On Wed, Mar 9, 2011 at 2:09 PM, Dagobert Michelsen <[email protected]> wrote: > Hi Phil, > > > From the task to clean package names in release "dublin" we are targetting > I estimate about 500 renames. Handling this on an individual basis is IMHO > prone to errors while always using intermediate packages is a simple, clean > and easy way to achieve consistent updating.
It may be "simple, clean and easy" from the perspective of someone generating a bunch of packages to dump out. But from the perspective of integrating into our package database, its very messy. Even setting aside the issue of potentially having 500 "junk" entries in our databases... which isnt so nice...., there's the issue of file collisions. if you have CSWoldpkg, and CSWnewpkg, and they both have the same files... you cant register "newpkg" until "oldpkg" has been cleaned out of the collision database! > Following the "I"-approach consistently would probably require rebuilding > the whole set and pushing 500 packages at once which would be much harder > to do right and to verify. I would say just the opposite. Using the "i" approach, would lead to doing smaller sets of packages at one time, instead of dumping 100 into the update stream all at once. Seems like the reason we have so many different perl packages, is that only a handful are actually *shared*. Most of the time, perl CPAN modules are used by only 1 or two packages, I think. (please note I said "most", but not all :-) _______________________________________________ maintainers mailing list [email protected] https://lists.opencsw.org/mailman/listinfo/maintainers .:: This mailing list's archive is public. ::.
