From: linux@horizon.com
To: jonsmirl@gmail.com, linux@horizon.com
Cc: git@vger.kernel.org
Subject: Re: Change set based shallow clone
Date: 8 Sep 2006 14:42:15 -0400 [thread overview]
Message-ID: <20060908184215.31789.qmail@science.horizon.com> (raw)
In-Reply-To: <9e4733910609071923tf1c49f6o70419e961e9eb66f@mail.gmail.com>
Thanks for making suggestions. It's easier to knock straw men down
than make them. But forgive me if I have fun pointing out the holes...
>> Okay. Now, the server hasn't heard of one or more of those commit
>> objects, because they're local changes. What then?
>
> Toss them, if they don't exist on the server the server is going to be
> able to send any objects for them.
Er... that makes "git pull git://git.other-hacker.personal/dir"
impossible. There's a reason that git can handle this as it presently
exists:
a-b-c-d-e-f <-- HEAD
/
Hacker 1: o-o-o-o <-- origin
Hacker 2: o-o-o-o-o-o <-- origin
\
v-w-x-y <-- HEAD
Suppose hacker 2, who happened to sync with upstream more recently,
wants to pull from hacker 1, in the hopes of building
a-b-c-d-e-f
/ \
Hacker 2: o-o-o-o-o-o z <-- HEAD
\ /
v-w-x-y
This works now, and should continue to work.
> Client would track this incrementally and not recompute it each time.
Yes, this is probably possible. I haven't worked it out, but given a
cache of precomputed (commit,depth) numbers, you can trace back from the
new heads until you hit a cache entry.
> If you follow the links in what looks to be a dangling object sooner
> or latter you will run into the root object or a 'not present' object.
> If you hit one of those the objects are not dangling and should be
> preserved.
I don't understand. It seems like you're saying that any commit without an
ancestor, and all objects recursively pointing to it, are not garbage.
How would anything ever get declared garbage in this case?
> Here is another way to look at the shallow clone problem. The only
> public ids in a git tree are the head and tag pointers. Send these to
> the client. Now let's modify the git tools to fault the full objects
> in one by one from the server whenever a git operation needs the
> object. Dangling references would point to 'not-present' objects.
Er... that would fault in a gigabyte the first time someone ran gitk,
or several other history-browsing commands. Don't you need a way to say
"tell the user this isn't present and will take an hour to download"?
Well, thinking, there are actually two shallow clone possibilities:
1) Don't load the unwanted commit objects
2) Clone the commit objects, but not their trees.
The latter would let you browse the commit history, at least.
next prev parent reply other threads:[~2006-09-08 18:42 UTC|newest]
Thread overview: 101+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-09-07 19:52 Change set based shallow clone Jon Smirl
2006-09-07 20:21 ` Jakub Narebski
2006-09-07 20:41 ` Jon Smirl
2006-09-07 21:33 ` Jeff King
2006-09-07 21:51 ` Jakub Narebski
2006-09-07 21:37 ` Jakub Narebski
2006-09-07 22:14 ` Junio C Hamano
2006-09-07 23:09 ` Jon Smirl
2006-09-10 23:20 ` Anand Kumria
2006-09-08 8:48 ` Andreas Ericsson
2006-09-07 22:07 ` Junio C Hamano
