You might as well kill the entire priority business from packages altogether and rely entirely on the overrides. The priority is, after all, not really a property of a package but a property of the distribution.
As long as there is no practical way for a package maintainer to verify the correctness of the choice of priority and no reaction or penalty for making an incorrect choice, many priorities will always be wrong. A first step to sort this out might be to document what technical effect the different priorities have. AFAICT, the major users would be debootstrap and the installer, but policy doesn't mention that. Perhaps the policy isn't the right place, but for example the Developers' Reference doesn't comment either. In fact, the section on priorities in the Developers' Reference shows how silly and wasteful the priorities business is from the perspective of a package maintainer. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
