Re: [PATCH] commit-reach: do not parse and iterate minima

[Jonathan Tan <jonathantanmy@google.com>]

Derrick Stolee <derrickstolee@github.com> writes:
> On 3/23/2022 5:08 PM, Jonathan Tan wrote:
> >  test_expect_success 'fetch --update-shallow' '
> > +	git init a-submodule &&
> > +	test_commit -C a-submodule foo &&
> 
> You add a submodule here so you can trigger the bug in this test, but...
> 
> >  test_expect_success 'fetch --update-shallow into a repo with submodules' '
> > -	git init a-submodule &&
> > -	test_commit -C a-submodule foo &&
> 
> This test was already named to be specific to submodules. The test names
> could probably use an update to describe what's really the difference
> between the two tests.
> 
> Is there a reason we couldn't trigger the failure only in this second
> test, allowing us to still test --update-shallow when submodules are NOT
> present?
> 
> Thanks,
> -Stolee

I tried but somehow couldn't figure out the correct server and client
state to trigger this. If I can't think of anything else, I might just
"cp -r" both repos before the "fetch --update-shallow" and that might do
the trick.