On Mon, Aug 24, 2009 at 11:38 PM, Peter Eisentraut<peter_e@gmx.net> wrote:
> On m=E5n, 2009-08-24 at 23:07 +0100, Greg Stark wrote:
>> I think it's a conceit to think we always fix bugs immediately.
>
> I completely agree with that one. =A0The claim that we don't need a bug
> tracker because most bugs get fixed immediately is bogus because a) it's
> not true, and b) it doesn't help people *track* bugs.
Well to be fair I think when people say that they're thinking of
projects like Mozilla which have thousands and thousands of bugs in
their bugzilla which keep getting moved from one release target to the
next until someone is tasked with closing them all because the code's
been rewritten five times and probably doesn't still have the problem.
That's mostly a problem with GUI apps where users submit tons of bugs
which will never be reproducible and may be caused by all kinds of
desktop ui interactions -- even things like viruses. I do think some
bug trackers make the problem worse by being designed around that
workflow, and I want to avoid seeing us being pushed down that road.
But we do have pending bugs. Sometimes Tom posts a patch but doesn't
commit it because he doesn't really like the fix, sometimes we aren't
sure how to solve it or there are just more important things going on.
The recent "avoid redundant detoasting" had been a pending problem
without a good solution for a long time. Tom didn't forget about that
one but there are probably a dozen or so like it and I certainly don't
remember them all. And my list probably isn't the same as Tom's list.
I would expect us to be like debian where we have a handful of bugs
which come in each week. We exchange a few emails and we close the
bug. But every now and then one comes in that we neither want to close
nor fix right away. When it comes time for a point release or a major
release we check that list to see if any of them are things we really
want to commit before releasing.
--=20
greg
http://mit.edu/~gsstark/resume.pdf
