In <20100202133013.gi23...@prunille.vinc17.org>, Vincent Lefevre wrote: >On 2010-02-01 21:41:16 -0600, Boyd Stephen Smith Jr. wrote: >> This all means that it is possible to run into some unique breakage for >> your mix of packages. Bugs should only be filed when they can be found in >> a pure- oldstable (security bugs only), pure-stable (RC or security bugs >> only), pure- testing (only during a freeze; bugfixes should normally come >> through unstable), or pure-unstable setting. > >There's no such thing as a pure-unstable setting.
Yes, there is. >Package upgrades >can come in an unpredictable order, that depends on the architecture, >the mirror and so on. That doesn't prevent a pure-unstable setting from existing. It just guarantees that (a) it changes frequently and (b) it varies from architecture to architecture. pure-testing also changes fairly frequently and can vary from architecture to architecture. Even pure-oldstable varies from architecture to architecture. It shouldn't depend too much on the mirror. The way the mirroring process works you should only get an "older" pure-unstable, not one that never existed on the main server. >Dependencies should ensure that the system is consistent (as >long as one doesn't downgrade or install non-official packages or >packages from experimental). Otherwise this is a bug. That would be great, and my experience tell me that it is normally true. However, DDs don't have time to test (or fix) every combination of packages and versions that are available -- even from only official sources excluding experimental. If you do file a bug, a DD will probably be willing to bump the dependency to a versioned dependency whenever the next upload to unstable is done. That doesn't help your stable/testing mixed system though -- not for a minimum of ten days. -- Boyd Stephen Smith Jr. ,= ,-_-. =. b...@iguanasuicide.net ((_/)o o(\_)) ICQ: 514984 YM/AIM: DaTwinkDaddy `-'(. .)`-' http://iguanasuicide.net/ \_/
signature.asc
Description: This is a digitally signed message part.