Re: [PATCH 01/17] describe: fix accidental oid/hash type-punning

From: Jeff King <peff@peff.net>
Date: Jun 20, 2019, 18:25 UTC
Message-ID: <20190620182527.GB18704@sigill.intra.peff.net>
In-Reply-To: <xmqq7e9gp68e.fsf@gitster-ct.c.googlers.com>

On Thu, Jun 20, 2019 at 09:32:49AM -0700, Junio C Hamano wrote:

Show 14 quoted lines

> Jeff King <peff@peff.net> writes:
> 
> > The find_commit_name() function passes an object_id.hash as the key of a
> > hashmap. That ends up in commit_name_neq(), which then feeds it to
> > oideq(). Which means we should actually be the whole "struct object_id".
> >
> > It works anyway because pointers to the two are interchangeable. And
> > because we're going through a layer of void pointers, the compiler
> > doesn't notice the type mismatch.
> 
> Wow.  Good eyes.  I wouldn't have noticed this (and for the reasons
> you stated, it is very tricky for any clever compiler to notice it).
> 
> Impressed.

It only looks impressive in retrospect. It became very obvious when updating the reference to sha1hash(peeled->hash) in the same line, and then scratching my head about why this one was looking at the hash member. But history rewriting lets me re-order it to make myself look good. Thanks, "rebase -i"! :)

-Peff