From Fedora Project Wiki

Bug Triage Meeting :: 2009-05-26

Agenda: https://www.redhat.com/archives/fedora-test-list/2009-May/msg01189.html

Attendees

  • adamw
  • comphappy
  • iarlyy
  • mcepl
  • thomasj
  • SMParrish
  • tk009
  • Viking-Ice

Meeting Recap

The metrics are up and working. comphappy will continue to work on missing functionality, automation, bug fixes and confirm the accuracy of the statistics (they may not be correct). This code is not ready for general public testing or stress testing at this time. comphappy will announce when external testing can begin. The code is available at http://git.fedorahosted.org/git/triage.git.

The BugZappers will be setting the severity field (only) on all F12 (rawhide) bugs as part of the regular triage process going forward. Setting this field on F11 bugs is optional. At a future time (TBD), we will evaluate the impact of this procedure and modify or change it as necessary. Adam Williamson will make the modifications to the wiki regarding this new procedure.


IRC Transcript

adamw well, ok May 26 08:02
adamw i see no tk009 or poelcat May 26 08:02
adamw so...show of hands for the bugzappers meeting? May 26 08:02
adamw anyone? anyone? bueller? May 26 08:03
adamw iarlyy: comphappy: ping May 26 08:03
adamw Viking-Ice: ping May 26 08:04
comphappy yeah I am here but leaving shortly May 26 08:04
adamw ok, then, quick hack: first topic, triage metrics May 26 08:04
adamw comphappy: you have the floor, explain for the vast audience our current status :D May 26 08:04
comphappy not really I am tieing my tie as a chat May 26 08:04
adamw curses May 26 08:05
adamw fine, i'll just copy/paste what we got last night from you :) May 26 08:05
Viking-Ice pong May 26 08:06
adamw haha, i caught you. now you are officially present for the bugzappers meeting. May 26 08:06
*SMParrish here May 26 08:07
*iarlyy iarlyy here x) May 26 08:07
adamw alrighty May 26 08:08
adamw we have two bugzappers and a captured viking May 26 08:08
adamw works for me! May 26 08:08
adamw so, as I said, topic one: triage metrics. i shall now turn into comphappy BEFORE YOUR VERY EYES May 26 08:09
adamw <comphappy_> adamw: I think it is almost back up, I spent a good portion of the day patching it up, and packaging it. I will keep poking at it tonight May 26 08:09
adamw <comphappy_> adamw: the database on publictest is being filled with todays data right now May 26 08:09
adamw <comphappy_> adamw, tk009: ok we have moved forward, todays data was collected and made it up onto the graph on the front page May 26 08:09
mcepl here May 26 08:09
adamw <comphappy_> adamw, tk009: I changed a few lines and fixed the defaultdict issue May 26 08:10
adamw <comphappy_> should be working now May 26 08:10
adamw http://publictest14.fedoraproject.org/triageweb/ May 26 08:10
iarlyy is it about metrics of bugzilla in real time? May 26 08:10
adamw <comphappy_> I need to do a little more work so that everything is automated with cron May 26 08:10
adamw that should be simple but it not in the rpm May 26 08:10
adamw <comphappy_> after that am going to let it run for a while see what people break May 26 08:10
adamw <comphappy_> once I do the cron work I will roll a new rpm spec May 26 08:11
adamw which should happen tomorrow night May 26 08:11
adamw and then all of that goes into git May 26 08:11
adamw <comphappy_> give me until wed before it goes outside of bugzappers May 26 08:11
adamw ...and that's the greatest hits May 26 08:11
adamw iarlyy: yes, it's a system for dynamically producing various metrics that are of interest to the bugzappers group, from bugzilla May 26 08:11
adamw the test instance you can access at the above URL isn't running on live bugzilla yet, but on a snapshot of a single day's data, which is why the contents seem so sparse May 26 08:12
mcepl looks cool, some data would be useful though ;-) May 26 08:12
adamw so basically, we have significant progress and a working version of the system up for testing, we will enter a fairly short 'beta test' period now. May 26 08:12
adamw anyone have any points on the triage metrics? May 26 08:13
iarlyy points?! you mean suggestions or anything related? May 26 08:13
adamw yep May 26 08:13
adamw just anything to add May 26 08:13
adamw or ask May 26 08:14
adamw (though i can only answer questions i happen to know the answer to :>) May 26 08:14
iarlyy has it some relation with our group "triagers" May 26 08:14
iarlyy on fas May 26 08:14
adamw yes, brennan mentioned that the list of triagers is derived from the 'triagers' group in fas (not the 'fedorabugs' group) May 26 08:15
iarlyy like, how much bugs iarlyy triaged May 26 08:15
