Package: bugs.debian.org Severity: normal Consider bugs such as #588087 and #593086 which I reported some years ago against then-current gcc-4.4 and which were recently closed when gcc-4.4 was removed from Debian.
However, the bugs haven't actually been fixed. I've checked that both still exist in the current (as of this writing) gcc-4.9.2. Now, I think I understand why e.g. gcc-4.4 and gcc-4.9 must be distinct packages in Debian. But their contents are not really distinct (such as gcc vs. clang) -- they're still different versions of the same code base. So the default assumption should be that a bug that hasn't been fixed in one version is still present in newer versions (just as is the case for other software that doesn't need distinct packages for different versions). Therefore, I'd ask that for such packages, when old versions are removed, open bugs are not automatically closed, but automatically moved to a newer version of the software. Besides gcc this should apply to many libraries, among others. Sure, this risks occasional false positives, but this seems much better than the many false negatives we have currently. (Just as an example: Together with the two above-mentioned bugs, a few dozen other bugs were automatically closed, as can be seen in the close message; I'd bet many of them are actually still open in current gcc versions. FWIW, I've reported my two bugs upstream meanwhile, so it's not about these, but about the general issue.) This is important to me because many of the bugs I've reported in Debian so far have been sitting idly in the bug database for years (and looking at other packages, this doesn't seem specific to my reports, but rather the norm in Debian, unfortunately). At least the database serves as a kind of documentation of existing bugs (also for other users who don't need to report the same issues again). However, when bugs are simply removed after some time on a (IMHO) technicality, even this little usefulness disappears, and I really wonder why I should report bugs at all anymore.
