From Fedora Project Wiki

Enhancing Bugbot

  • Followup discussion from previous meeting
  • Tailor output to be specific to bug triage
  • Possibly have a separate bot on a separate channel

ACTION: poelcat to propose what new format would look like

Wiki Updating

Identifying Duplicate Bugs

Bugzilla Reporting

ACTION: jmn to investigate gnome type implementation in Fedora which can better address: 1. identifying duplicate bugs 1. better reports

Suggested Triage Commitment

  • Followup to past discussion about whether it makes sense to ask people to commit to triage for a certain length of time
  • Some ideas here: http://fedoraproject.org/wiki/BugZappers/Joining #4
  • Commit for a full release cycle--from one GA date to another
  • Fedora 9 development cycle
  • Create two wiki pages

1. Current triagers 1. Past triagers

  • Refresh current triagers page at start of each release cycle
  • Create a list of functional areas people can self-assign themselves to

1. DECISON: All in attendance thought it was a good idea 1. ACTIONS: a. poelcat to create wiki pages for maintaining a triager list a. jds2001 to create a list of major functional areas and comps groups

Bug Lifecycle Discussion

1. having to have a separate bug for each issue AND release 1. bodhi auto-closing ALL bugs without the developer being able to opt out

  • lmacken may be able to work around this issue in Bodhi

DECISION: For now drop requirement of limiting one bug to one release

IRC Transcript

-!- poelcat changed the topic of #fedora-meeting to: Bug Triage Meeting 12:00
poelcat Hello Bug Triage Party People! 12:00
poelcat who is around? 12:00
* nirik is sorta here... trying to deal with the avalanche of his email. 12:01
* jmn says yo 12:02
poelcat nirik: any luck on finding out how to modify bugbot? 12:03
poelcat jmn: howdy 12:03
-!- poelcat changed the topic of #fedora-meeting to: Bug Triage Meeting -- http://fedoraproject.org/wiki/BugZappers/Meetings/Agenda-2008-Jan-30 12:03
nirik poelcat: I talked with mkant some. I think we should come up with a list of what we need it to do... 12:03
* mefoster would like to help out with bug triage if I can 12:03
nirik I haven't had a chance to look at filtering the existing channel 12:03
poelcat mefoster: welcome! 12:03
-!- snavin_ [n=snavin@124.43.41.175] has quit [Client Quit] 12:03
poelcat nirik: i'll take action to create a simple 'requirements doc' 12:04
nirik poelcat: ok, that would be great. I am sure we can get it to report what we want... either to fedora-qa, or some new channel. 12:04
poelcat okay looking at items to followup from previous meetings 12:05
poelcat we just covered one 12:05
poelcat :) 12:05
poelcat jds2001 can't make it today 12:05
-!- smooge [n=smooge@canopus.unm.edu] has joined #fedora-meeting 12:05
poelcat anyone know if inode0 has done anything with the 'university outreach' angle? 12:06
poelcat i'll take that as a "no" 12:07
poelcat next open topic was canned queries to locate bugs needing triage 12:07
poelcat snecklifter updated this page http://fedoraproject.org/wiki/BugZappers/FindingBugs 12:08
smooge what do you mean? 12:08
smooge university outreach 12:08
poelcat smooge: at a previous meeting we were talking about ways to find people to engage in helping w/ bug triage 12:09
-!- nphilipp [n=nils@155.56.26.52] has quit ["Leaving"] 12:09
poelcat the thought was that university students might be a good group of folks to talk to 12:09
smooge ah. I don't think I have seen anything from in0de on this 12:09
poelcat smooge: okay. jds2001 said he was looking into it too... i'll check w/ him later 12:10
-!- EvilBob [n=bob@fedora/pdpc.sustaining.BobJensen] has quit [Remote closed the connection] 12:11
poelcat next followup item was wiki updating (my action item) 12:11
* jds2001 is here-ish 12:11
poelcat http://fedoraproject.org/wiki/BugZappers 12:11
poelcat is the front page 12:11
poelcat adding newer links to the bottom 12:11
jds2001 poelcat: i killed the stock responses link at the bottom - i think it's redundant 12:12
poelcat and trying to split off like information into streamlined pages 12:12
poelcat jds2001: yeah, I've got a stock response for that :) 12:12
poelcat jk 12:12
poelcat that's fine 12:12
poelcat i think the tools page needs to be rounded out more 12:13
