git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH] Ignore dirty submodule states during stash

From
Vasily Titskiy <qehgt0@gmail.com>
Date
May 16, 2016, 15:46 UTC
Message-ID
<20160516154606.GA8797@gmail.com>
In-Reply-To
<CAGZ79kaTss6ctZDCiRP2wjuxH+rJ79RKFLM79_FJN+37Bed+HQ@mail.gmail.com>
Hi Stefan,
On Sun, May 15, 2016 at 11:37:20PM -0700, Stefan Beller wrote:
Show 5 quoted lines
> On Sun, May 15, 2016 at 7:07 PM, Vasily Titskiy <qehgt0@gmail.com> wrote:
> > Do not save states of submodules as stash should ignore it.
> 
> Can you explain why this is a good idea?
> (It is not obvious to me either way.)
Actually, submodules are already ignored by stash, but not fully (it was introduced in commit 6848d58).
Current behavior is counter-intuitive, for example (if one has a project with a submodule):
 $ cd sub1
 $ edit .. commit .. edit .. commit. Alternatively, just checkout some other commit
 $ cd .. # back to main project
 $ git stash save
   No local changes to save
 $ # so, stash declares there are no changes
 $ edit main.cpp
 $ # For example, I need to update my working tree to latest master
 $ git stash save # save local changes of 'main.cpp'...
 $ git pull --recurse-submodules && git submodule update --recursive # update to latest
 $ git stash pop # I expect to get stashed changes for 'main.cpp', but...
   warning: Failed to merge submodule sub1 (commits don't follow merge-base)
   Auto-merging sub1
   CONFLICT (submodule): Merge conflict in sub1

So, this is the issue. Instead of getting my local changes, I got a conflict (and stash is not poped out). The root cause is the 'stash' command does not know how to deal with submodules, but currently it tries to save the state of submodules, and even tries to re-apply the state (and it fails of course). The proposed solution fixes this behaviour.

All internal tests work fine with the change.
> 
> Do we need a test/documentation updates for this?
I don't think so, 'stash' have never claimed submodule support.
Show 27 quoted lines
> 
> >
> > Signed-off-by: Vasily Titskiy <qehgt0@gmail.com>
> > ---
> >  git-stash.sh | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/git-stash.sh b/git-stash.sh
> > index c7c65e2..b500c44 100755
> > --- a/git-stash.sh
> > +++ b/git-stash.sh
> > @@ -116,7 +116,7 @@ create_stash () {
> >                         git read-tree --index-output="$TMPindex" -m $i_tree &&
> >                         GIT_INDEX_FILE="$TMPindex" &&
> >                         export GIT_INDEX_FILE &&
> > -                       git diff --name-only -z HEAD -- >"$TMP-stagenames" &&
> > +                       git diff --name-only --ignore-submodules -z HEAD -- >"$TMP-stagenames" &&
> >                         git update-index -z --add --remove --stdin <"$TMP-stagenames" &&
> >                         git write-tree &&
> >                         rm -f "$TMPindex"
> > --
> > 2.1.4
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe git" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
Previous: Stefan BellerNext: Stefan Beller
Message 3 of 6 in “Ignore dirty submodule states during stash”
  1. Ignore dirty submodule states during stashVasily Titskiy, May 16, 2016
  2. Stefan BellerMay 16, 2016
  3. Vasily TitskiyMay 16, 2016
  4. Stefan BellerMay 16, 2016
  5. Vasily TitskiyMay 17, 2016
  6. Johannes SchindelinMay 16, 2016

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.