On Sun, May 01, 2005 at 12:38:41AM +0200, Jeroen van Wolffelaar wrote: > On Sat, Apr 30, 2005 at 06:45:26PM -0300, Otavio Salvador wrote: > > But you remove the package from testing doesn't mean we won't have > > users with it installed since it was present there so, IMHO, the > > Conflict is need. > > The bug is in the other package, packages are not required to work > around other bugs in other packages, that'd be a gigantic mess of > workarounds. If dash breaks using my package for whatever reason, I'm > not going to add a conflict: dash (with non-fixed version or whatever), > dash needs to fix it. > > Ditto here, and the fix is removing the package.
That's a pretty maintainer-centric view. For your _users_, it doesn't matter which of the packages was "guilty" - and a Conflicts is cheap. See e.g. #218717 and #220983 for examples where the other packages was technically wrong, but a Conflicts in libc6 was the only possible solution. I hope you agree that a Conflicts in a non-"guilty" package is better than email loss for your users? > --Jeroen cu Adrian -- "Is there not promise of rain?" Ling Tan asked suddenly out of the darkness. There had been need of rain for many days. "Only a promise," Lao Er said. Pearl S. Buck - Dragon Seed -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]