adamw which is somewhat significant, as it means we have to either incorporate the triagers group into the new member SOP somehow to ensure it continues to accurately reflect the group of active triagers, or come up with a different way for the metrics system to get such a list May 26 08:15
iarlyy or how much bugs iarlyy triaged in the last month May 26 08:16
adamw yes May 26 08:16
adamw big brother is watching you :) May 26 08:16
iarlyy heh :D May 26 08:16
adamw anything else? May 26 08:17
comphappy adamw: one point before I leave I will talk with jds2001 and get that FAS group automated May 26 08:17
adamw comphappy: i'm not sure that's the best way, thinking about it May 26 08:18
comphappy that way no one has to worry or add it to SOP May 26 08:18
comphappy I have ideas but I must go May 26 08:18
iarlyy much things was changed since my last access, is excellent. May 26 08:18
adamw ok, we can discuss it at a later point, thanks a lot for your work on this May 26 08:18
adamw next (and last) topic - severity/priority May 26 08:19
adamw from my pov, this is now quite simple: no-one voted within the deadline, so we should just go ahead, with mcepl's system May 26 08:19
adamw i will happily take an action item to make all the appropriate changes to the wiki for this May 26 08:20
adamw and announce it to -test-list / -devel-list May 26 08:20
adamw anyone anything to add on that? May 26 08:21
adamw hey tk009 May 26 08:21
tk009 =) May 26 08:22
tk009 sorry day job had me May 26 08:22
iarlyy ya tk009 x) May 26 08:22
adamw tk009: we've done a triage metrics update, and we're onto severity/priority now May 26 08:22
adamw you have anything to add on either topic? May 26 08:22
*mcepl is surprised that when no-one commented, I won and not Adam; but no complaints May 26 08:22
tk009 on the metrics no May 26 08:23
adamw mcepl: you won based on last week's meeting :) it was discussed there and those present favoured your proposal May 26 08:23
mcepl OK, as I said, no complaints May 26 08:23
adamw so we said that if no-one votes on the list, we'll go with the Cepl Method May 26 08:23
mcepl The Cepl Method, I beg your pardon ;-) May 26 08:24
adamw :) May 26 08:24
tk009 lol May 26 08:24
*mcepl is not sure whether definite article means anything here ... May 26 08:24
adamw it makes it sound more impressive :) May 26 08:24
adamw or we could call it The Cepl Procedure, it sounds like a vaguely disturbing hollywood movie from 1963 :) May 26 08:25
tk009 so will we be doing this to all triaged bugs or as poelcat suggested, jsut to rawhide? May 26 08:25
mcepl and of course I have to say thanks to Academy, my parents, and SAP guy who actually invented it ;-) May 26 08:25
adamw tk009: good point, any preferences? maybe we can 'require' it for f12 and later, and it's optional for f11 and earlier May 26 08:26
adamw so if you want to keep things simple and just do it for all bugs you triage, you can May 26 08:26
iarlyy sorry what is the Cepl Procedure? May 26 08:26
mcepl I would say, all bugs from now on, it is easier to get it into blood as part of the routine May 26 08:26
adamw iarlyy: one of the two ways of handling the severity/priority fields - mcepl's method is for us (bugzappers) to only set the severity field, versus mine/beland's proposal that we set both...see the mailing list threads on this topic for more details May 26 08:27
adamw tk009: iarlyy: what's your take on all bugs or just f12+? May 26 08:28
iarlyy hmm, I remember x) thanks May 26 08:28
tk009 I am fine either way May 26 08:28
tk009 that doesn't help much huh May 26 08:29
adamw hehe May 26 08:29
tk009 I would go the way everyone else is May 26 08:29
tk009 if we think all, I am fine with that May 26 08:29
adamw well only mcepl's expressed a definite opinion :) May 26 08:30
adamw iarlyy? May 26 08:30
iarlyy about severity? May 26 08:30
adamw about whether to do it on all bugs from now on, or just f12+ bugs May 26 08:31
mcepl iarlyy: the canonical source is http://article.gmane.org/gmane.linux.redhat.fedora.testers/70982 May 26 08:31
tk009 iarlly should we set severity on rawhide or all bugs we triage? May 26 08:31
iarlyy all bugs we triage sounds better for me. May 26 08:32
adamw ok, so two +1s for all bugs and two "i don't minds" May 26 08:32
adamw sounds like we can go with all bugs May 26 08:32
adamw anything further on the topic? May 26 08:33
iarlyy yep +1 May 26 08:33
tk009 just that we are clear on no revision wars with the reporter or dev May 26 08:33
tk009 if they change it for any reason after we have we leave it alone May 26 08:34
tk009 that is our policy yes? May 26 08:34
adamw for the developer, yes. May 26 08:34
adamw for the reporter, no. May 26 08:34
adamw we outrank the reporter, the developer outranks us =) May 26 08:34
adamw (for this particular case) May 26 08:35
adamw this is one bit i'm not sure which way to proceed, in fact - quite a lot of people wanted to have reporters blocked from changing the priority/severity fields at all May 26 08:35
