On Wed, Jun 1, 2011 at 7:48 PM, Ray Rashif <[email protected]> wrote:
> On 1 June 2011 23:36, Jakob Gruber <[email protected]> wrote: > > On 06/01/2011 05:26 PM, D. Can Celasun wrote: > >> > >> Actually, it is still possible. Here's how it'd work: > >> > >> - TU changes package name from foo to bar. > >> - This automatically triggers an out-of-date notification (and an > >> explanation comment) for all packages that depend on foo. > >> - Everyone updates their packages to reflect the changes. > >> > >> Now all votes, comments and even notification lists are preserved > without > >> doing a single database query. I really don't think it gets more KISS > than > >> that. > > > > I beg to differ. The way it is right now is way more KISS. > > So far, my favorite suggestion is either 1) creating a way to transfer > votes > > and comments or 2) keeping everything as it is. > > How about marking a checkbox upon upload where it would present a text > box to rename from an older package to the current one? > > x Rename package from |________________| > > If this from field MATCHES the pkgname being submitted then as usual > nothing would be done (submission as usual as if nothing had been > marked; package would be created if it does not exist). > > If it does NOT MATCH, then a rename function (such as the one this > patch demonstrates) would be applied to that older pkg and then the > submission continues. Of course, the older package must be owned by > the currently logged in user. If the older pkg entered does not exist > in the db then the submission fails. If a package gets wrongly renamed > then the user can go through the same steps to rename that. No TU > involvement here. > > > -- > GPG/PGP ID: 8AADBB10 > I think package name modification without TU involvement is a bad idea. There is a reason only TUs can delete a package and I think this functionality should also be considered as such.
