Re: [GSoC][PATCH] unit-tests: add tests for oidset.h

From: Junio C Hamano <gitster@pobox.com>
Date: Sep 26, 2024, 18:28 UTC
Message-ID: <xmqqy13ei819.fsf@gitster.g>
In-Reply-To: <CAP8UFD2yTMNmx0n1jhOu7dz_4XeOyTy1iLmRWYmuf9QJf75hsQ@mail.gmail.com>

Christian Couder <christian.couder@gmail.com> writes:

Show 20 quoted lines

> On Sat, Aug 24, 2024 at 7:20 PM Ghanshyam Thakkar
> <shyamthakkar001@gmail.com> wrote:
>>
>> Add tests for oidset.h library, which were not previously present using
>> the unit testing framework.
>
> It might be interesting to also say if there are tests for oidset in
> the end-to-end tests, not just in the unit test framework. Also I
> think oidset.h is more an API than a library.
> ...
> If initializing the hash algo fails here, it is likely because it
> already failed when get_oid_arbitrary_hex() (which initializes it) was
> called in the tests before this one. So I think it might be even
> better to move the above hash algo initialization code to setup() and
> make setup() error out in case the initialization fails. Then setup()
> could pass 'hash_algo' to all the functions it calls, even if some of
> them don't use it.
> ...
>
> Thanks.

While reviewing the "What's cooking" list of topics after tagging -rc0 of this development cycle, I noticed that this topic from late August has been expecting but not yet seeing an update.

As discussed elsewhere on the "Project Tracking" thread, I am in favor of formally adopting a policy to discard a topic from 'seen' after being inactive for 3 weeks, without having seen a clear consensus that it is good enough to be moved to 'next'. Interested parties are still free to revive the topic even after such a discard event.

    Side note: The definition of being "inactive" for the purpose of
    the policy is that nobody has discussed the topic, no new
    iteration of the topic was posted, and no responses to the
    review comments were given.

I'll discard this one by the end of this week unless the topic sees any activity. It looks to me that the project decided that a longer term direction to adopt "clar" as the unit-tests framework, so this patch would need to be written even if it were perfect in the old world order anyway.

Thanks.