Excerpts from Maciej Bliziński's message of Sun Sep 04 03:40:41 -0400 2011:
> > This is a technically valid catalog as all dependencies are met > > (chkcat should be happy). I'm currently detecting this as a "Silent > > Name Change." This is an instance of 'un-stubbing' which I think > > should be considered an error. What do others think? > > In general, is catalogname change with no pkgname change an error? I think we've generally treated this as an error condition. A package name change was (and is) worse since that is what Solaris itself cares about, but we've tried to avoid catalog name changes. > > In this specific case, was the name change intentional? It looks a > > bit funny to me to start calling CSWpmdbi pm_dbi2 (version number is > > 1.616) all of a sudden after stubbing it out. Can someone provide > > insight on this? > > I vaguely remember that a more complicated rename case required you > to change the catalogname where you didn't want to, and that the > number 2 was used. I don't recall the specifics, it was something > like two renames of the same package. In cases where we do want to allow this, the tool will need an 'ignore errors on foo' option so that it can be driven manually. As most errors will halt the mirror push, this would allow us to register changes in the catalog, cache this catalog and then have the rest of the job re-run. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 _______________________________________________ maintainers mailing list [email protected] https://lists.opencsw.org/mailman/listinfo/maintainers .:: This mailing list's archive is public. ::.
