From: Shawn Pearce <spearce@spearce.org>
To: Jeff King <peff@peff.net>
Cc: git <git@vger.kernel.org>, Vicent Marti <vicent@github.com>
Subject: Re: [PATCH 10/19] pack-bitmap: add support for bitmap indexes
Date: Wed, 30 Oct 2013 10:27:26 +0000 [thread overview]
Message-ID: <CAJo=hJudyxLP-pSu-=Py6tSHsnbODDFCZ6JxAoRPtfe3opf=5A@mail.gmail.com> (raw)
In-Reply-To: <20131030081023.GK11317@sigill.intra.peff.net>
On Wed, Oct 30, 2013 at 8:10 AM, Jeff King <peff@peff.net> wrote:
> On Fri, Oct 25, 2013 at 01:55:13PM +0000, Shawn O. Pearce wrote:
>>
>> Yay! This is similar to the optimization we use in JGit to send the
>> entire pack, but the part about sending a leading prefix is new. Do
>> you have any data showing how well this works in practice for cases
>> where offset is before than length-20?
>
> Actually, I don't think it kicks in very much due to packfile ordering.
> You have all of the commits at the front of the pack, then all of the
> trees, then all of the blobs. So if you want the whole thing, it is easy
> to reuse a big chunk. But if you want only the most recent slice, we can
> reuse the early bit with the new commits, but we stop partway through
> the commit list. You still have to handle all of the trees and blobs
> separately.
>
> So in practice, I think this really only kicks in for clones anyway.
>
> In theory, you could find "islands" of ones in the bitmap and send whole
> slices of packfile at once. But you need to be careful not to send a
> delta without its base. Which I think means you end up having to
> generate the whole sha1 list anyway, and check that the other side has
> each base before reusing a delta (i.e., the normal code path).
>
> In fact, I'm not quite sure that even a partial reuse up to an offset is
> 100% safe. In a newly packed git repo it is, because we always put bases
> before deltas (and OFS_DELTA objects need this). But if you had a bitmap
> generated from a fixed thin pack, we would have REF_DELTA objects early
> on that depend on bases appended to the end of the pack. So I really
> wonder if we should scrap this partial reuse and either just have full
> reuse, or go through the regular object_entry construction.
Yes. This is why JGit only does whole pack file reuse (minus the 12
byte header and 20 byte SHA-1 trailer). Any other case has so many
corner cases that we just punt into the object entry construction path
and rely on individual object reuse.
next prev parent reply other threads:[~2013-10-30 10:28 UTC|newest]
Thread overview: 87+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-10-24 17:59 [PATCH 0/19] pack bitmaps Jeff King
2013-10-24 17:59 ` [PATCH 01/19] sha1write: make buffer const-correct Jeff King
2013-10-24 18:00 ` [PATCH 02/19] revindex: Export new APIs Jeff King
2013-10-24 18:01 ` [PATCH 03/19] pack-objects: Refactor the packing list Jeff King
2013-10-24 18:01 ` [PATCH 04/19] pack-objects: factor out name_hash Jeff King
2013-10-24 18:01 ` [PATCH 05/19] revision: allow setting custom limiter function Jeff King
2013-10-24 18:01 ` [PATCH 06/19] sha1_file: export `git_open_noatime` Jeff King
2013-10-24 18:01 ` [PATCH 07/19] compat: add endianness helpers Jeff King
2013-10-26 7:55 ` Thomas Rast
2013-10-30 8:25 ` Jeff King
2013-10-30 17:06 ` Vicent Martí
2013-10-24 18:02 ` [PATCH 08/19] ewah: compressed bitmap implementation Jeff King
2013-10-24 23:34 ` Junio C Hamano
2013-10-25 3:15 ` Jeff King
2013-10-26 7:55 ` Thomas Rast
2013-10-24 18:03 ` [PATCH 09/19] documentation: add documentation for the bitmap format Jeff King
2013-10-25 1:16 ` Duy Nguyen
2013-10-25 3:21 ` Jeff King
2013-10-25 3:28 ` Duy Nguyen
2013-10-25 13:47 ` Shawn Pearce
2013-10-30 7:50 ` Jeff King
2013-10-30 10:23 ` Shawn Pearce
2013-10-30 16:11 ` Vicent Marti
2013-10-30 16:14 ` Vicent Marti
2013-10-24 18:03 ` [PATCH 10/19] pack-bitmap: add support for bitmap indexes Jeff King
2013-10-25 13:55 ` Shawn Pearce
