A web interface to git

[Posted April 28, 2005 by corbet]

Further evidence that the the kernel source code management situation is slowly stabilizing: there is now a web interface to the kernel.org git repositories. Most people, perhaps, will be interested in Linus's tree, where the latest patches merged into the mainline can be viewed, but there are several developer trees available as well. (Thanks to Steven Cole).

A web interface to git

Posted Apr 28, 2005 15:23 UTC (Thu) by thompsot (guest, #12368) [Link] (2 responses)

It seems that unless the kernel developers hit a major snag in git's capabilities, we're looking at the next defacto standard in it's early stages.

Standard?

Posted Apr 29, 2005 18:39 UTC (Fri) by proski (subscriber, #104) [Link] (1 responses)

The obvious snag is the network throughput. I think we are more likely to see something with the backend of Mercurial and multiple frontends. Or we are going to see a "git server" that would eliminate the need to download the whole repository (but it would be a step back to the client-server model of CVS and Subversion). Or git will be changed to pull old revisions on demand only.

Remember that the complete history of Linux (or at least the full set of revisions that existed in BitKeeper) has not been imported into the git repository, and I doubt that git would be able to handle it in its present form.

Standard?

Posted Apr 29, 2005 19:19 UTC (Fri) by dlang (guest, #313) [Link]

there is already work being done on a git-aware server so this will be done sooner rather then later, but it will not take things back to the CVS central server model becouse you are not depending strictly on one server and only being able to pull from that server.

one of the big things about git is that each snapshot is completely independant of the prior ones, this means that you don't have to have (or pull) older versions if you don't want to.

there have been people who have pulled the entire kernel history into git (it took 3 - 3.6G of disk for the result IIRC), so git can handle the full history.

Linus has said that at this point he really doesn't care about the older stuff (as it is available through other means) and so he doesn't intend to fully populate his repository with all the older stuff. personally I expect this to change over time, but unless people really need all of that history in git and post good reasons to back the request up it's unlikly to happen soon.

A web interface to git

Posted Apr 28, 2005 15:23 UTC (Thu) by StevenCole (guest, #3068) [Link]

Just to clarify, the thanks to me were for giving Jon a heads up about the recent announcement.

The big thanks must go to the developers. Kay Sievers and Christian Gierke are listed in the code for gitweb.cgi.

This looks great!

A web interface to git

Posted Apr 28, 2005 15:43 UTC (Thu) by Seegras (guest, #20463) [Link] (5 responses)

Linus tree? Naah, that one will come out as official kernel when its ready (tm). The interesting things are David Millers Sparc fixes of course ;).

A web interface to git

Posted Apr 28, 2005 20:30 UTC (Thu) by spot (guest, #15640) [Link] (4 responses)

Yes, but only for all 7 of us Linux/SPARC users. ;)

A web interface to git

Posted May 5, 2005 9:31 UTC (Thu) by nix (subscriber, #2304) [Link] (3 responses)

8, please. :)

(Besides, nobody uses Linux on SPARC anymore, I hope. UltraSPARC, sure...)

A web interface to git

Posted May 5, 2005 16:36 UTC (Thu) by bferrell (subscriber, #624) [Link] (2 responses)

9 and I have two of 'em!

A web interface to git

Posted May 7, 2005 2:35 UTC (Sat) by clump (subscriber, #27801) [Link] (1 responses)

Ooh, let me be number 10. *And* I actually use SPARC/32. I would like some UltraSparc 64-bit but alas... I currently have none. Linux on SPARC is great, and many thanks to all that make it happen.

A web interface to git

Posted May 13, 2005 14:20 UTC (Fri) by nix (subscriber, #2304) [Link]

I'm surprised Linux on SPARC32 still works, as it hasn't had an active maintainer since the 2.2 days.

A web interface to git

Posted Apr 28, 2005 20:02 UTC (Thu) by macemoneta (guest, #2717) [Link]

There's also an alternative interface at:

http://grmso.net:8090/

Now this is _far_ better than the bkbits web site

Posted Apr 29, 2005 6:44 UTC (Fri) by ctg (guest, #3459) [Link]

So, after only a week or two since the "disaster".. I can see signs of
things being better than they were before... (which, as a believer in the
FOSS development paradigm, I fully expected).

The SCM is dead! Long live the SCM!!

Andrew Morton's git?

Posted Apr 29, 2005 9:44 UTC (Fri) by nhoxanh (guest, #17931) [Link] (3 responses)

anybody knows where Andrew's tree is?

Andrew Morton's git?

Posted Apr 29, 2005 13:12 UTC (Fri) by corbet (editor, #1) [Link] (1 responses)

He never had a BK tree, and I doubt he'll have a git tree. The way he operates requires a different sort of tool.

Andrew Morton's git?

Posted Apr 29, 2005 22:32 UTC (Fri) by iabervon (subscriber, #722) [Link]

He seems to be interested in trying out git, at least. He was asking on the git mailing list how to use git for his workflow, and Linus had some suggestions.

Andrew Morton's git?

Posted Apr 29, 2005 18:00 UTC (Fri) by bos (guest, #6154) [Link]

Andrew used to use a set of patch management scripts that he wrote himself. These turned into (or inspired, I can't remember) quilt. I don't recall whether he uses quilt at the moment.

In effect, Andrew's "tree" is what he publishes as the -mm kernel patch sets. That's really what he works in terms of - piles of patches.