RE: Why does send-pack call pack-objects for all remote refs?

[Daniel Koverman <dkoverman@predictiveTechnologies.com>]

Your interpretation of my email was correct. As you picked up on, I
had a fundamental misunderstanding of what pack-objects was doing.
Thanks for the explanation, I have a much better idea of what is
going on now.

Given my use pattern, it may be reasonable for me to patch in an
option to compute

    git rev-list --objects $my_topic --not $subset_of_remote_refs

capitalizing on my knowledge of this particular repository to come
up with heuristics for picking a reasonable subset. This will
come with the risk of sometimes producing an unnecessarily large
(potentially an obscenely large) packfile. You have thoroughly
convinced me that an option like that will not generalize and would
be unsuitable for main line git.

It is also good to know that 2000 remote refs is insane. The lower
hanging fruit here sounds like trimming that to a reasonable
number, so I'll try that approach first.

Thanks again, Junio and Peff.

Daniel

-----Original Message-----
From: Jeff King [mailto:peff@peff.net] 
Sent: Monday, December 07, 2015 5:57 PM
To: Daniel Koverman
Cc: Junio C Hamano; git@vger.kernel.org
Subject: Re: Why does send-pack call pack-objects for all remote refs?

On Mon, Dec 07, 2015 at 02:41:00PM -0800, Junio C Hamano wrote:

> Also it was unclear if you are working with a shallow repository.
> The performance trade-off made between the packsize and the cycles
> is somewhat different between a normal and a shallow repository,
> e.g. 2dacf26d (pack-objects: use --objects-edge-aggressive for
> shallow repos, 2014-12-24) might be a good starting point to think
> about this issue.

Also note that for a while the "aggressive" form was used everywhere. I
think that started in fbd4a70 (list-objects: mark more commits as edges
in mark_edges_uninteresting - 2013-08-16), and was fixed in 1684c1b
(rev-list: add an option to mark fewer edges as uninteresting,
2014-12-24).

So it was present from v1.8.4.2 up to v2.3.0.

-Peff