Em Sáb, 2005-07-16 às 17:22 +0200, Joerg Jaspert escreveu: > I vote against it. > Nice example just arrived yesterday: "Just" an soname change, > maintainer didnt fix his scripts, no files installed in .debs. Simple, > nice, example against automated addition of files. This could happen to new upstream versions of non-lib packages and should better be handled by a lintian check, anyway, no? What I'm personaly interested in working on adding to katie is a check to hint a rejection is needed for a package because it depends on a package not in the archive: a problem which often happens when you maintain a library which changes soname and a non-lib package which depends on it and forgets that it will be stuck in NEW. Oops. =) The first case (missing files inside the deb) can be handled by a new Debian revision, but the second one can be a matter of downgrading an upstream version and using an epoch, which sux terribly. See ya, -- kov@debian.org: Gustavo Noronha <http://people.debian.org/~kov> Debian: <http://www.debian.org> * <http://www.debian-br.org>
Attachment:
signature.asc
Description: This is a digitally signed message part