> -----Original Message-----
> From: Hyrum K Wright [mailto:hyrum_at_hyrumwright.org]
> Sent: vrijdag 3 juni 2011 21:12
> To: Subversion Development
> Subject: 1.7.0-alpha1 tarballs up for testing/signing
>
> All,
>
> At long request, we now have tarballs for the first 1.7.0 pre-release:
> 1.7.0-alpha1 ("phoenix-hatchling"). These were cut from trunk, and
> the magic rev is r1130808. The buildbots were green at that rev, and
> all tests pass for me locally. You can find them here:
> http://people.apache.org/~hwright/svn/1.7.0-alpha1/
>
> This is an alpha release, and there are obviously known issues. The
> purpose is to give distributors and downstream consumers a concrete
> milestone against which to test, as well as vet our release processes
> on ASF infrastructure. As you may expect, these prereleases are
> intended for intrepid users and testers only.
>
> As a reminder, there have been a few changes in the release artifacts
> since the 1.6.x series. We no longer distribute a deps tarball, but
> instead provide a get-deps.sh script for folks who want to fetch the
> dependencies. There is no longer a 'to-tigris' directory in the
> candidate directory. We no longer computer md5sums of the tarballs,
> but continue to provide sha1sums to verify integrity. (There's
> probably other stuff I've forgotten.)
>
> I hope to get signatures from folks by the end of Tuesday, June 7, and
> then do the public announcement. Historically, we've not required a
> full set of signatures for alphas and betas, but it would be nice to
> get at least some coverage besides just my own. As you send in your
> signature announcement, please note any issues you found with the
> alpha release (these will be included in the public announcement).
> Please send sigs here:
> http://work.hyrumwright.org/pub/svn/collect_sigs.py
By switching a development version of SharpSvn to trunk I identified the
following problems since this tag:
r1131052 - svn_repos_recover3 segfaults with a NULL callback.
r1131284 - svn_client_diff4 and svn_client_diff_peg4 pass invalid arguments
to their diff5 variants, thereby ignoring the ignore_content_type flag.
These problems don't impact our svn client, but are pretty serious for old
api users.
(I suggest not holding the release for them, but they belong in the release
notes)
I have a few more problems I'm investigating, but none of them look very
serious at this time.
r1131086 is also a direct result of this investigation: 'svn info
svn://localhost' used to provide 'Path: localhost' and now provides 'Path:
.'.
This difference is caused by switching from svn_path_basename to the more
correct svn_uri_basename.
Bert
Received on 2011-06-05 01:15:54 CEST