Re: [PATCH 09/14] hash: expose hash context functions to Rust

From: Junio C Hamano <gitster@pobox.com>
Date: Oct 30, 2025, 21:52 UTC
Message-ID: <xmqqh5vg3vm4.fsf@gitster.g>
In-Reply-To: <aQPbsX-J4eKvRTar@fruit.crustytoothpaste.net>

"brian m. carlson" <sandals@crustytoothpaste.net> writes:

Show 16 quoted lines

> On 2025-10-29 at 16:32:50, Junio C Hamano wrote:
>> "brian m. carlson" <sandals@crustytoothpaste.net> writes:
>> 
>> > +struct git_hash_ctx *git_hash_alloc(void)
>> > +{
>> > +	return malloc(sizeof(struct git_hash_ctx));
>> > +}
>> 
>> Not an objection, but this looked especially curious to me because
>> it has been customary to use xmalloc() for a thing like this.  Going
>> forward, is our intention that we'd explicitly handle OOM allocation
>> failures ourselves, at least in the Rust part of the code base?
>
> No, I'll change this to use `xmalloc`.  Rust handles allocation itself
> and just panics on OOM, so we will not want to handle allocation
> failures ourselves.

Thanks.

And re-reading what I wrote, it does not make much sense, as we would want the integration go in both direction. I should try hard to get out of this mentality of talking about C-part and Rust-part of the system. What is allocated in one side needs to be able to go to the other side and then come back seamlessly.