Re: ab/fsck-unexpected-type (and "cat-file replace handling and optimization")

On Tue, Oct 05 2021, Jeff King wrote:

On Tue, Oct 05, 2021 at 10:47:26PM +0200, Ævar Arnfjörð Bjarmason wrote:

quoted

quoted

* ab/fsck-unexpected-type (2021-10-01) 17 commits

[...]
Note that Jeff's just-submitted cat-file series[3] will conflict with
this, as they both adjust the same "garbage" object tests. The semantic
conflict is minimal/none, but the textual one is probably annoying
(e.g. his 1/5 uses a variable I split/renamed).

Jeff: Depending on what Junio thinks of queuing ab/fsck-unexpected-type
for next what do you think about rebasing your series on top, or perhaps
take a look at the v10[4] of it/ack it in case that helps with that (since
you've been looking at some related code just now...).

The conflict is pretty easy to resolve: just keep my new cleanup tests,
but swap out the variable name. The combined diff is below for
reference (this is more readable than a remerge diff, but I guess a
remerge diff could actually be applied).

I don't mind rebasing on top if that's easier for Junio, but in that
case it may make sense to float the test cleanup to the front of the
series.

I can also just change mine to do the --batch-all-objects tests in a
separate repository (which is what the existing ones do). That has the
minor advantage that we know all objects in the repository, so rather
than picking out the interesting object with perl, we could generate the
full expected output.

Thanks, your plan sounds better. I hadn't tried the merge when I sent
that E-Mail, just saw your series & that it would conflict, wanted to
give you/Junio a headsup.