17:03   MootBot Meeting started at 17:03. The chair is liw.
17:03   MootBot Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE]
17:04   liw     we don't have much of an agenda prepared
17:04   liw     [TOPIC] post-release bug status
17:04   MootBot New Topic:  post-release bug status
17:04   heno    I just posted a straw man one now
17:04   liw     https://wiki.ubuntu.com/QATeam -- right, at the bottom of the page now
17:05   heno    my favourite bug graph this week http://people.ubuntu.com/~brian/testing_graphs/plots/update-manager-month-unconfirmed.png
17:05   liw     what's the situation with bugs now, almost a week after the release?
17:05   liw     uh oh, that doesn't look good
17:05   heno    poor mvo really gets flooded just after release
17:05   liw     mvo?
17:06   heno    we should help out with these I think
17:06   bdmurray        I think some of them are invalid, just temporary issues with mirrors, etc
17:06   heno    bdmurray: can you try some bughelper duplication-finding magic on those?
17:06   mvo     liw: hello
17:06   liw     do I read that graph correctly: it shows the number of new, untriaged bugs each day?
17:07   heno    are there lots of apport-generated ones?
17:07   mvo     that is quite possible
17:07   mvo     yes, those are mostly auto-generated
17:07   mvo     we have e.g. ~50 duplicates for a adept bug, let me dig the bugnumber
17:08   heno    liw: right, see http://people.ubuntu.com/~brian/testing_graphs/update-manager.html for the interactive version
17:08   mvo     https://bugs.edge.launchpad.net/ubuntu/+source/update-manager/+bug/154771
17:08   bdmurray        153500 maybe?
17:08   ubotu   Launchpad bug 154771 in update-manager "Kubuntu 7.10 Upgrading to Kubuntu 7.10 and after gives error and says Distribution is Up To Date" [Undecided,Confirmed]
17:08   liw     is there automation for finding duplicated, automatically reported bugs?
17:08   mvo     bdmurray: yes, that is the other one
17:08   heno    http://people.ubuntu.com/~brian/testing_graphs/ for other packages
17:08   MootBot LINK received:  http://people.ubuntu.com/~brian/testing_graphs/ for other packages
17:09   mvo     the problem with the update-manager bugs is that each update problem gets to this package first
17:09   bdmurray        liw: I believe ther is for crash reports
17:09   pedro_  liw, for stacktraces yes, don't know if there's one for that kind of issues
17:09   heno    mvo: so you don't sound too worried
17:09   mvo     even if its really a bug in e.g. a postinst
17:10   mvo     its certainly not good
17:10   mvo     our upgrade is fragile, mostly because of the fact that for ~2000 installed package ~4000 maintainer scripts gets run and some are really fragile
17:11   mvo     it used to be a upgrade error if e.g. invoke-rc.d can't stop a daemon
17:11   mvo     -> maintainer script failre, upgrade error
17:11   liw     mvo, do you think the fragility could be lessened by more vigorous pre-release testing using automated tools?
=== asac___ is now known as asac_
17:11   bdmurray        mvo: did you find anything out about the medibuntu repo ones?
17:12   heno    mvo: I guess you and liw should talk a bit about upgrade testing at UDS
17:12   mvo     same for .e.g python. if python-central is not able to compile py modules (e.g. because of locally installed stuff) the upgrade may fail
17:12   mvo     liw: to a certain extend yes
17:12   mvo     liw: one big issue is that we do not test universe packages enough for upgradability
17:12   heno    mvo: what's the limiting factor on test coverage; coding tests or HW to run on?
17:12   Lutin   bdmurray: what are the medibuntu-related ones ?
17:12   cjwatson        we made python a bit less fragile in the gutsy cycle by at least ensuring that debconf doesn't fail to configure in that case
17:12   liw     mvo, we should test universe packages as well, certainly
17:13   mvo     liw: but then its a problem with our paradigma, maintainer script fails -> upgrade fails
17:13   heno    imagine if we wanted to run a silly amt of permutations of upgrade scenarios
17:13   mvo     heno: time
17:13   heno    mvo: as in dev time?
