diff mbox series

rebase: save autostash entry into stash reflog on --quit

Message ID 353a67567a90aea8a90bce1de05d005c61b3b670.1588066252.git.liu.denton@gmail.com (mailing list archive)
State New, archived
Headers show
Series rebase: save autostash entry into stash reflog on --quit | expand

Commit Message

Denton Liu April 28, 2020, 9:31 a.m. UTC
In a03b55530a (merge: teach --autostash option, 2020-04-07), the
--autostash option was introduced for `git merge`. Notably, when
`git merge --quit` is run with an autostash entry present, it is saved
into the stash reflog. This is contrasted with the current behaviour of
`git rebase --quit` where the autostash entry is simply just dropped out
of existence.

Adopt the behaviour of `git merge --quit` in `git rebase --quit` and
save the autostash entry into the stash reflog instead of just deleting
it.

Signed-off-by: Denton Liu <liu.denton@gmail.com>
---

Notes:
    This patch is based on 'dl/merge-autostash'.

 Documentation/git-rebase.txt |  3 ++-
 builtin/rebase.c             |  1 +
 t/t3420-rebase-autostash.sh  | 20 ++++++++++++++++++++
 3 files changed, 23 insertions(+), 1 deletion(-)

Comments

Junio C Hamano April 28, 2020, 7:35 p.m. UTC | #1
Denton Liu <liu.denton@gmail.com> writes:

> In a03b55530a (merge: teach --autostash option, 2020-04-07), the
> --autostash option was introduced for `git merge`. Notably, when
> `git merge --quit` is run with an autostash entry present, it is saved
> into the stash reflog. This is contrasted with the current behaviour of
> `git rebase --quit` where the autostash entry is simply just dropped out
> of existence.
>
> Adopt the behaviour of `git merge --quit` in `git rebase --quit` and
> save the autostash entry into the stash reflog instead of just deleting
> it.

The goal is wrothy, I would think, but I do not think we would
explain it in terms of "stash reflog".  It is true that what "stash
list" shows, the list of stasn entries, happens to be implemented as
reflog entries of a single ref, but the end users would not view
them as reflog entries, I suspect.  Do you know anybody who would do
"git reflog stash@{now}"?

It is less bad to explain with "reflog of stash ref" in the log
message, meant to be read by our future developers, but ...


> diff --git a/Documentation/git-rebase.txt b/Documentation/git-rebase.txt
> index f7a6033607..7d0c89a184 100644
> --- a/Documentation/git-rebase.txt
> +++ b/Documentation/git-rebase.txt
> @@ -256,7 +256,8 @@ See also INCOMPATIBLE OPTIONS below.
>  --quit::
>  	Abort the rebase operation but HEAD is not reset back to the
>  	original branch. The index and working tree are also left
> -	unchanged as a result.
> +	unchanged as a result. If a temporary stash entry was created
> +	using --autostash, it will be saved to the stash reflog.

... let's not do so for end-user facing documentation.  "..., it
will be stashed away".  Or we may not even want to say anything; any
"--autostash" user would expect that the changes that were stashed
before "rebase" started would not be discarded, and this change may
just be a bugfix.

> diff --git a/builtin/rebase.c b/builtin/rebase.c
> index bc4fc69906..71aec532b1 100644
> --- a/builtin/rebase.c
> +++ b/builtin/rebase.c
> @@ -1556,6 +1556,7 @@ int cmd_rebase(int argc, const char **argv, const char *prefix)
>  		goto cleanup;
>  	}
>  	case ACTION_QUIT: {
> +		save_autostash(state_dir_path("autostash", &options));
>  		if (options.type == REBASE_MERGE) {
>  			struct replay_opts replay = REPLAY_OPTS_INIT;

Nice to see a fix as simple a this one ;-)
Denton Liu April 29, 2020, 12:23 a.m. UTC | #2
Hi Junio,

On Tue, Apr 28, 2020 at 12:35:15PM -0700, Junio C Hamano wrote:
> > diff --git a/Documentation/git-rebase.txt b/Documentation/git-rebase.txt
> > index f7a6033607..7d0c89a184 100644
> > --- a/Documentation/git-rebase.txt
> > +++ b/Documentation/git-rebase.txt
> > @@ -256,7 +256,8 @@ See also INCOMPATIBLE OPTIONS below.
> >  --quit::
> >  	Abort the rebase operation but HEAD is not reset back to the
> >  	original branch. The index and working tree are also left
> > -	unchanged as a result.
> > +	unchanged as a result. If a temporary stash entry was created
> > +	using --autostash, it will be saved to the stash reflog.
> 
> ... let's not do so for end-user facing documentation.  "..., it
> will be stashed away".  Or we may not even want to say anything; any
> "--autostash" user would expect that the changes that were stashed
> before "rebase" started would not be discarded, and this change may
> just be a bugfix.

