From: Jeff King <peff@peff.net>
Date: Wed, 22 Nov 2017 03:17:01 GMT
Subject: Re: [PATCH 5/5] sha1_file: don't re-scan pack directory for null sha1
Message-ID: <20171122031701.GA16789@sigill>
In-Reply-To: <xmqqr2srxeu2.fsf@gitster.mtv.corp.google.com>

On Wed, Nov 22, 2017 at 10:49:25AM +0900, Junio C Hamano wrote:

> WRT existing codepaths that pass 0{40} and refuses to notice a
> potential repository corruption (from getting a NULL for a non null
> object name), I think we would need a sweep of the codebase and fix
> them in the longer term.  As long as that will happen someday, either
> approach between "we know 'no loose object? let's redo the packs' is
> the part that matters performance-wise, so let's do a short-cut only
> for that" and "we know that callers that comes with 0{40} want to get
> NULL back" should be OK, I would think.

I agree. Let's go with the "v2 5/5" I posted then.

I'll try to work up a patch for the fetch.c case I found tomorrow, but I
suspect there are many more. But that's largely orthogonal to the
series.

-Peff