On 14 August 2013 10:24, Pablo Beltran <pbeltr...@gmail.com> wrote: > Hi!, > > Please, see the log below. The newer revision r333117 is older (1999) than > the its ancestor r333113 (2005). > > I guess that someone imported an older CVS repository in the 2005 year by > using the cvs2svn tool, hence the r333117 date comes from the original CVS > repo . > > Should it be considered as a bug or as an acceptable strange use case? > > I'm developing a Subversion based > product<https://marketplace.atlassian.com/1211294>and I would like to know > the official answer. > > Thanks, > Pablo. > > svn log http://svn.apache.org/repos/asf/xalan/java@333117 > ------------------------------------------------------------------------ > r333117 | (no author) | 1999-11-09 17:50:01 +0100 (Tue, 09 Nov 1999) | 1 > line > > New repository initialized by cvs2svn. > ------------------------------------------------------------------------ > r333113 | bayard | 2005-11-13 21:29:12 +0100 (Sun, 13 Nov 2005) | 1 line > > preparing migration > ------------------------------------------------------------------------ > > I don't consider this a bug because it accurately reflects the date that revision was authored. If you wanted to, you can go back and change the timestamp because it is just another revision property, but revisions that are out of chronological order seems like a natural consequence of migrating more than one project into a single SVN repository. The only problem is, as it says in the manual [1], that Subversion's ability to correctly convert revision dates into real revision numbers may be impaired by non-chronological revisions.
[1] http://svnbook.red-bean.com/en/1.7/svn-book.html#svn.tour.revs.dates -- Mat Booth Software Engineer | WANdisco http://www.wandisco.com
