On Tue, Jan 20, 2015 at 12:17:33AM -0500, Samuel Bronson wrote: > As you know, bugs.debian.org doesn't eat the backport changelogs, so > it has no idea about what bugs might be present in any given BPO > version of a package. Even when those bugs are merely inherited from > the backported version of the package. > > Consequently, apt-listbugs will, at best, be able to find out about > bugs in backports if they were reported as found in that exact > backport version of the package -- which, if the reporter followed the > backports documentation, wouldn't even exist. > > It would be nice if you could process those changelogs, even if only > for apt-listbugs' benefit.
I heard a second-hand report today that this was deliberately not being done on the grounds that some package maintainers don't want bugs against backports to be filed in the BTS. I found the link to version tracking in particular to be surprising, and I want to check that there isn't a degree of "telephone game" happening here. My intent in sending this email isn't to rehash that particular argument, which I know has gone back and forward a fair bit. There's a separate bug (https://bugs.debian.org/897934) for some of the problems that would perhaps make it more palatable for backports bugs to be handled in the BTS. However, leaving that aside: is there any reason for that debate to block BTS tracking of backports versions in particular? Samuel makes a good argument here that this would be useful from the point of view of users of backports, and I can well imagine that it would also be helpful for package maintainers who also maintain the corresponding backports. I find it improbable that simply fleshing out the version graph a bit more would result in a significant increase of bug reports against backports. If I were a more active BTS admin then TBH I'd be inclined to just do this, but since I'm pretty inactive I thought it best to ask around. -- Colin Watson [cjwat...@debian.org]