-!- axelilly [n=jdf@208-51-242-037.vitamix.com] has joined #fedora-meeting 12:13
poelcat in particular, does anyone know if there is a good resource somewhere about "locating duplicates"? 12:13
poelcat several people have mentioned that this is one of the keys to being a good triager 12:14
-!- fab_away [n=bellet@bellet.info] has joined #fedora-meeting 12:14
jmn Can we snarf the bugzilla.gnome.org script to do that? 12:14
jds2001 it is. I tend to use the specific search 12:14
jds2001 which searches summaries *and* comments 12:14
jds2001 with some well-chosen keywords :) 12:14
jmn http://bugzilla.gnome.org/duplicates.cgi 12:15
jds2001 hmm, i'm banned from gnome bugzilla :P 12:15
* poelcat added a new section to the bottom of http://fedoraproject.org/wiki/BugZappers/FindingBugs 12:16
poelcat for duplicates writeup 12:16
axelilly jmn: it looks like the script at gnome bugzilla would be very usefull in detecting and reporting duplicates 12:16
poelcat jmn: could you help us getting it working? 12:17
jmn The entire http://bugzilla.gnome.org/page.cgi?id=reports.html page is really useful. 12:17
jmn poelcat: I have failed at this once before, but would be happy to try again :). 12:17
poelcat jmn what were the pain points? 12:17
jds2001 since i can't see it, what does it do? 12:17
* jds2001 sent mail to bugmaster@gnome.org asking why our NAT address is banned :P 12:18
jmn poelcat: access and resource conflicts 12:18
axelilly jds2001: lists Bug#, Dupe Count, Change in last 7 days, etc 12:18
poelcat jds2001: must be your ISP ;-) 12:19
jds2001 lol 12:19
nirik FWIW don't they get a lot more dups because of bug buddy? or do those not go there? 12:19
jds2001 they get a *ton* of dupes via bug-buddy 12:19
jds2001 and they have scripts to handle that, iirc 12:19
poelcat jmn: "access" to what? 12:19
jmn http://bugzilla.gnome.org/reports/weekly-bug-summary.cgi is another really useful report for sensing trends. 12:19
jds2001 hmm, I've been compiling that sorta data manually jmn 12:20
axelilly are those custom reports they wrote or are they part of bugzilla? 12:20
jmn poelcat: Let's have this conversation privately. I don't want to out my friends and co-workers publicly 12:20
poelcat lol 12:20
jds2001 custom - both gnome's and rh's bugzillas are hacked up beyond recognition :) 12:20
-!- kital [n=Joerg_Si@fedora/kital] has joined #fedora-meeting 12:21
-!- JSchmitt [n=s4504kr@fedora/JSchmitt] has quit ["Konversation terminated!"] 12:21
axelilly so, will someone need to contact gnome to ask for those scripts? 12:22
poelcat okay so looks like we can figure out a few things by digging deeper into gnome setup 12:22
poelcat 1) identify dups 12:22
poelcat 2) better reports 12:22
jmn axelilly: Yeah, I'll take that, unless someone has a contact and wants to wrestle me for it (I surrender easily). 12:23
poelcat jmn: sold to the highest bidder! 12:23
* jmn bids $-20 12:24
poelcat next topic 12:24
-!- JSchmitt [n=s4504kr@p54B1221F.dip0.t-ipconnect.de] has joined #fedora-meeting 12:24
poelcat in past meeting we talked about asking people to commit to triage for a certain length of time 12:24
poelcat I sketched something out here: http://fedoraproject.org/wiki/BugZappers/Joining #4 12:24
poelcat any thoughts/ideas/criticisms? 12:25
jds2001 looks good to me :) 12:25
axelilly looks good to me 12:25
jds2001 notting had a good point that they do need to know triagers from joe schmoe 12:25
axelilly so, the time period would last for the whole cycle? 12:26
nirik sure, can give it a try and see if we need adjustment 12:26
poelcat axelilly: that was the simplest thing I could think of 12:26
poelcat we want a way to flush the list periodically so it doesn't get stale 12:26
poelcat w/o firing everyone and starting over :) 12:26
axelilly sounds good, it might be good to point out that theses are the commited zappers, but any other zappers are encouraged to help out at will, even if not currently commited. 12:27
nirik I just want a list so I can see how many people we have (approx)... ie, do we have enough to matter? 12:27
mefoster As a prospective triager: how would sign-up actually happen? Editing a wiki page, or ... ? 12:27
poelcat mefoster: yes, good point 12:27
poelcat i think this is critical to a lot of things in Fedora... it is hard to attract and retain people if there isn't critical mass 12:28