17:13   bdmurray        bug 153948 is an example
17:13   ubotu   Launchpad bug 153948 in rosetta "Launchpad translations take a plain '%' char as a format string when it's not a format string" [High,Confirmed] https://launchpad.net/bugs/153948
17:13   mvo     cjwatson: yes, I think we do better than before (with invoke-rc.d as well)
17:13   mvo     bdmurray: no, sorry
17:13   bdmurray        Lutin: bug 153958 is an example
17:13   ubotu   Launchpad bug 153958 in update-manager "update-manager problem when upgrading from Feisty to Gutsy" [Undecided,Incomplete] https://launchpad.net/bugs/153958
17:13   mvo     heno: dev time to setup the tests
17:14   mvo     heno: the currently available auto-upgrade tester could test universe as well, the code is there
17:14   Lutin   bdmurray: this url has been deprecated for months
17:14   mvo     heno: but it takes time to generate the base image for the upgrade test and especially in universe quite a few packages fail to install
17:14   heno    I wonder how many of the bugs we are getting involve universe packages
17:14   bdmurray        Lutin: well, apparently not everyone knows that
17:15   liw     it would be possible to set up piuparts for example to break invoke-rc.d and see what happens, might be an interesting thing to do
17:15   mvo     heno: the other issue is that the upgrade takes a lot of time with  a lot of packages
17:15   mvo     heno: I did tests with ~12000 installed packages and dpkg eat ~500 mb of real mem
17:15   Lutin   bdmurray: unfortunately :/
17:16   liw     mvo, that's thankfully a problem that can be solved by throwing hardware at it
17:16   mvo     liw: if invoke-rc.d is broken, every upgraded daemon fails and that makes the upgrade fail (or at least it gives scary errors)
17:16   heno    mvo: right, I'm thinking it would be cool to combine this with other tests, like server testing that needs HW anyway
17:16   mvo     right
17:16   liw     mvo, right, the point would be to see how much stuff breaks by that kind of thing
17:17   mvo     right
17:17   heno    so just to wrap up the topic:
17:17   Lutin   bdmurray: does this kind of bug _break_ the upgrade ? I wouldn't believe it
17:17   mvo     I didn't managed to analyze the bugreports fully yet, but from what i have seen over the weekend it seem a lot of errors are releated to getting the upgrades because of server overload (especially mirrors) and then packages breaking during the upgrade
=== asac_ is now known as asac
17:18   heno    bdmurray, pedro_, ogasawara: would you say the bug report impact from the release has been noticeable, but not alarming?
17:18   ogasawara       heno: yes
17:18   pedro_  i thought it would be worse
17:19   pedro_  we indeed get a lot of reports during the weekend but nothing alarming
17:19   heno    ok, great
17:19   heno    next topic?
17:19   bdmurray        my concern is us fallling further behind in the next 2 weeks though
17:19   liw     [AGREED] the bug report impact from the release has been noticeable, but not alarming
17:19   MootBot AGREED received:  the bug report impact from the release has been noticeable, but not alarming
17:20   liw     [TOPIC] mobile platform test cases
17:20   MootBot New Topic:  mobile platform test cases
17:20   heno    bdmurray: with UDS and AllHands getting in the way, true
17:21   heno    any one here have experience with the mobile platform?
17:21   liw     not me
17:21   ogasawara       heno: I don't
17:21   pedro_  i dont...
17:22   bdmurray        I've used a mobile device before. ;)
17:22   heno    we need some better test plans there; But I'll email some people
17:23   heno    (it was worth asking :) ) ; ok, next topic
17:23   liw     [TOPIC] testing plans for hardy
17:23   MootBot New Topic:  testing plans for hardy
17:24   heno    I seem to remember we had various long-term issues we wanted to discuss last week, but cut short
17:24   heno    hence the general testing and bug topics on the agenda
17:25   heno    has everyone going to UDS looked at the proposed topics and subscribed?
