On Fri, Aug 02, 2002 at 07:47:05PM +0200, Santiago Vila wrote: > Should I buy the argument that new libraries would not enter testing? > *That* would be FUD. Well, sure, if you start making people work harder because you're obsessed with this one solution, you could probably make it work. You'll need to worry about ensuring that at some point libfoo1 can be gotten rid of (since it won't disappear from testing until no packages use it, and while it's in testing newly uploaded packages will happily be rebuilt on it), you'll need to worry about packages that Build-Depend: on packages that aren't in testing at all yet, and you'll still have problems with all the packages which aren't able to change their name whenever their shlibs or equivalent changes (eg, apache-common). You also have the problem that, done this way, nothing will be built against new libraries until they've entered testing, thus ensuring that they never get tested enough to see if they *should* go in testing until they are. Additionally, you're doing two things I'd consider fundamentally wrong. One is you're changing the way unstable works for people who don't give a rat's whisker about testing, which is something I've been trying to avoid. The other is that rather than focussing on getting problems resolved by **fixing the bugs in the packages** you're focussing on random technical solutions to make it not matter if bugs don't get fixed. That seems somewhat backwards to me: the whole point of testing, from a release POV, is to make it clear to everyone which bugs are being showstoppers. Please, please, _please_ do some actual research before continuing this thread. Cheers, aj -- Anthony Towns <firstname.lastname@example.org> <http://azure.humbug.org.au/~aj/> I don't speak for anyone save myself. GPG signed mail preferred. ``If you don't do it now, you'll be one year older when you do.''
Description: PGP signature