[Justin Pryzby] > Are you talking about the case that the file is at some defualt-ish > "version" 1, then updated to v2, then to v3, and then the admin > manually "updates" in such a way that it happens to be identical to > v2?
I'm saying we can't tell whether you modified the file since it was last touched by gpmconfig. gpmconfig did not record this information. When we switched to ucf, we did not at the same time record what your old file's md5sum was. But if you *had* made local modifications to the file by hand, recording the md5sum would have been the wrong thing to do anyway. There's a reason we switched to ucf. To get away from this madness. > > What md5sum? The one corresponding to the woody /etc/gpm.conf on > > your computer? > woody? I haven't run woody on this machine in forever .. possibly > never. The file might have been "the one woody used" though, I > guess. The interval between woody and sarge is not all that interesting. gpm was changed to use debconf and ucf very late in the sarge release cycle. So when I say sarge, I mean sarge as of the last 6 months or less before its release. If you were running sarge before that, your gpm version at the time would have been very similar to woody's. This discussion isn't going anywhere. We have marked the bug 'wontfix' for, we think, good reasons. I think the best way to change our minds is to post a tested and working patch. I have my doubts that such a patch is even *possible*, without hitting a sort of inverse of this bug (hitting a false negative instead of a false positive, and overwriting local admin changes silently). Which would be quite a bit worse.
signature.asc
Description: Digital signature