Re: developer bug fixing statistics now available

Martin Jansen wrote: > On Sat Dec 16, 2006 at 11:3112AM +0100, Lukas Kahwe Smith wrote: >> Martin Jansen wrote: >>> On Sat Dec 16, 2006 at 12:0559AM -0800, Justin Patrin wrote: >>>> compared to you, of course, and I suspect I won't stay there long. I >>>> find it very likely that many people didn't assign bugs to themselves >>>> before closing, as I did. >>> Exactly my thoughts. When I first looked at my bug rank (#68) I was >>> pretty surprised, because I have closed much more bugs in the past >>> years. But then I realised that those numbers are based on the >>> "assigned to" field and suddenly they make much more sense. >> Not sure if this is really practical. We do not log who last changed the >> status to a closed state? > > I don't know much about the bugtracker's DB schema, but the following > seems to be what we are looking for: > > mysql> SELECT COUNT(c.id) AS closed_bugs > -> FROM bugdb_comments c, users u, bugdb b > -> WHERE b.status = 'Closed' AND b.ts2 = c.ts > -> AND c.email = CONCAT(u.handle, '@php.net') > -> AND u.handle = 'mj' > -> GROUP BY u.handle > -> ORDER BY ts DESC; > +-------------+ > | closed_bugs | > +-------------+ > | 149 | > +-------------+ > 1 row in set (1.51 sec) > > mysql> > > Greg, can you confirm that this is correct? Hi, This query is not quite accurate, which you can tell by changing the Group by to: GROUP BY c.id HAVING closed_bugs > 1 For me, there are 2 bugs that are counted twice. In addition, I've often marked bugs as closed that were fixed by someone else because they forgot to mark them as closed. Assignment is far more clear and accurate. I understand why some think this is unfair, but I think the benefits outweigh the inconvenience. For instance, should there be a change of developer (something that will happen more often as PEAR ages), tracking down who actually fixed a problem that has resurfaced is much easier if the bug is assigned to the developer. The query above increases the number of bugs I've "fixed" to 357, which is 40 more than I've fixed. I have always assigned myself to bugs that I've fixed. I ran another query to be sure. Examples of bugs I've closed that I didn't fix: http://pear.php.net/7290 http://pear.php.net/8176 (this one is assigned to the real fixer, davidc) http://pear.php.net/9006 (ditto) http://pear.php.net/6322 (also the same) http://pear.php.net/5018 (you sense a pattern?) http://pear.php.net/4352 (this is fun) http://pear.php.net/1895 (***) http://pear.php.net/3417 (** *) Your query, however, is useful for finding bugs that have been fixed but not yet assigned to anyone, I put up a page at http://pear.php.net/~greg/helpdevs.php that should simplify the search for closed bugs. This page takes forever to load, and can be simplified by passing in your handle as in http://pear.php.net/~greg/helpdevs.php?handle=mj Yes, I secured the script, so XSS is not possible, and database manipulation is also not possible Greg