Hmm, in this case, git-merge.txt may need an update as well. From
'dl/merge-autostash', 

	'git merge --abort' is equivalent to 'git reset --merge' when
	`MERGE_HEAD` is present unless `MERGE_AUTOSTASH` is also present in
	which case 'git merge --abort' applies the stash entry to the worktree
	whereas 'git reset --merge' will save the stashed changes in the stash
	reflog.

and

	--quit::
		Forget about the current merge in progress. Leave the index
		and the working tree as-is. If `MERGE_AUTOSTASH` is present, the
		stash entry will be saved to the stash reflog.

both need to be amended to remove the reference to the "stash reflog".

When I was writing this documentation, I wanted to distinguish between
the temporary autostash entry and the actual stash since the autostash
entry isn't pushed to the stash unless there are conflicts or it's
explicitly saved. I'm not sure that something like "If a temporary stash
entry was created using --autostash, it will be stashed away" works very
well since the word "stash" is overloaded here to mean "a random stash
commit" and "stashed away in _the_ stash". Unfortunately, I'm also
having trouble coming up with a suitable phrasing of my own.

I dunno, perhaps I'm overthinking this too and your suggested rewording
sounds good and I'm just being too picky.

Thanks,

Denton
Phillip Wood April 30, 2020, 9:41 a.m. UTC | #3
Hi Denton

Thanks for working on this, the implementation and test look fine

Best Wishes

Phillip

On 28/04/2020 10:31, Denton Liu wrote:
> In a03b55530a (merge: teach --autostash option, 2020-04-07), the
> --autostash option was introduced for `git merge`. Notably, when
> `git merge --quit` is run with an autostash entry present, it is saved
> into the stash reflog. This is contrasted with the current behaviour of
> `git rebase --quit` where the autostash entry is simply just dropped out
> of existence.
> 
> Adopt the behaviour of `git merge --quit` in `git rebase --quit` and
> save the autostash entry into the stash reflog instead of just deleting
> it.
> 
> Signed-off-by: Denton Liu <liu.denton@gmail.com>
> ---
> 
> Notes:
>     This patch is based on 'dl/merge-autostash'.
> 
>  Documentation/git-rebase.txt |  3 ++-
>  builtin/rebase.c             |  1 +
>  t/t3420-rebase-autostash.sh  | 20 ++++++++++++++++++++
>  3 files changed, 23 insertions(+), 1 deletion(-)
> 
> diff --git a/Documentation/git-rebase.txt b/Documentation/git-rebase.txt
> index f7a6033607..7d0c89a184 100644
> --- a/Documentation/git-rebase.txt
> +++ b/Documentation/git-rebase.txt
> @@ -256,7 +256,8 @@ See also INCOMPATIBLE OPTIONS below.
>  --quit::
>  	Abort the rebase operation but HEAD is not reset back to the
>  	original branch. The index and working tree are also left
> -	unchanged as a result.
> +	unchanged as a result. If a temporary stash entry was created
> +	using --autostash, it will be saved to the stash reflog.>  --apply:
>  	Use applying strategies to rebase (calling `git-am`
> diff --git a/builtin/rebase.c b/builtin/rebase.c
> index bc4fc69906..71aec532b1 100644
> --- a/builtin/rebase.c
> +++ b/builtin/rebase.c
> @@ -1556,6 +1556,7 @@ int cmd_rebase(int argc, const char **argv, const char *prefix)
>  		goto cleanup;
>  	}
>  	case ACTION_QUIT: {
> +		save_autostash(state_dir_path("autostash", &options));
>  		if (options.type == REBASE_MERGE) {
>  			struct replay_opts replay = REPLAY_OPTS_INIT;
>  
> diff --git a/t/t3420-rebase-autostash.sh b/t/t3420-rebase-autostash.sh
> index b97ea62363..ca331733fb 100755
> --- a/t/t3420-rebase-autostash.sh
> +++ b/t/t3420-rebase-autostash.sh
> @@ -184,6 +184,26 @@ testrebase () {
>  		git checkout feature-branch
>  	'
>  
> +	test_expect_success "rebase$type: --quit" '
> +		test_config rebase.autostash true &&
> +		git reset --hard &&
> +		git checkout -b rebased-feature-branch feature-branch &&
> +		test_when_finished git branch -D rebased-feature-branch &&
> +		echo dirty >>file3 &&
> +		git diff >expect &&
> +		test_must_fail git rebase$type related-onto-branch &&
> +		test_path_is_file $dotest/autostash &&
> +		test_path_is_missing file3 &&
> +		git rebase --quit &&
> +		test_when_finished git stash drop &&
> +		test_path_is_missing $dotest/autostash &&
> +		! grep dirty file3 &&
> +		git stash show -p >actual &&
> +		test_cmp expect actual &&
> +		git reset --hard &&
> +		git checkout feature-branch
> +	'
> +
>  	test_expect_success "rebase$type: non-conflicting rebase, conflicting stash" '
>  		test_config rebase.autostash true &&
>  		git reset --hard &&
>
Phillip Wood April 30, 2020, 9:45 a.m. UTC | #4
Hi Denton