poelcat anything else on this topic? 12:29
jds2001 mefoster: i think that editing the wiki page would be good. 12:29
nirik also, would it make sense to add areas in there? 12:29
-!- DanseMakabre [i=pat@gandalf.ds5.agh.edu.pl] has joined #fedora-meeting 12:29
nirik ie, Joe Bob - libvirt, pruit bugs 12:30
-!- DanseMakabre [i=pat@gandalf.ds5.agh.edu.pl] has quit [Read error: 104 (Connection reset by peer)] 12:30
nirik or Phred Jones - general 12:30
-!- J5 [n=quintice@nat/redhat/x-175f9d6400de1513] has quit [Read error: 110 (Connection timed out)] 12:30
jds2001 nirik: probably 12:30
-!- ldimaggi_ [n=ldimaggi@nat/redhat/x-984dde3fa818a91a] has quit [Read error: 110 (Connection timed out)] 12:30
poelcat nirik: i fear if we list by package we'll have a list that scrolls like the bugzilla bug entry page :) 12:31
jds2001 lol 12:31
poelcat s/page/srpm list 12:31
jds2001 maybe major functional area 12:31
nirik yeah, true... by comps groups? 12:31
-!- fab_away is now known as fab 12:31
jmn +1 to major functional areas/comp groups 12:32
mefoster If reproducing the bug is important, it's also relevant what hardware you have 12:32
jds2001 it's not required to reproduce it....if you can, that's great, though not absolutely required. 12:32
-!- ldimaggi_ [n=ldimaggi@nat/redhat/x-2d726bbc7bcbcf9b] has joined #fedora-meeting 12:33
poelcat at least people would know my house isn't worth robbing 12:33
jds2001 hehe 12:33
poelcat who can stub out this wiki page? 12:34
-!- echapin [n=echapin@nat/redhat/x-00399ec7bc1bc2cb] has joined #fedora-meeting 12:34
poelcat okay. who can send me a list of the major functional areas and comps groups? 12:36
jmn Sorry, stopped paying attention for a second. I can take that. 12:36
poelcat jmn: thanks! 12:36
poelcat thoughts and comments to flame fest created by: http://fedoraproject.org/wiki/JohnPoelstra/BugLifeCycle 12:37
mefoster What was the big issue in the flamefest? 12:38
-!- joadams [n=joadams@nat/redhat/x-b77cd0b1ca50e254] has joined #fedora-meeting 12:38
poelcat one bug for one release for one issue 12:39
-!- llaumgui [n=llaumgui@213.161.192.247] has quit [Read error: 113 (No route to host)] 12:39
poelcat mostly the one bug for one release vs. covering multiple releases 12:39
-!- echapin [n=echapin@nat/redhat/x-00399ec7bc1bc2cb] has left #fedora-meeting ["Leaving"] 12:39
poelcat i'm not sure however, that is our call to make 12:39
-!- mthompson [n=Michelle@nat/redhat/x-4146ed8f664a9607] has joined #fedora-meeting 12:40
poelcat also question raised as to what correct CLOSED:______ state for bugs auto-closed by bodhi 12:40
-!- echapin [n=echapin@nat/redhat/x-1d2b56bed7530da7] has joined #fedora-meeting 12:41
jds2001 poelcat: why is it not our call? 12:43
jds2001 poelcat: lmacken is going to investigate automation of that. 12:43
nirik well, it affects maintainers workflows a lot too... 12:43
poelcat jds2001: who decides whether or not a maintainer has to clone bugs for each release? 12:43
axelilly I'm confused why you would have one bug over several releases? Is it if it wasn't fixed in the first released it was noticed in? 12:44
jds2001 poelcat: if lmacken succeeds, they won't have to. Bodhi will do that for them. 12:44
-!- blitzkrieg3 [n=james@65.215.14.158] has joined #fedora-meeting 12:44
jds2001 axelilly: so you can fix it separately in each branch 12:44
jds2001 axelilly: i'm confused as well, but when we changed it, flamefest ensued. 12:45
poelcat axelilly: and be explicit about which releases you fixed or didn't 12:45
poelcat nirik: are there any existing FESCo/pkg maintainer guidelines about bugzilla usage? 12:46
nirik not that I know of... 12:46
poelcat nirik: we couldn't find any which is why jds2001 created the lifecycle 12:46
nirik yep. Agreed... and this needs to get hashed out... but getting people to change the way they do things is hard... especially when it means more work for them. 12:47
jds2001 i think alot of it is that folks are resistant to change... 12:47
jds2001 but i'm not entirely sure. 12:47
wwoods If a bug exists in rawhide and F8 12:47
wwoods it's a different procedure to push out a fix for each 12:47
wwoods for rawhide you can likely just, you know, move to latest upstream 12:48
