On Fri, Aug 02, 2002 at 04:20:52PM +0200, Santiago Vila wrote: > On Fri, 2 Aug 2002, Raphael Hertzog wrote: > > Le Thu, Aug 01, 2002 at 10:48:36PM +0200, Santiago Vila ?crivait: > > > Yes, packages should be compiled using libraries in testing. I fully > > > agree, since it would help to solve the following > > That's a solution, I won't comment. :-) > > [ But if you read the previous thread, you'll see that aj explained that > > it is not possible otherwise no new libraries would get in testing. ] > It is not only possible, but necessary, if we want to help everybody > to respect the GPL closely when it talks about distribution of source code. Sorry, this is FUD, nothing more. If you want to respect the GPL closely enough to make everything exactly rebuildable you need to rebuild the entire archive everytime there's a new version of libc or gcc uploaded. That's never going to happen unless you take someone to court and get an order in your favour, and that'll probably only result in the death of Debian, so *shrug*. Arguing that we should place license terms above what benefits our users is a pretty daft idea, too, especially where the GPL's concerned. > Every distribution that we release should have build-depends within > the same distribution. Yes, it should, and we're going to aim to do that for sarge, as Bdale'll probably reveal any time now. This doesn't require any changes to anything except the policy for the testing scripts though. For reference, what you're missing is any real conception of what the real problems are that keep packages out of testing for extended periods are. And no, I'm not going to play the game by substituting my own guesses for yours. Optimisation based on guesswork is stupid and counterproductive. Measure, then decide. Cheers, aj -- Anthony Towns <email@example.com> <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