
> Hi all,
> 
> I've just found that:
> - given you have an empty staging area
> - and you have only untracked files in your working dir
> - when you do `git stash --untracked`
> - then `git stash show` gives you an empty output => stash looks empty
> 
> My first thought was "oh god, my files are lost!"
> Second thought: "Jeez I found a bug in git! cool!"
> Then I found that actually `git stash apply` restores the apparently lost
> files
> So I think it's a UX issue.
> It cost me a few lost files already, as I thought "an empty stash? uhm..
> can't remember what/when I stashed.. whatever.. let's just delete it and
> clean up a little bit this mess of stashes".
> 
> 
> Here are the reproducible steps:
> 
>    1. create new fresh git repo in $REPO_DIR
>    2. create a couple of files/dirs and commit
>    3. edit src/MyClass.java and commit
>    4. create dir src/new-dir with one file inside
>    5. edit file.txt and stage it
>    6. stash => stashes staged changes; only untracked files are left
>    7. stash -u => stashes untracked changes => working dir is clean
>    8. stash list
>    9. git stash show -p => empty output
>    10. git stash apply (restore stashed untracked files)

Hi,

I think problem lies with  show_stash() which just shows the
diff between working tree and the base tree, it ignores the
untracked files. A quick and dirty fix can be to just show
the diff between the untracked files and a NULL commit.
Here's the patch, it works all right but can be implemented
much better. I will try to find a better approach tomorrow.

diff --git a/git-stash.sh b/git-stash.sh
index d4cf818..7088584 100755
--- a/git-stash.sh
+++ b/git-stash.sh
@@ -304,6 +304,8 @@ show_stash () {
        assert_stash_like "$@"

        git diff ${FLAGS:---stat} $b_commit $w_commit
+       empty_tree=$(git hash-object -t tree /dev/null)
+       git diff ${FLAGS:---stat} ${empty_tree} $u_commit
 }

 #

Cheers,
Tanay


--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to