On Tue, Jun 03, 2008 at 08:26:32PM +0000, Josselin Mouette wrote: > Le mardi 03 juin 2008 à 21:06 +0200, Johannes Wiedersich a écrit : > > As I've understood it so far, testing is for 'people trying to help the > > developers by testing the software prior to release'. If too much of > > testing becomes _either_ unusable _or_ unavailable, it won't be possible > > to use it for testing and evaluation. > > Except during freeze phases, you won’t help developers much by reporting > issues on a testing system, unfortunately. A very large portion of these > bugs are already fixed in unstable and this is merely causing noise. To add some numbers, http://bts.turmzimmer.net/graph-large.png shows that at least 2/3 of the RC bugs in testing are fixed in unstable, and I bet it's rather 3/4 because I suspect that many bugs have wrong version informations and that in fact more affect testing than what we know. Though OTOH, unstable has many more bugs than testing that are _not_ in testing for sure :) But I know no numbers here, as the one buscan gives are for all packages, not only the ones that are also in testing. -- ·O· Pierre Habouzit ··O madcoder@debian.org OOO http://www.madism.org
Attachment:
pgphxNdY9knHk.pgp
Description: PGP signature