Re: [PATCH v3 1/1] midx.c: fix an integer overflow

From: Jeff King <peff@peff.net>
Date: Mar 24, 2020, 06:01 UTC
Message-ID: <20200324060125.GA610977@coredump.intra.peff.net>
In-Reply-To: <20200323222515.779477-1-damien.olivier.robert+git@gmail.com>

On Mon, Mar 23, 2020 at 11:25:15PM +0100, Damien Robert wrote:

Show 9 quoted lines

> When verifying a midx index with 0 objects, the
>     m->num_objects - 1
> overflows to 4294967295.
> 
> Fix this both by checking that the midx contains at least one oid,
> and also that we don't write any midx when there is no packfiles.
> 
> Update the tests so that we check that `git multi-pack-index write` does
> not write an midx when there is no object.

Thanks, both sides of this make sense.

> ---
> Since I did not receive any guidelines, I did not upload an midx with no
> object to check in the tests. I just modified the current tests to check
> that we don't produce an midx if there is no objects.

I'd be OK with just this, but adding a binary t/t5319/zero-objs.midx would be fine by me, too.

One minor style nit:

Show 10 quoted lines

> @@ -1124,6 +1130,13 @@ int verify_midx_file(struct repository *r, const char *object_dir, unsigned flag
>  				    i, oid_fanout1, oid_fanout2, i + 1);
>  	}
>  
> +	if (m->num_objects == 0) {
> +		midx_report(_("the midx contains no oid"));
> +		// remaining tests assume that we have objects, so we can
> +		// return here
> +		return verify_midx_error;
> +	}

We prefer /**/ for comments, like:

  /*
   * Remaining tests assume that we have objects, so we can
   * return here.
   */

-Peff