17:25   liw     I'm trying to find the log from the previous meeting
17:26   liw     I've looked at and subscribed to all relevant UDS topics I know of
17:26   heno    you need to subscribe to the specs to be scheduled for the meetings
* ogasawara needs to subscribe then
17:26   heno    yeah, please do that before UDS starts
17:26   liw     http://kryten.incognitus.net/mootbot/meetings/ubuntu-meeting.log.20071017_1719.html -- that should be our previous meeting
17:26   MootBot LINK received:  http://kryten.incognitus.net/mootbot/meetings/ubuntu-meeting.log.20071017_1719.html -- that should be our previous meeting
17:27   heno    I'll go through the list later today and subscribe other distro people to things that seem relevant as well
17:28   liw     nothing much in that log, though
17:28   liw     I'll re-iterate my intention of concentrating, to the extent possible, on automating tests, starting with using existing tools of various kinds
17:29   liw     anyone else?
17:29   heno    right; system-level stuff before new year (est.) and then make a start on gui-based testing after new year
17:29   liw     anything else?
17:29   liw     yeah, like that
17:30   heno    I have another topic: QA resource page
17:31   liw     let's do the remainign one in the wiki first
17:31   liw     [TOPIC] bug management plans for hardy
17:31   MootBot New Topic:  bug management plans for hardy
17:31   liw     I have the feeling we should try to keep the number of untriaged bugs and the number of open bugs as low as possible -- it's easier to do triage if there's only a small number of bugs to work on
17:31   liw     but I haven't participated in triage at all yet
17:32   heno    a noble goal
17:33   liw     let's see if it is a practical one as well
17:33   liw     anything else about bug management?
17:33   heno    that either means cloning more triagers, tightening policy (closing more bugs), making it more difficult to file bugs, etc.
17:33   bdmurray        I think writing some bughelper clue files for previous high profile bugs to identify duplicates and cleanup some packages might be worthwhile
17:34   heno    which may all be valid choices
17:34   heno    bdmurray: 'clean up some packages' -> the clue files for them?
17:36   bdmurray        Like mvo mentioned an update-manager bug earlier, lets write a clue file for that and other update-manager bugs and see how many dupes we can find and what affect that has on the number of update-manager bugs
17:37   heno    bdmurray: right, this is a good time for that as we're not under the pressure of a release
17:37   davmor2 Sorry I'm late :(
17:37   bdmurray        well, we also need to keep an eye out for sru candidate bugs
17:37   bdmurray        but I think some of the reports may help with that
17:38   davmor2 SRU?
17:38   bdmurray        stable release update
17:38   davmor2 ta
17:40   liw     anything else on this topic?
17:41   heno    next?
17:41   liw     [TOPIC] QA resource page
17:41   MootBot New Topic:  QA resource page
17:41   liw     heno?
17:41   heno    I just posted this page https://wiki.ubuntu.com/QATeam/Resources
17:42   heno    I'd like us to bring up a few on-line resources about QA at these meetings each week
17:42   heno    there is lots we can all learn out there
17:42   heno    as a new team I think that would be valuable
17:43   liw     I agree
17:43   heno    We could all just add links to that page, but I'd like them to be mentioned here in some way
17:43   heno    a long list of links is not that useful
17:44   liw     mentioning them here would be good
17:44   liw     also, perhaps adding a few words of description to each link?
17:44   liw     [LINK] https://wiki.ubuntu.com/QATeam/Resources
17:44   heno    I wonder if we could ask each person to bring one item to the meeting each week and say two lines about it?
17:44   MootBot LINK received:  https://wiki.ubuntu.com/QATeam/Resources
17:45   liw     I'd rather have people mention anything they happen to find
17:45   heno    it could be an article, research paper, good blog entry about QA, video, etc.
17:45   liw     in a repeating topic that happens every week
17:45   heno    ok, that works too
17:45   davmor2 sounds good :)
17:46   liw     I'll add that topic to the agenda now so we don't forget
17:46   bdmurray        And what is the purpose?  I'd have to stop doing something else to do this.
