Re: git push failing when push.recurseSubmodules on-demand and git commit --amend was used in submodule.

[Stefan Beller <sbeller@google.com>]

On Tue, Jan 31, 2017 at 5:10 PM, Carlo Wood <carlo@alinoe.com> wrote:
> On Tue, 31 Jan 2017 14:08:41 -0800
> Stefan Beller <sbeller@google.com> wrote:
>
>> On Sun, Jan 29, 2017 at 5:00 PM, Junio C Hamano <gitster@pobox.com>
>> wrote:
>> >  2. If the amend is good and ready to go, "git add" to update the
>> >     superproject to make that amended result the one that is needed
>> >     in the submodule.
>>
>> yup.
>
> But that is what I am doing. The amended commit IS already
> added to the superproject (and pushed to the remote).
>
> Please have a look at my script, this happens here:
>
> # Correct that in the parent too:
> pushd parent
> git add subm
> git commit -m 'Updated subm.'
> popd

And if you were to use ammend here, too; there would be no problem;
In the parent there are now two commits, the first one pointing at the
first (unammended) commit in the submodule, the second pointing to
the corrected commit.

>
> The commit from before the amend was added to the super
> project (but never pushed) but has now been completely
> replaced. I still think this is a flaw in git. It shouldn't
> not complain and simply push.

The problem here is in the design.
When "on-demand" is set, the parent repo determines which
sumbodules needs pushing, then runs a plain "git push" in them
and then checks again.

The push operation in the submodule did not push out the un-amended
commit (which is correct IMO).

The parent in the second check determines that there is a commit, that
is not pushed, though. Maybe we need an option there "on-demand-but-no-recheck"
as a weaker promise what Git can deliver there.

>
> --
> Carlo Wood <carlo@alinoe.com>