* Frank Lichtenheld schrieb am 04.10.05, um 16:39 Uhr: > On Tue, Oct 04, 2005 at 04:16:22PM +0200, Thomas Schmidt wrote: > > But which makes me wonder is why packages with such depends need to be > > hinted into testing manually - it used to work very well without > > manual hinting in the past - since when and why was this changed in > > the testing migration scripts? Did i miss some kind of announcement > > about this? It this supposed to be just temporally or not? > > No, I don't know of any regressions in britney in this point. > I obviously can't say anything about past transitions (because > of lack of information) but the hint was necesarry because > you can't update vdr in testing without breaking all the plugins > (because of they conflict with the vdr in unstable) but you also > can't update the plugins on their own because they depend on the > vdr in unstable. AFAIK britney was never able to handle situations > like that. I guess during previous transition there was > always only one of these things true, but not both at the same time? Well it seems that you are right - i was under the (wrong) impression that such a migration to testing was done sometime in the past, but since the VDR Team maintains the package there was no migration of a new upstream-version of vdr to testing. (Which requires all plugins to be rebuilt against this particular new upstream-version - hence this Depends: vdr >= <currentupstream>, Conflicts: vdr >> <currentupstream>-9999) It would be very nice if britney could check if the problems in testing could be prevented (package2 would be uninstallable when package1 would be updated) by checking if there is a new version of package2 in unstable and if both packages could enter testing together. Regards, Thomas P.S.: No need to CC me, i am subscribed to the list. :) -- Thomas Schmidt, Debian VDR Team http://pkg-vdr-dvb.alioth.debian.org/
Attachment:
signature.asc
Description: Digital signature