apt-listbugs should not list fixed bugs. I get: serious bugs of udev (215-17+deb8u1 → 220-7) <Resolved in some Version> b1 - #787364 - ifup@.service does not start any more with systemd 220 (Fixed: systemd/220-3) Merged with: 787263 Summary: udev(1 bug) Are you sure you want to install/upgrade the above packages? [Y/n/?/...] while this bug is fixed in 220-3, which is < 220-7, and not found in versions larger than 220-3. So, what's the point of listing it? Note: on the bug page: Found in versions systemd/220-2, systemd/220-1 Fixed in version systemd/220-3 Done: Martin Pitt <mpitt@debian.org>
On Thu, 18 Jun 2015 18:16:23 +0200 Vincent Lefevre wrote: [...] Hello Vincent, thanks for your bug report! Indeed, apt-listbugs is designed to not list bugs which appear to be fixed in the target package version. Mmmmh, this does not seem to be true. Running: $ apt-listbugs -d list udev/220-7 [...] Start checking: 787364 .. checking ver 220-7 against fixed:systemd/220-3 found:systemd/215-18 systemd/219-10 systemd/220-1 systemd/220-2 systemd/220-4 systemd/220-5 ... bug found between the latest fixed version and this version: I am buggy in conclusion, for bug 787364 comparing fixed:[systemd/220-3], found:[systemd/215-18 systemd/219-10 systemd/220-1 systemd/220-2 systemd/220-4 systemd/220-5], cur_ver: and new_ver:220-7 results in removal:false [...] I see that the bug appears to be found in a version larger than 220-3 (that is to say, 220-5). The point of listing it is that it does not appear to be fixed in version 220-7 ... :-? At <https://bugs.debian.org/787364> I see: Found in versions systemd/220-1, systemd/220-2, systemd/215-18, systemd/220-4, systemd/220-5, systemd/219-10 Fixed in version systemd/220-3 Done: Martin Pitt <mpitt@debian.org> And the linked version tracking graph seems to confirm that version 220-7 is buggy: <https://bugs.debian.org/cgi-bin/version.cgi?package=udev;collapse=1;absolute=0;info=1;fixed=systemd%2F220-3;found=systemd%2F220-1;found=systemd%2F220-2;found=systemd%2F215-18;found=systemd%2F220-4;found=systemd%2F220-5;found=systemd%2F219-10> Please clarify why we seem to see two different situations...
[...] I was looking at the merged bug 787263, which still contains: Found in versions systemd/220-2, systemd/220-1 I thought that merged bugs shared the same "found" list, otherwise this doesn't seem to really make sense.
On Fri, 19 Jun 2015 00:29:15 +0200 Vincent Lefevre wrote: [...] Apparently, the version tracking info is not necessarily equal for all the bugs that are merged with each other. Quoting from <https://www.debian.org/Bugs/server-control#merge>: [...] | When reports are merged opening, closing, marking or unmarking as | forwarded and reassigning any of the bugs to a new package will have | an identical effect on all of the merged reports. [...] Nothing is said about adding found or fixed versions...
This page has contradictory information. If this were the case, forcemerge would not propagate found and fixed versions. But they are propagated as seen here: https://bugs.debian.org/cgi-bin/bugreport.cgi?msg=86;bug=764322 and looking at the details on each bug page.
[...] If that documentation page is unclear, this should be reported as a separate bug against the correct (pseudo-)package, maybe against the www.debian.org pseudo-package or rather against the bugs.debian.org pseudo-package... I propose that this bug report is reassigned to debbugs, requesting a modification of the BTS to automatically join found and fixed version lists, when merging bugs and to keep those lists in sync among merged bugs. Of course, the documentation should be changed accordingly. I'll reassign this bug report myself, if you don't object. Please let me know what you think. Thanks!
I completely agree. Thanks,
Control: tags -1 - moreinfo Control: reassign -1 debbugs 2.4.1.1 Here we go! Dear debbugs maintainers, Vincent Lefevre reported a bug against apt-listbugs (which I maintain). During the conversation that followed, it was found out that the Debian BTS does not guarantee that merged bug reports share the same version tracking info. In other words, it is possible that two merged bugs have different "found" and "fixed" version lists. One example is bug #787364, which is merged with bug #787263. The former [1] is Merged with 787263 Found in versions systemd/220-1, systemd/220-2, systemd/215-18, systemd/220-4, systemd/220-5, systemd/219-10 Fixed in version systemd/220-3 while the latter [2] is Merged with 787364 Found in versions systemd/220-2, systemd/220-1 Fixed in version systemd/220-3 [1] https://bugs.debian.org/787364 [2] https://bugs.debian.org/787263 The result is that the BTS claims that one bug is fixed in current unstable (systemd/220-7), while the other bug is still present. Does this make sense? I mean, the two reports are merged with each other, hence they are considered as two duplicate reports for the same issue. This issue is either present or absent in a given version of the package under consideration: how come the BTS claims that this issue is present and absent at the same time in the same version? Unless I misunderstand something, this should not happen. If you agree that this is an inconsistency, I think that debbugs should guarantee that merged bug reports share the same "found" and "fixed" version lists. This could be done by automatically joining the "found" version lists of the bug reports being merged (just like what is done with tags). Then, any alteration performed to the "found" version list of one bug report should be automatically propagated to the other merged bug reports. The same should hold for "fixed" version lists. Obviously, the documentation [3] should be updated accordingly. [3] such as: https://www.debian.org/Bugs/server-control#merge Please fix this bug. Thanks for your time! Bye.
On Sat, 20 Jun 2015 10:35:48 +0200 Francesco Poli wrote: [...] [...] I forgot to re-title the bug report... I am doing so now. Bye.