Re: bug reporting/fixing stats

Helgi Þormar Þorbjörnsson wrote: > On Wed, 18 Apr 2007 00:34:46 -0500, Gregory Beaver wrote: > >> Lukas Kahwe Smith wrote: >>> Gregory Beaver wrote: >>>> Lukas Kahwe Smith wrote: >>>>> Gregory Beaver wrote: >>>>>> Hi all, >>>>>> >>>>>> The bug fixing stats at >>>>>> http://pear.php.net/bugs/stats.php#devs have >>>>>> been updated to implement two interesting feature requests by Tim >>>>>> Jackson. Now, you can see bug fixing activity of the last month per >>>>>> developer, as well as bug reporting/commenting activity, which can >>>>>> be used as a rough gauge of developer activity in the bug tracker. >>>>> hmm .. 35 comments on bug reports .. seems a bit low for me. >>>> Have you changed email recently? I ran a script that updated the >>>> stats based on the current email address. If you let me know the >>>> other addresses you've used to open/comment on bugs, I'll update the >>>> db. >>> ah good point .. i probably used a few different email addresses ... >>> anyways .. that seems like a flaw in the "system" .. >> One of the reasons statistics like these are best used as side >> information or to corroborate other information. The awareness that >> statistics don't give the complete picture is important. Not knowing >> any of the statistics does not make one wiser in my opinion, otherwise I >> wouldn't have implemented the feature. >> >> I'm serious - if anyone sees a clear discrepancy, I would like to know >> past emails I can use to update the database. It's a 1-liner query per >> email address, and so well worth the effort on my end. >> >> Greg > > Would it be too much problem to just have a list of emails attached to > the users table ? :) comma sep field, well probably a lot more hazzle > than your way, I'll mail you my emails later tonight Hi Helgi, In terms of attaching a list of emails, no, don't think we're going to do that :). It would require substantial changes to the way things are mailed, and you would get copies of bugs, patches, etc. sent to all email addresses. I don't think there is anyone who wants that to happen, and the problem of statistics only applies to the past. All future bugs are auto-correlated by handle with a user, so that the email address can change and the bug will still be credited to the proper person. > Another interesting thing to implement, I've closed countless bugs in the > past, pretty sure it might be over what the is shown on the site > currently, wouldn't it be best that if I (my email address) closed a bug > report but no one is assigned than it will be counted that I closed it ? > Might give use somewhat more accurate stats if this is not being done > already :) I already ran a script that assigned closed bugs to the last comment, and then manually went through the 1000 bugs that were left over that had not been assigned, and was able to determine the fixer for all but about 13 bugs. So, the fixed bugs tally is as accurate as it will get, unless the people who fixed the few remaining unassigned closed bugs go in and assign it to themselves. Yes, it took a long time, and no, I don't want to repeat that effort in the near future if possible :) Greg