17:46   heno    these are not required reading of course, just recommended :)
17:46   heno    bdmurray: if you don't feel it helps you then you don't need to
17:47   heno    you might want to glance at what is being posted though
17:47   bdmurray        I'd certainly like to, but I'm quite uncertain where to fit it in
17:47   liw     I follow programming.reddit.com and they fairly often have testing stuff there, so I'd be more than happy to mention anything there
17:48   heno    bdmurray: you mean about finding links or looking at other people's submissions?
17:48   bdmurray        heno: both. ;)
17:48   davmor2 I think an useful links that further the cause can only be a good thing :)
17:48   heno    right, the google tech talks are quite long, about an hour
17:50   heno    bdmurray: you could post an informative bug search from your weekly work that new community people might find useful
17:51   bdmurray        heno: okay, I'll see what I can come up with
17:53   liw     so: anyone is welcome to bring what they have, and anyone is welcome to make what use they can of what is brought -- are we in agreement? :)
17:53   heno    I guess this is something I'm used to fro my university days; lunch-time talks
17:53   heno    I think so
17:53   liw     any other topics?
17:53   heno    any other business?
17:53   heno    :)
17:54   heno    oh, the open week sessions went really well!
17:54   bdmurray        When will the next meeting be?
17:54   heno    bdmurray, pedro_, liw: well done!
17:54   liw     bdmurray, good point
17:54   heno    hm, we should schedule one during UDS
17:54   davmor2 yes,  Is there a way that we can log our hardware fully?  So that it gets added to bugs and reports save having to add it each time?
17:54   heno    with in-person and voip
17:55   liw     heno, would it be best to do that scheduling on site when we know what the uds schedule looks like?
17:55   heno    right, should we punt that decision to Monday?
17:55   liw     davmor2, the reportbug(-ng) program could perhaps be adapted to report bugs to launchpad
17:56   liw     heno, I'd say we should punt to Monday
17:56   heno    davmor2: I'd like to start using hwtest for this
17:56   davmor2 any links for hwtest?
17:56   heno    it produces an XML file that we can store different places
17:57   heno    it should be an ubuntu app
* heno looks for it
17:57   heno    when did we remove that from the menus?
17:58   bdmurray        I thought it was run once thing
17:58   davmor2 okay cool.  I'm thinking that in testing senario's it may speed things up if every tester with intel cpu is having issues where as everyone with amd's are fine same for 32 and 64 bit etc.
17:58   bdmurray        That is why it isn't in the menu
17:59   bdmurray        However, I don't recall seeing it when doing Gutsy installs
17:59   heno    I think it was there in the old days, but removed for the reason you stated
18:00   heno    let's take that question to the mailing list ...
18:00   liw     anything else for this meeting?
18:00   heno    bdmurray: should that be announced somewhere, should we just start using it?
18:01   davmor2 heno are you on about ubuntu device database?
18:01   bdmurray        heno: that what?
18:01   heno    davmor2: yes
18:01   heno    bdmurray: ubuntu-qa mailing list
18:01   davmor2 it's hidden in system tools
18:01   heno    (sorry, unclear)
18:02   davmor2 ubuntu-qa has a mailing list ?
18:02   bdmurray        davmor2: yeah, I haven't looked at it in a bit and no is subscribed but we thought we would start using it for this
18:03   bdmurray        so I need to double check the config then we can start using it
18:03   heno    bdmurray: can you post some info to the bugsquad list and place a link in the wiki, etc. when it's ready?
18:04   heno    or we can just subscribe and go?
18:04   stgraber        sorry, was afk for quite a long time :)
18:05   bdmurray        why don't we just start using it and see what happens
18:05   heno    davmor2: let's play with that app a bit and brainstorm on the new list
18:06   heno    ok cool
18:06   davmor2 where's the sign up point?
18:06   heno    anything else?
18:06   liw     how does one subscribe?
18:06   bdmurray        https://lists.ubuntu.com/mailman/listinfo/ubuntu-quality
18:07   davmor2 request sent :)
18:08   pedro_  subscribed :-)
18:08   heno    ok, I think we're done
18:08   liw     subscribed
18:08   liw     #endmeeting
18:08   MootBot Meeting finished at 18:08.

MeetingLogs/QATeam/20071024 (last edited 2012-12-21 09:47:35 by javier-lopez)