2006-09-07 22:40 ` Jakub Narebski
2006-09-08 3:54 ` Martin Langhoff
2006-09-08 5:30 ` Junio C Hamano
2006-09-08 7:15 ` Martin Langhoff
2006-09-08 8:33 ` Junio C Hamano
2006-09-08 17:18 ` A Large Angry SCM
2006-09-08 14:20 ` Jon Smirl
2006-09-08 15:50 ` Jakub Narebski
2006-09-09 3:13 ` Petr Baudis
2006-09-09 8:39 ` Jakub Narebski
2006-09-08 5:05 ` Aneesh Kumar K.V
2006-09-08 1:01 ` linux
2006-09-08 2:23 ` Jon Smirl
2006-09-08 8:36 ` Jakub Narebski
2006-09-08 8:39 ` Junio C Hamano
2006-09-08 18:42 ` linux [this message]
2006-09-08 21:13 ` Jon Smirl
2006-09-08 22:27 ` Jakub Narebski
2006-09-08 23:09 ` Linus Torvalds
2006-09-08 23:28 ` Jon Smirl
2006-09-08 23:45 ` Paul Mackerras
2006-09-09 1:45 ` Jon Smirl
2006-09-10 12:41 ` Paul Mackerras
2006-09-10 14:56 ` Jon Smirl
2006-09-10 16:10 ` linux
2006-09-10 18:00 ` Jon Smirl
2006-09-10 19:03 ` linux
2006-09-10 20:00 ` Linus Torvalds
2006-09-10 21:00 ` Jon Smirl
2006-09-11 2:49 ` Linus Torvalds
2006-09-10 22:41 ` Paul Mackerras
2006-09-11 2:55 ` Linus Torvalds
2006-09-11 3:18 ` Linus Torvalds
2006-09-11 6:35 ` Junio C Hamano
2006-09-11 18:54 ` Junio C Hamano
2006-09-11 8:36 ` Paul Mackerras
2006-09-11 14:26 ` linux
2006-09-11 15:01 ` Jon Smirl
2006-09-11 16:47 ` Junio C Hamano
2006-09-11 21:52 ` Paul Mackerras
2006-09-11 23:47 ` Junio C Hamano
2006-09-12 0:06 ` Jakub Narebski
2006-09-12 0:18 ` Junio C Hamano
2006-09-12 0:25 ` Jakub Narebski
2006-09-11 9:04 ` Jakub Narebski
2006-09-10 18:51 ` Junio C Hamano
2006-09-11 0:04 ` Shawn Pearce
2006-09-11 0:42 ` Junio C Hamano
2006-09-11 0:03 ` Shawn Pearce
2006-09-11 0:41 ` Junio C Hamano
2006-09-11 1:04 ` Jakub Narebski
2006-09-11 2:44 ` Shawn Pearce
2006-09-11 5:27 ` Junio C Hamano
2006-09-11 6:08 ` Shawn Pearce
2006-09-11 7:11 ` Junio C Hamano
2006-09-11 17:52 ` Shawn Pearce
2006-09-11 2:11 ` Jon Smirl
2006-09-09 1:05 ` Paul Mackerras
2006-09-09 2:56 ` Linus Torvalds
2006-09-09 3:23 ` Junio C Hamano
2006-09-09 3:31 ` Paul Mackerras
2006-09-09 4:04 ` Linus Torvalds
2006-09-09 8:47 ` Marco Costalba
2006-09-09 17:33 ` Linus Torvalds
2006-09-09 18:04 ` Marco Costalba
2006-09-09 18:44 ` linux
2006-09-09 19:17 ` Marco Costalba
2006-09-09 20:05 ` Linus Torvalds
2006-09-09 20:43 ` Jeff King
2006-09-09 21:11 ` Junio C Hamano
2006-09-09 21:14 ` Jeff King
2006-09-09 21:40 ` Linus Torvalds
2006-09-09 22:54 ` Jon Smirl
2006-09-10 0:18 ` Linus Torvalds
2006-09-10 1:22 ` Junio C Hamano
2006-09-10 3:49 ` Marco Costalba
2006-09-10 4:13 ` Junio C Hamano
2006-09-10 4:23 ` Marco Costalba
2006-09-10 4:46 ` Marco Costalba
2006-09-10 4:54 ` Junio C Hamano
2006-09-10 5:14 ` Marco Costalba
2006-09-10 5:46 ` Junio C Hamano
2006-09-10 15:21 ` linux
2006-09-10 18:32 ` Marco Costalba
2006-09-11 9:56 ` Paul Mackerras
2006-09-11 12:39 ` linux
2006-09-10 9:49 ` Jakub Narebski
2006-09-10 10:28 ` Josef Weidendorfer
-- strict thread matches above, loose matches on Subject: below --
2006-09-09 10:31 linux
2006-09-09 13:00 ` Marco Costalba
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=20060908184215.31789.qmail@science.horizon.com \
--to=linux@horizon.com \
--cc=git@vger.kernel.org \
--cc=jonsmirl@gmail.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).