From: karthik nayak <karthik.188@gmail.com>
Date: Wed, 27 Nov 2024 12:23:07 GMT
Subject: Re: [PATCH v2] builtin: allow passing custom data to sub-commands
Message-ID: <CAOLa=ZRj8SiZu3AiBTGHwp+SiuBNBj3VAsVG9PppTjyKuz9oyA@mail.gmail.com>
In-Reply-To: <CAOLa=ZQgKv4SEPRqQix8pJo1nr30Fpg2oF-LdDX1ynHeSjDcGw@mail.gmail.com>

karthik nayak <karthik.188@gmail.com> writes:

> Junio C Hamano <gitster@pobox.com> writes:
>
>> karthik nayak <karthik.188@gmail.com> writes:
>>
>>> Could you elaborate on why you think this is an over-optimization? We
>>> don't loose functionality with this, nor do we have to have add
>>> additional code to handle the typecasting to `struct repository *` if
>>> needed. But would definitely like to resolve anything I missed.
>>
>> I do not know if it is an optimization, but if you know it is of a
>> specific type, casting back and force with "void *" loses type
>> safety, no?
>
> That is true of course, I just thought that is the price to pay in C for
> having to deal with generic data propagation. In the end I'm okay with
> both scenarios, I see merits both ways. I do tend to incline on the
> generic version in v2, since it is a lot more expandable.

That said, I see v1 is merged into `next` [1]. I think we should keep it
that way and like you said modify it in the future if needed to add
varargs or another `void *` argument.

Thanks

[1]: https://github.com/gitster/git/commit/668c0913f5803ebcc3b0b1f13e6cd49503194869