Thursday, 16 March 2006 23:04, Hendrik Sattler wrote: > That's exactly the reason why so many bugs rot in the BTS forever. On > big sources, it is almost impossible for the Debian-Maintainer on all > changes to close rather upstream related bugs. > What does he do then: he asks the only person that eventually > reported it. That person says that is uses Debian stable only and > does not want to test Debian stable. The problem at hand is present in unstable and is easy to confirm. The package maintainer does not need the original reporter to confirm that a bug has been fixed. The main problem is with bugs that the package maintainer could never confirm in the first place. > So you want upstream bugs in the BTS closed after 2-3 years? Get > real! Those requesting such non-sense are the same persons that > complain about long-standing bugs in the BTS and that many bugs are > left unanswered. I want bugs in the BTS closed when, on balance, one can assume that they have been fixed. If that takes two or three years, so be it. The value of fixing bugs is not to have a clean BTS page. > Upstreams does not look at the Debian BTS, so where's the gain? Let me count the ways: - It is non-trivial for users to determine whether a bug is an upstream bug or one caused by distribution-specific modifications or packaging - It is easier for the package maintainer to determine whether the bug was introduced by the distribution-specific activities than it is for the upstream maintainer to do so (particularly if users report bugs directly to upstream and fail to mention that they are using a downstream binary) - Users can't always judge whether or not a bug is release-critical, and if it is, Debian should know about it, which they might not if BDO is bypassed - It is unreasonable to expect users to find, familiarise themselves with, and in many cases register for the upstream BTS of every piece of software that they use (in other words, the centralised BTS is an extension of centralised package management) - Many upstream maintainers won't even touch bugs from reporters who don't use upstream software - Package maintainers generally have more experience in creating effective reports containing the information their associated upstream maintainer will be looking for - Users and backport maintainers rely on BDO being a reasonably accurate reflection of the current state of packages I can accept that, pragmatically, reporting to bugs.debian.org is not always the most effective way to get things done, but it should always be a perfectly acceptable course of action, and going directly to upstream should never mean bypassing BDO altogether. -- Alex Nordstrom http://lx.n3.net/ Please do not CC me in followups; I am subscribed to debian-kde.
Attachment:
pgpwvyZUg0jzP.pgp
Description: PGP signature