Subversion 1.5.6 tomorrow - Please review :)

[Bert Huijben <rh...@sharpsvn.net>]

Hi,

(CET, yesterday)
20:14 <@hwright> Relatedly, everybody review and vote on stuff for
1.5.6, because I *will* be rolling on Wednesday, 1 fi
x or many. :)
20:14 <@gstein> i.e. one week to pull together fixes and create a
tarball. one week to get it signed.
20:14 <@gstein> hwright: 'k
20:17 <@markphip> not to beat this to death, but getting votes in STATUS
has probably been the biggest release burden
20:18 <@hwright> looks like 1.5.6 will be the "Bert and Paul" release
20:18 <@gstein> mm? if somebody doesn't vote, then something doesn't go
in. seems fine?

So, full committers, please review:

Easy to review:
* r35566
This commit makes calling "svn log --limit 10 http://server" about 40%
faster when run against a Subversion < 1.4 server on the other side of
the world, because only one ra session is opened when no additional
revision properties are required.

* r35733
Without this patch all Subclipse operations that pass paths into javahl
use an expensive true path name operation on Windows.


And a bit harder to review:

* r35297, r35367: Allow unordered merge info range lists
And * r35466, r35712, r35713

These commits fix several issues in merge info parsing that can be
triggered by converting old svnmerge.py merge info into Subversion
internal merge info. 

* r35516 is a fix to svnmerge-migrate-history.py that makes sure no new
invalid merge info can enter repositories.

Thanks,
	Bert

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1219977