www-infrastructure-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Dennis E. Hamilton" <orc...@apache.org>
Subject RE: mirroring git.apache.org
Date Thu, 27 Dec 2012 01:35:53 GMT
I suspect that Shawn would be happy with the list at <http://git.apache.org> if it were
in a text file instead of XHTML 1.0.  (It is very systematic and the title and URL of each
repository could be scraped without too much trouble.  A two-column text file with just those
entries would be even easier.)

That should serve well enough for the purpose of mirroring.

 - Dennis

-----Original Message-----
From: Brett Porter [mailto:brett@porterclan.net] On Behalf Of Brett Porter
Sent: Wednesday, December 26, 2012 16:04
To: infrastructure-dev@apache.org
Subject: Re: mirroring git.apache.org

Hi Shawn,

For projects that have moved to using Git, you'll find the same index here: https://git-wip-us.apache.org/repos/asf?a=project_index

The rest of the projects at git.apache.org are read-only mirrors of Subversion repositories.

In the coming months, git-wip-us will likely become the production service and be relocated.
It may be that the read-only repos would be available under the same interface at that point
and you could grab the list from there. I'll let others speak about whether that's the case,
and how you might get a better list in the mean time.


On 26/12/2012, at 6:56 AM, Shawn Pearce <sop@google.com> wrote:

> Google would like to host a public mirror of git.apache.org on the
> same speedy Git infrastructure we use to host Android[1], our public
> mirror of git.kernel.org[2], and our public mirror of
> git.eclipse.org[3]. The new mirror would be available at
> https://apache.googlesource.com/.
> I am trying to locate an HTTP URL we can periodically fetch to get the
> list of available repositories. For example kernel.org gives us [4]
> and eclipse.org gives us [5]. The only source I can locate for
> git.apache.org is an XHTML index document that is much more difficult
> to parse than the plain text formats we get from other sites. Is there
> a plain text project listing we can parse? If not can infra make one
> available?
> Once setup the mirror will be served from 6 data centers (3 in the US,
> 2 in Europe, 1 in Asia) using Google's geo-location routing. Our
> infrastructure typically does `git clone` in less than half the time
> of normal Git hosting sites.
> [1] https://android.googlesource.com
> [2] https://kernel.googlesource.com
> [3] https://eclipse.googlesource.com
> [4] http://git.kernel.org?a=project_index
> [5] http://eclipse.org/projects/git-repos.php

Brett Porter

View raw message