On 29/04/2020 01:23, Denton Liu wrote:
> Hi Junio,
> 
> On Tue, Apr 28, 2020 at 12:35:15PM -0700, Junio C Hamano wrote:
>>> diff --git a/Documentation/git-rebase.txt b/Documentation/git-rebase.txt
>>> index f7a6033607..7d0c89a184 100644
>>> --- a/Documentation/git-rebase.txt
>>> +++ b/Documentation/git-rebase.txt
>>> @@ -256,7 +256,8 @@ See also INCOMPATIBLE OPTIONS below.
>>>  --quit::
>>>  	Abort the rebase operation but HEAD is not reset back to the
>>>  	original branch. The index and working tree are also left
>>> -	unchanged as a result.
>>> +	unchanged as a result. If a temporary stash entry was created
>>> +	using --autostash, it will be saved to the stash reflog.
>>
>> ... let's not do so for end-user facing documentation.  "..., it
>> will be stashed away".  Or we may not even want to say anything; any
>> "--autostash" user would expect that the changes that were stashed
>> before "rebase" started would not be discarded, and this change may
>> just be a bugfix.
> 
> Hmm, in this case, git-merge.txt may need an update as well. From
> 'dl/merge-autostash', 
> 
> 	'git merge --abort' is equivalent to 'git reset --merge' when
> 	`MERGE_HEAD` is present unless `MERGE_AUTOSTASH` is also present in
> 	which case 'git merge --abort' applies the stash entry to the worktree
> 	whereas 'git reset --merge' will save the stashed changes in the stash
> 	reflog.
> 
> and
> 
> 	--quit::
> 		Forget about the current merge in progress. Leave the index
> 		and the working tree as-is. If `MERGE_AUTOSTASH` is present, the
> 		stash entry will be saved to the stash reflog.
> 
> both need to be amended to remove the reference to the "stash reflog".

I agree that it would be good to have the documentation for rebase and
merge match in this case

> When I was writing this documentation, I wanted to distinguish between
> the temporary autostash entry and the actual stash since the autostash
> entry isn't pushed to the stash unless there are conflicts or it's
> explicitly saved. I'm not sure that something like "If a temporary stash
> entry was created using --autostash, it will be stashed away" works very
> well since the word "stash" is overloaded here to mean "a random stash
> commit" and "stashed away in _the_ stash". Unfortunately, I'm also
> having trouble coming up with a suitable phrasing of my own.

It's tricky to distinguish between the two, when I was reviewing the
merge --autostash patches I was wary of the mention of the stash reflog
but could not come up with anything better to distinguish between
remembering the stash oid and saving it to the list of stashes. I'm not
too bothered what we end up doing so long as it is consistent across the
commands.

Best Wishes

Phillip

> I dunno, perhaps I'm overthinking this too and your suggested rewording
> sounds good and I'm just being too picky.
> 
> Thanks,
> 
> Denton
>
diff mbox series

Patch

diff --git a/Documentation/git-rebase.txt b/Documentation/git-rebase.txt
index f7a6033607..7d0c89a184 100644
--- a/Documentation/git-rebase.txt
+++ b/Documentation/git-rebase.txt
@@ -256,7 +256,8 @@  See also INCOMPATIBLE OPTIONS below.
 --quit::
 	Abort the rebase operation but HEAD is not reset back to the
 	original branch. The index and working tree are also left
-	unchanged as a result.
+	unchanged as a result. If a temporary stash entry was created
+	using --autostash, it will be saved to the stash reflog.
 
 --apply:
 	Use applying strategies to rebase (calling `git-am`
diff --git a/builtin/rebase.c b/builtin/rebase.c
index bc4fc69906..71aec532b1 100644
--- a/builtin/rebase.c
+++ b/builtin/rebase.c
@@ -1556,6 +1556,7 @@  int cmd_rebase(int argc, const char **argv, const char *prefix)
 		goto cleanup;
 	}
 	case ACTION_QUIT: {
+		save_autostash(state_dir_path("autostash", &options));
 		if (options.type == REBASE_MERGE) {
 			struct replay_opts replay = REPLAY_OPTS_INIT;
 
diff --git a/t/t3420-rebase-autostash.sh b/t/t3420-rebase-autostash.sh
index b97ea62363..ca331733fb 100755
--- a/t/t3420-rebase-autostash.sh
+++ b/t/t3420-rebase-autostash.sh
@@ -184,6 +184,26 @@  testrebase () {
 		git checkout feature-branch
 	'
 
+	test_expect_success "rebase$type: --quit" '
+		test_config rebase.autostash true &&
+		git reset --hard &&
+		git checkout -b rebased-feature-branch feature-branch &&
+		test_when_finished git branch -D rebased-feature-branch &&
+		echo dirty >>file3 &&
+		git diff >expect &&
+		test_must_fail git rebase$type related-onto-branch &&
+		test_path_is_file $dotest/autostash &&
+		test_path_is_missing file3 &&
+		git rebase --quit &&
+		test_when_finished git stash drop &&
+		test_path_is_missing $dotest/autostash &&
+		! grep dirty file3 &&
+		git stash show -p >actual &&
+		test_cmp expect actual &&
+		git reset --hard &&
+		git checkout feature-branch
+	'
+
 	test_expect_success "rebase$type: non-conflicting rebase, conflicting stash" '
 		test_config rebase.autostash true &&
 		git reset --hard &&