Ok done, now your original dox history is back ;-) All files have first been restored from 6430 in a new dir, then updated, then new dir has been renamed to previous one (src).
Now I let you add new comments once you checked enough in your local svn dir, that svn add (dir) and svn move (files to new add dir at least) should correspond to unitary commits, probably we should never mix add, move, rename svn commands between commits to be (history) safe. Fabien > > Fabien Costantini wrote: > > > > > Al, I used svn move several times, before the commit, this may have > > > broken something ? > > > More precisely, I first created an images subdir, svn added it, then svn > > > moved the images, then also wanted to generalize this and then svn moved > > > the dox files, then svn moved again images to src. > > > I only used svn commands (svn add and move) but several times before the > > > commit. Maybe this helped to bug the svn... > > > > Ah, that might explain, why it happened. > > > > Now you made me curious ;-P > You made me curious too : > I googled the svn move loose history, > it seems that if you use simultaneous add (directory) and move before a > commit you loose the history. > > I also locally test svn copy between r6430 and when I blame myself (svn > blame) I still don't get the history info :-( > > So beware of copy between to revisions, it may not work as expected. > > I will add a test directory (src_doc in documentation) then will commit and > then will test locally again if svn move this time will keep history. > I currently test with r6430 just to validate the principle locally, it is > then possible to make a backup of current .dox changes Doxyfile Makefile, > revert to r6430, apply the new patches will all modifications, commit this > new version ? > > _______________________________________________ fltk-dev mailing list [email protected] http://lists.easysw.com/mailman/listinfo/fltk-dev
