----- Original Message ----- > On Tuesday 2013-12-03 21:15 -0800, Lawrence Mandel wrote: > > I'm taking a stronger stance and suggesting that we should be able to > > wontfix bugs that likely aren't worth anyone's time or attention. As a > > concrete example, what is the value in keeping the following bugs open? > > > bug 3246 - Core::Layout:Block and Inline P3 opened 14 years 9 months ago > > I don't think this should be wontfixed; it's a valid bug, and I > think worth fixing, although as part of other architectural changes.
The last meaningful comment in this bug was posted almost 13 years ago. It may be that this is worth fixing, in which case we're perhaps not doing a good job of finding this type of worthwhile bug in bugzilla. If this has simply been a lower priority than other work for ~15 years, perhaps it's time to admit that we're not going to fix this. > I'd like to be able to use Bugzilla to track the known issues in our > code rather than being forced to copy all the data into code > comments. (At least, I sometimes would. Other times I'd rather use > a version control system for tracking bugs.) I think that's a valid point. wontfix bugs should represent known issues in our code. The difference between wontfix and new simply being that we have acknowledged that we are not going to fix a particular issue. > > In fact, there at 6925 bugs across all Bugzilla products currently > > in the new or unconfirmed state that were opened more than 10 > > years ago. I would assert that if a bug hasn't been fixed in 10 > > years it probably isn't important enough to spend time on now. We > > can always reopen or refile if the issue becomes more pressing (by > > anyone's judgement). > > I don't think that's true; both priorities and costs really do > change over time. A good example of priorities changing over time > is https://bugzilla.mozilla.org/show_bug.cgi?id=63895 . When filed, > we may have been the only Web layout engine advanced enough for it > to make sense to report the bug; today all the others have caught up > and we're the only one with the bug. Very true. However, at least in this case, if the bug had been closed at some point, the issue still would have surfaced when it became relevant due to the filing of duplicate bug 924048. In fact, there are 42 duplicates of this bug that have been filed over the years. > I also think the idea that you should wontfix bugs as a function of > age just leads to messing with the bugs of components that have been > around for a long time. See also > http://dbaron.org/log/20080515-age-of-bugs . And it sends a bad > message to the people who are interested in seeing those components > improve, reporting and commenting in bugs, etc. Many of these old > bugs are actually bugs that people care about, and that Web > developers stumble into frequently. Some of them also contain > useful information about how to fix the problem described -- > information that wouldn't necessarily be there if they were > wontfixed and new bugs filed. I tend to think we should be putting > more effort into some of them than we currently are. I agree with this. Rereading my previous post I see that it looks like I meant that we should blanket close all old bugs. That was not my intention. I do think that we should consider wontfixing a very old bug that we don't think will be fixed at some point in the near future. Bugs that contain useful workarounds should perhaps remain open. > (If there's a valid aging threshold, I think it's bugs that have > been around long enough that they've shipped in a release. I think > it's a meaningful threshold because it proves they weren't bad > enough that we had to fix them in order to ship. But I think it's > far from saying they should be wontfixed.) I think it is reasonable to at least question whether we should wontfix bugs that shipped in, say, Firefox 3. Lawrence _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