2013-10-30 8:10 ` Jeff King
2013-10-30 10:27 ` Shawn Pearce [this message]
2013-10-30 15:47 ` Vicent Marti
2013-10-30 16:04 ` Shawn Pearce
2013-10-30 20:25 ` Jeff King
2013-10-24 18:04 ` [PATCH 11/19] pack-objects: use bitmaps when packing objects Jeff King
2013-10-25 14:14 ` Shawn Pearce
2013-10-30 8:21 ` Jeff King
2013-10-30 10:38 ` Shawn Pearce
2013-10-30 16:01 ` Vicent Marti
2013-10-24 18:06 ` [PATCH 12/19] rev-list: add bitmap mode to speed up object lists Jeff King
2013-10-25 14:00 ` Shawn Pearce
2013-10-30 8:12 ` Jeff King
2013-10-24 18:06 ` [PATCH 13/19] pack-objects: implement bitmap writing Jeff King
2013-10-25 1:21 ` Duy Nguyen
2013-10-25 3:22 ` Jeff King
2013-10-24 18:06 ` [PATCH 14/19] repack: stop using magic number for ARRAY_SIZE(exts) Jeff King
2013-10-24 18:07 ` [PATCH 15/19] repack: turn exts array into array-of-struct Jeff King
2013-10-24 18:07 ` [PATCH 16/19] repack: handle optional files created by pack-objects Jeff King
2013-10-24 18:08 ` [PATCH 17/19] repack: consider bitmaps when performing repacks Jeff King
2013-10-24 18:08 ` [PATCH 18/19] t: add basic bitmap functionality tests Jeff King
2013-10-24 18:08 ` [PATCH 19/19] pack-bitmap: implement optional name_hash cache Jeff King
2013-10-24 20:25 ` [PATCH 0/19] pack bitmaps Junio C Hamano
2013-10-25 3:07 ` Junio C Hamano
2013-10-25 5:55 ` [PATCHv2 " Jeff King
2013-10-25 5:57 ` [PATCH v2 01/19] sha1write: make buffer const-correct Jeff King
2013-10-25 6:02 ` [PATCH 02/19] revindex: Export new APIs Jeff King
2013-10-25 6:03 ` [PATCH v2 03/19] pack-objects: Refactor the packing list Jeff King
2013-10-25 6:03 ` [PATCH v2 04/19] pack-objects: factor out name_hash Jeff King
2013-10-25 6:03 ` [PATCH v2 05/19] revision: allow setting custom limiter function Jeff King
2013-10-25 6:03 ` [PATCH v2 06/19] sha1_file: export `git_open_noatime` Jeff King
2013-10-25 6:03 ` [PATCH v2 07/19] compat: add endianness helpers Jeff King
2013-10-25 6:03 ` [PATCH v2 08/19] ewah: compressed bitmap implementation Jeff King
2013-10-25 6:03 ` [PATCH v2 09/19] documentation: add documentation for the bitmap format Jeff King
2013-10-25 6:03 ` [PATCH v2 10/19] pack-bitmap: add support for bitmap indexes Jeff King
2013-10-25 23:06 ` Junio C Hamano
2013-10-26 0:26 ` Jeff King
2013-10-26 6:25 ` Jeff King
2013-10-28 15:22 ` Junio C Hamano
2013-10-30 7:00 ` Jeff King
2013-10-26 10:14 ` Duy Nguyen
2013-10-30 7:27 ` Jeff King
2013-10-25 6:03 ` [PATCH v2 11/19] pack-objects: use bitmaps when packing objects Jeff King
2013-10-26 10:25 ` Duy Nguyen
2013-10-30 7:36 ` Jeff King
2013-10-30 10:28 ` Duy Nguyen
2013-10-30 20:07 ` Jeff King
2013-10-31 12:03 ` Duy Nguyen
2013-10-25 6:04 ` [PATCH v2 12/19] rev-list: add bitmap mode to speed up object lists Jeff King
2013-10-25 6:04 ` [PATCH v2 13/19] pack-objects: implement bitmap writing Jeff King
2013-10-25 6:04 ` [PATCH v2 14/19] repack: stop using magic number for ARRAY_SIZE(exts) Jeff King
2013-10-25 6:04 ` [PATCH v2 15/19] repack: turn exts array into array-of-struct Jeff King
2013-10-25 6:04 ` [PATCH v2 16/19] repack: handle optional files created by pack-objects Jeff King
2013-10-25 6:04 ` [PATCH v2 17/19] repack: consider bitmaps when performing repacks Jeff King
2013-10-25 6:04 ` [PATCH v2 18/19] t: add basic bitmap functionality tests Jeff King
2013-10-28 22:13 ` SZEDER Gábor
2013-10-30 7:39 ` Jeff King
2013-10-25 6:04 ` [PATCH v2 19/19] pack-bitmap: implement optional name_hash cache Jeff King
2013-10-26 10:19 ` [PATCH 20/19] count-objects: consider .bitmap without .pack/.idx pair garbage Nguyễn Thái Ngọc Duy
2013-10-30 6:59 ` Jeff King
2013-10-30 17:36 ` Junio C Hamano
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='CAJo=hJudyxLP-pSu-=Py6tSHsnbODDFCZ6JxAoRPtfe3opf=5A@mail.gmail.com' \
--to=spearce@spearce.org \
--cc=git@vger.kernel.org \
--cc=peff@peff.net \
--cc=vicent@github.com \
/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).