From: Shawn O. Pearce <spearce@spearce.org>
Date: Thu, 05 Apr 2007 06:27:53 GMT
Subject: Re: [PATCH] clean up and optimize nth_packed_object_sha1() usage
Message-ID: <20070405062753.GC5436@spearce.org>
In-Reply-To: <alpine.LFD.0.98.0704041639070.28181@xanadu.home>

Nicolas Pitre <nico@cam.org> wrote:
> Let's avoid the open coded pack index reference in pack-object and use 
> nth_packed_object_sha1() instead.  This will help encapsulating index
> format differences in one place.
> 
> And while at it there is no reason to copy SHA1's over and over while a 
> direct pointer to it in the index will do just fine.

Ack++.  ;-)

This is a good change.  That API has been bothering me for a while.
 
> @@ -96,14 +96,14 @@ static int find_short_packed_object(int len, const unsigned char *match, unsigne
>  			last = mid;
>  		}
>  		if (first < num) {
> -			unsigned char now[20], next[20];
> -			nth_packed_object_sha1(p, first, now);
> +			const unsigned char *now, *next;
> +		       now = nth_packed_object_sha1(p, first);
>  			if (match_sha(len, match, now)) {

Uh, there is some sort of werid whitespace indent damage on that
second added line.  It doesn't line up with the first added line,
or with either of the context lines.

-- 
Shawn.