poelcat jds2001: was the issue of "one bug one release" part of the proposal FESCo voted on? 12:48
poelcat or do we need to write another one? 12:48
jds2001 poelcat: i think so, but f13 had issues. 12:48
wwoods for F8? maybe move to upstream, but you might be backporting patches instead.. and then you have to go through updates-testing 12:48
jds2001 f13 had valid issues, in that we need to allow maintainers to "opt-out" 12:49
jds2001 but if you want us to manage bugs for you, it's our way. 12:49
* nirik isn't sure still what the multiple bugs gets us... tracking for when a bug was fixed per release, but do we really care? 12:49
jds2001 am I not understanding f13's response? 12:50
wwoods so you've got two different patches and two different workflows to fix for two different releases. that's two different bugs, really 12:51
f13 nirik: many of our users do 12:51
f13 nirik: us fixing it in rawhide and then closing the bug doesn't help them on F-7 12:51
wwoods the arguments against this mostly come down to: "meh. lazy." 12:51
nirik f13: sure, but if they care, they can file a F7 bug? 12:51
nirik or if it was F7 to begin with, we shouldn't close it rawhide. 12:51
f13 nirik: what if they did? 12:51
f13 and it was fixed closed rawhide? 12:52
nirik I don't think thats an acceptable closing for that bug unless the reporter is ok with it. 12:52
f13 or what if the fix that was built across the branches didn't actually work on F-7 while it did in F-8 and devel, or it introduced some other regression on the branch? 12:52
f13 I agree that if we're going to /force/ multiple bugs, we should have easy to use tools to set them up 12:53
f13 but I don't want to force, just like I don't want to force everybody to use bodhi's methods. 12:53
nirik there can be all sorts of corner cases... in that one, fix F7 again, or close WONTFIX, or something... 12:53
f13 We have guidelines of how we think it should be done, which we urge you to follow. If you don't and continually piss of your users, we'lld eal with it. 12:53
nirik agreed. 12:54
mefoster Real-world example: I filed a Fedora 7 bug, it was closed RAWHIDE, fix isn't available yet outside RAWHIDE: https://bugzilla.redhat.com/show_bug.cgi?id=253201 12:54
buggbot Bug 253201: low, low, ---, Rex Dieter, CLOSED RAWHIDE, kdeartwork: Can see xscreensaver savers, but I can't use them 12:54
mefoster (Nothing against rdieter, by the way -- it was a fiddly thing to find!) 12:55
nirik yeah, might just be an oversight... since he is busy with kde4. ;) I would ping the bug again asking for f7 update for it... 12:55
rdieter mefoster: fix *was* made available to F-7/F-8 I thought. ?? 12:55
mefoster Not pushed yet. 12:55
poelcat so what does all this mean for what we tell bug triagers to do? 12:56
rdieter well, crap, you're right (I think I had intended to do it...). 12:56
nirik poelcat: can we re-work things so we have: 12:56
nirik a) an opt out for maintainers on the multiple bugs... and b) a automated way to file those bugs so we don't pile work on maintainers? 12:57
nirik if we do b) we might get most everyone out of a). 12:57
poelcat nirik: right... i think we've come full circle 12:58
poelcat and luke is looking into that according to jds2001 12:58
poelcat ? 12:58
jds2001 yep, maybe we move that outside bodhi to some other app 12:58
jds2001 not entirely sure what, though. 12:58
poelcat so we can move forward, okay if we table this topic? 12:58
-!- rahul_b [n=rbhalera@123.236.177.158] has joined #fedora-meeting 12:59
-!- hircus [n=michel@adsl-76-251-80-217.dsl.bltnin.sbcglobal.net] has quit ["Ex-Chat"] 12:59
* knurd wonders if we (EPEL) can start out meeting here in time 12:59
poelcat for now we will not include "one bug one release" as a triager step 12:59
poelcat ? 12:59
-!- spoleeba [n=one@fedora/Jef] has joined #fedora-meeting 12:59
* knurd wonders if we should move the EPEL meeting to #epel 12:59
jds2001 poelcat: yea, we're intruding on EPEL 12:59
jds2001 we can move to #fedora-qa 12:59
poelcat we have 10 seconds left :) 12:59
jds2001 lol 12:59
poelcat okay :) 13:00
knurd you move? we move? 13:00
poelcat i'll send unfinished items to fedora-test list for continued discussion 13:00
poelcat knurd: take it away 13:00
poelcat :) 13:00
poelcat all yorus 13:00

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

Next Meeting