adamw i'm not sure whether to request that change from the bz admins, or just leave it for now and revisit it if revision wars start happening May 26 08:35
tk009 I think they should be May 26 08:36
tk009 blacked that is May 26 08:36
tk009 blocked* May 26 08:36
adamw yeah, i guess i'm leaning that way May 26 08:36
iarlyy just a doubt: if a bug is opened as high and devel ( or we ) changes to medium, and reporter re-change to hight, what to do? keep it as high May 26 08:36
iarlyy ***high May 26 08:36
adamw well if reporters are blocked from setting it, they couldn't do that :) May 26 08:37
adamw i'll add to my action item to ask the bugzilla admins to block people outside the usual 'empowered groups' from adjusting priority/severity, perhaps with a link to a page explaining why May 26 08:38
adamw ok, anything else or shall we wrap this topic? May 26 08:40
iarlyy nop for me. May 26 08:41
tk009 sounds like it for that one May 26 08:41
adamw alright May 26 08:41
adamw which takes us to...open floor May 26 08:41
adamw take the mic if you want to raise anything (or just sing!) May 26 08:41
tk009 no singing May 26 08:41
tk009 so let it be written so let it be done May 26 08:41
iarlyy x) May 26 08:41
tk009 I dont have anything for the open floor myself May 26 08:42
Viking-Ice you outrank the reporter? May 26 08:42
mcepl yeah, I vote for blocking reporters ... they can comment May 26 08:42
Viking-Ice interesting May 26 08:42
adamw Viking-Ice: for the purposes of setting severity only May 26 08:42
mcepl possibly, they could be allowed to set it once, when reporting the bug May 26 08:42
tk009 speak your mind viking-ice May 26 08:43
mcepl (but of course that depends whether Dave is able to fix Bugzilla appropriately) May 26 08:43
adamw Viking-Ice: since it can only be done properly by someone with a decent overview, and it needs to be done consistently following the agreed policy...reporters are usually not in the position to satisfy either condition, otherwise bug reports would *already* have useful and accurate severity settings. which they don't :) May 26 08:43
mcepl Viking-Ice: Nemo debet esse judex in propria causa (to use the remains of my lawyer's education) ... no-one can be judge in his own case. That's nothing against bug reporters, but a general princip of universe May 26 08:45
Viking-Ice My mind on this subject ( severity/priority ) is very clear bug 495985 May 26 08:45
buggbot Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=495985 medium, medium, ---, itamar, NEW, [RFE ] Hide Severity/Priority field from everyone except the maintainer/developer May 26 08:45
*thomasj late today.. sorry.. work is horrible May 26 08:46
mcepl oh, so we do agree May 26 08:46
adamw thomasj: no problem :) May 26 08:47
adamw Viking-Ice: so, you seem to agree with the proposal (except you wanted to restrict it to maintainer only, not just maintainer+triagers) May 26 08:47
adamw we could restrict priority to maintainers-only, in fact, under the cepl procedure May 26 08:47
adamw and severity will be maintainers+triagers May 26 08:47
Viking-Ice adamw: or just have it maintainers only May 26 08:48
adamw well, the point of the current proposal is to have triagers set severity in a consistent way in order hopefully to help developers, and it's gone through an extremely long process to get here, so we're not going to dump it now :) May 26 08:49
Viking-Ice they are the only one that are qualified based on their workload etc on playing with both these fields May 26 08:49
adamw but if the trial's a failure and we dump it, it would indeed make sense to then limit that field to maintainers only as well May 26 08:49
mcepl Viking-Ice: well, OTOH, we can assume at least minimal intelligence on part of bug zappers that they won't play with it unnecessarily. May 26 08:50
thomasj adamw, thanks, by the way, i spoke to Ray about the closing Bugs and he is fine with it. Looks like i'm settled and you can add me to the Component-Page if you want. But thats just on a side note. May 26 08:50
adamw thomasj: great, you can add yourself to the page :) May 26 08:50
thomasj adamw, oh, ok :) May 26 08:50
adamw Viking-Ice: there will be a policy for how bugzappers should set severity, and we're going to review the system in a while to see if it's working smoothly and providing a benefit to maintainers. May 26 08:51
tk009 we have come to far with this to stop now. the impact of this will be faily low as poelcat pointed out. May 26 08:51
adamw okay...anything else? May 26 08:54
tk009 not from me May 26 08:54
iarlyy nop May 26 08:54
adamw alrighty then May 26 08:55
*adamw bangs gavel May 26 08:55
adamw meeting adjourned! May 26 08:55
tk009 EOM =) May 26 08:55

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!