I haven't received an email (mpi-all-commits) confirming my commit (re-commit) - r4970 (~ 1/2 hr back) .

Regards,
Jayesh

-----Original Message-----
From: mpich2-core-bounces@mcs.anl.gov [mailto:mpich2-core-bounces@mcs.anl.gov] On Behalf Of Dave Goodell
Sent: Wednesday, July 15, 2009 1:50 PM
To: mpich2-core@mcs.anl.gov
Subject: Re: [mpich2-core] spotty SVN history

Ken thinks he has fixed this and my spot checks all look good so far, 
but keep your eyes open for a little while when dealing with SVN.  
Also, we had to blow away Jayesh's commit r4970 in the process, so he's going to re-commit it.  Don't be surprised by the "double" commit.

-Dave

On Jul 15, 2009, at 1:03 PM, Dave Goodell wrote:

> Watch your SVN interactions today very carefully.  Systems has
> upgraded the SVN server (and trac) to new versions, but they seem to
> have lost and/or mangled some history in the process.
>
> For example, I committed r4954 against trunk at 2009-07-13 15:26:57
> and it doesn't appear correctly in the svn history anymore.  Instead
> it ends up showing me change a commit by Kumud (r4907) on the "kumudb"
> dev branch.  At least, it shows me the log message for that commit. 
> It doesn't have any diff for the whole repository for that change.  If
> you try to view the same changeset from the trac website [1] it shows
> you the right message, but it again thinks that no files or properties
> changed.  I know this change was originally recorded correctly because
> I have a commit email that shows the property change occurring.
>
> Ken is still out at lunch, but I'm planning to follow up with him
> about this once he returns.  In the mean time, I should be able to
> answer revision history questions accurately from my local git-svn
> repository as long as you don't need to know about any property
> changes (including svn:externals).  Feel free to ask if you want to
> double-check an svn query.
>
> -Dave
>
> [1] https://trac.mcs.anl.gov/projects/mpich2/changeset/4954