Dmitry Mikhin <dmitrym@acres.com.au> writes:
> It seems that in assigning CVS revisions to SVN revisions and
> adjusting the times, the script may break the sequential order of
> SVN revisions, so that later revisions have earlier time stampt:
>
> ... subversion r 8612 trunk, built to RPM on RH9 ...
>
> ... converted a CVS repo ...
>
> [dmitrym_at_manjushri svn]$ svn co svn+ssh://manjushri/owwe/trunk owwe
> [dmitrym@manjushri svn]$ cd owwe
> [dmitrym@manjushri owwe]$ svn log > 2
>
> ... see below ...
>
>
> So, revisions 161 and 162 are later in time then revisions 163-165 (although
> the difference is negligible and unlikely to cause problems).
>
> Similar with revisions 62, 63.
Thank you. This might be a race condition in the resyncing code in
cvs2svn... Can you file an issue on this, including the CVS repository
data needed to reproduce the problem? (The more minimal the data set,
the better, but anything helps.)
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Feb 12 18:13:00 2004