From: Junio C Hamano <junkio@cox.net>
Date: Sun, 10 Jul 2005 23:10:17 GMT
Subject: Re: [PATCH 0/2] Support for packs in HTTP
Message-ID: <7vzmsus1eu.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <Pine.LNX.4.21.0507101539220.30848-100000@iabervon.org>

Daniel Barkalow <barkalow@iabervon.org> writes:

> This series has one patch which is ready to go in and one that's not
> (although it's a reasonable phony for the current state of the git world).

I like the general direction in which this patch is leading us.

But before going further, I'd like to see a consensus on the
pack naming convention.  The "sha1 of packed object names" was
originally introduced to easily avoid the pack name collisions,
but not enforced, so a user could do the following and still
expect things to work:

    $ n=`git-pack-objects pk <list-of-objects`
    $ mv pk-$n.pack .git/objects/pack/pk.pack
    $ mv pk-$n.idx .git/objects/pack/pk.idx

The first part of this patch makes things stricter, and your
packfile under .git/objects/pack _must_ be named pack-X{40}.pack
(I am not saying this is a bad thing).  So I would suggest
either:

  - droping the packname parameter from git-pack-objects.  Make
    the packs always named pack-X{40}.pack (or just X{40}.pack);
    also have verify-pack to verify the name of the packfile,
    and make sure X{40} part of the name matches what it claims
    to contain;

  - or drop sha1_pack_name() and let the user name the pack any
    way he wants.

I am moderately in favor of the former.