On 26 Aug., 04:35, John H Palmieri <jhpalmier...@gmail.com> wrote: > On Thursday, August 25, 2011 6:08:54 PM UTC-7, leif wrote: > > John, is this still current, i.e., do we still need this (especially > > aborting in case there are uncommitted changes)? > > I don't know about "need", but for example uncommitted changes in at least > one of the scripts repo or the main Sage repo (or maybe both) causes > upgrading to fail in strange ways. So the point of the check is to fail > gracefully, not fail very badly.
Well, uncommitted changes are committed upon root repo (re)installation anyway. (Same for scripts and the library IIRC; extcode for sure.) We should just ignore lines starting with "? " there, and IMHO only issue a warning if other changes remain; or prompt the user whether to commit or continue, or exit, unless a to-be-implemented "force" option was given. > > Note that len(output)>0 doesn't necessarily mean there are changes > > that should be committed; any file unknown to Mercurial will show up > > there. > > > If only "hg status" had an exit status of zero if no changes, something > > else otherwise. I couldn't really read that... > I think that the upgrade process is fraught with peril, and could be cleaned > up significantly. Handling uncommitted changes (or committed changes which > need to be merged with the new spkg) isn't handled well at all, I think in > any repo. Hmmm. Suggestions? Should we require *all* repos to be clean, and test all of them for uncommitted changes *there*, proceeding as suggested above in case they aren't? -leif -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
