Re: [PATCH v4] reflog-walk: don't segfault on non-commit sha1's in the reflog

[Eric Sunshine <sunshine@sunshineco.com>]

On Tue, Jan 5, 2016 at 8:28 PM, Eric Sunshine <sunshine@sunshineco.com> wrote:
> On Tue, Jan 5, 2016 at 8:20 PM, Dennis Kaarsemaker
> <dennis@kaarsemaker.net> wrote:
>> On di, 2016-01-05 at 20:05 -0500, Eric Sunshine wrote:
>>> Hmm, this test is successful for me on OS X even without the
>>> reflog-walk.c changes applied.
>>>
>>> And this test actually fails (inversely) because it's expecting a
>>> failure, but doesn't get one since the command produces the expected
>>> output.
>>
>> That's... surprising to say the least. What's the content of 'actual',
>> and which git.git commit are you on?
>
> % cat t/trash\ directory.t1410-reflog/actual
> b60a214 refs/tests/tree-in-reflog@{0}: Restoring to commit
> 140c527 refs/tests/tree-in-reflog@{1}: Forcing tree
> b60a214 refs/tests/tree-in-reflog@{2}: Creating ref
> %
>
> This is with only the t/t1410-reflog.sh changes from your patch
> applied atop current 'master' (SHA1 7548842).

By the way, the segfault does occur for me on Linux and FreeBSD.

And, in all cases, on all tested platforms, with the full patch
applied, both tests behave sanely (in the expected fashion). So, even
though the crash doesn't manifest everywhere, the fact that the tests
are meaningfully testing it on the "affected" platforms may mean that
it's not worth worrying about why it doesn't segfault on OS X.

(Of course, practicality aside, one might want to satisfy one's
intellectual curiosity about why it behaves differently on OS X.)