From: Nicolas Pitre <nico@fluxnic.net>
To: Duy Nguyen <pclouds@gmail.com>
Cc: Jeff King <peff@peff.net>,
Jonathan Tan <jonathantanmy@google.com>,
Git Mailing List <git@vger.kernel.org>
Subject: Re: [PATCH 2/2] index-pack: prefetch missing REF_DELTA bases
Date: Mon, 20 May 2019 19:04:14 -0400 (EDT) [thread overview]
Message-ID: <nycvar.YSQ.7.76.1905201803520.1558@knanqh.ubzr> (raw)
In-Reply-To: <CACsJy8AkhKX57RYL1Z+HZHqKbAKKOcLoRkgwg8bSnk+DW2+Nmg@mail.gmail.com>
On Sat, 18 May 2019, Duy Nguyen wrote:
> On Fri, May 17, 2019 at 3:55 PM Jeff King <peff@peff.net> wrote:
> >
> > On Fri, May 17, 2019 at 02:20:42PM +0700, Duy Nguyen wrote:
> >
> > > On Fri, May 17, 2019 at 12:35 PM Jeff King <peff@peff.net> wrote:
> > > > As it turns out, index-pack does not handle these complicated cases at
> > > > all! In the final fix_unresolved_deltas(), we are only looking for thin
> > > > deltas, and anything that was not yet resolved is assumed to be a thin
> > > > object. In many of these cases we _could_ resolve them if we tried
> > > > harder. But that is good news for us because it means that these
> > > > expectations about delta relationships are already there, and the
> > > > pre-fetch done by your patch should always be 100% correct and
> > > > efficient.
> > >
> > > Is it worth keeping some of these notes in the "third pass" comment
> > > block in index-pack.c to help future readers?
> >
> > Perhaps. I started on the patch below, but I had trouble in the commit
> > message. I couldn't find the part of the code that explains why we would
> > never produce this combination, though empirically we do not.
Good question indeed.
> That still has some value even if your commit ends up with a question
> mark. There's not much to dig out of 636171cb80 (make index-pack able
> to complete thin packs., 2006-10-25). Adding Nico, maybe he still
> remembers...
What about this comment in fix_unresolved_deltas():
/*
* Since many unresolved deltas may well be themselves base objects
* for more unresolved deltas, we really want to include the
* smallest number of base objects that would cover as much delta
* as possible by picking the
* trunc deltas first, allowing for other deltas to resolve without
* additional base objects. Since most base objects are to be found
* before deltas depending on them, a good heuristic is to start
* resolving deltas in the same order as their position in the pack.
*/
Doesn't that cover it?
In pack-objects, another comment says:
* Depth value does not matter - find_deltas() will
* never consider reused delta as the base object to
* deltify other objects against, in order to avoid
* circular deltas.
Sorry if I'm not of any help here. Although I used to have my brain
wrapped around this code pretty tightly, it's been quite a while, and
the code did change as well since then.
Nicolas
next prev parent reply other threads:[~2019-05-20 23:04 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-05-14 21:10 [PATCH 0/2] Partial clone fix: handling received REF_DELTA Jonathan Tan
2019-05-14 21:10 ` [PATCH 1/2] t5616: refactor packfile replacement Jonathan Tan
2019-05-15 8:36 ` Johannes Schindelin
2019-05-15 18:22 ` Jonathan Tan
2019-05-14 21:10 ` [PATCH 2/2] index-pack: prefetch missing REF_DELTA bases Jonathan Tan
2019-05-15 8:46 ` Johannes Schindelin
2019-05-15 18:28 ` Jonathan Tan
2019-05-17 18:33 ` Johannes Schindelin
2019-05-15 23:16 ` Jeff King
2019-05-16 1:43 ` Junio C Hamano
2019-05-16 4:04 ` Jeff King
2019-05-16 18:26 ` Jonathan Tan
2019-05-16 21:12 ` Jeff King
2019-05-16 21:30 ` Jonathan Tan
2019-05-16 21:42 ` Jeff King
2019-05-16 23:15 ` Jonathan Tan
2019-05-17 1:09 ` Jeff King
2019-05-17 1:22 ` Jeff King
2019-05-17 4:39 ` Jeff King
2019-05-17 4:42 ` Jeff King
2019-05-17 7:20 ` Duy Nguyen
2019-05-17 8:55 ` Jeff King
2019-05-18 11:39 ` Duy Nguyen
2019-05-20 23:04 ` Nicolas Pitre [this message]
2019-05-21 21:20 ` Jeff King
2019-06-03 22:23 ` Jonathan Nieder
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=nycvar.YSQ.7.76.1905201803520.1558@knanqh.ubzr \
--to=nico@fluxnic.net \
--cc=git@vger.kernel.org \
--cc=jonathantanmy@google.com \
--cc=pclouds@gmail.com \
--cc=peff@peff.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).