> Russ: Policy basically already does that, doesn't it? It's not > normative (maybe it should be), but that's how I always read 10.9.1: I didn't think it forbade it. If that was the intent, I think it should be clearer. > Russ: I'd personally be happy to strengthen that to say that you > should only use dpkg-statoverride in maintainer scripts if you're > handling dynamic UID/GID file ownership. I don't see any immediate > downside to doing so; other changes should be handled directly by the > permissions set in the *.deb. That would be great. Then you could close this bug. > martin: I misunderstood the original intent and thought it was about > static uids/gids only. I brought up two things I thought were problems in this report. Then Russ explained to me exactly why chown/chmod would cause problems in the postinst script. > Don: I think this is a holdover from when xcdroast asked a debconf > question; it's probably a bug that that code is still there... file > it! I was going to, but when I re-read the debian policy, I realized that it doesn't really say you can't use dpkg-statoverride for static uid/gid, only you must for dynamic uid/gid, so I figured I would get the policy changed first. > Don: I don't know about an xsendmail package; sendmail itself doesn't > do this. (Or at least, it doesn't any more.) I checked again, and the file is xsendmail, but it is in the xmail package. > Don: The reason why I'm asking this question is because before policy > is changed into a must requirement, someone should have found out > which packages will be instantly RC buggy. When I originally filed this bug, I didn't intend to make packages which use dpkg-statoverride unncessarily to be RC buggy. I suppose, though, that that is what would happen if this became a required directive. I'm pretty sure there are a lot of packages that do this. Dozens, maybe? There are the two I mentioned, and probably w3m-img, which Russ mentioned above (video is static gid). Also, I think that the debian policy manual should explain why dpkg-statoverride is necessary when using dynamic uid/gids, and why it is necessary when debconf changes permissions. It's probably obvious to some, but it wasn't to me, and I am sure plenty of other people reading it. -Brandon
Attachment:
signature.asc
Description: PGP signature