Calendar:QA Chat:2007-07-19:Log

[18:40] <ctalbert> Hi folks, sorry I'm late.  Got stuck in traffic
[18:41] <ctalbert> Is the QA Chat happening?
[18:44] <Andreas> Is there any important at the agenda?
[18:45] <ctalbert> Not to much.  Someone figured out the issues behind the disapearing events bugs
[18:45] <ctalbert> We have only one qa discussion bug
[18:46] <ctalbert> We are scheduled to have a test day next tuesday.  Any thoughts on topics for it?
[18:46] <mschroeder> Hi Clint
[18:46] <ctalbert> Hi Martin
[18:47] <ctalbert> mschroeder: Have you had any luck looking into the chrome automated test framework?
[18:47] <mschroeder> ctalbert: No. I didn't have the time.
[18:47] <mschroeder> but it's on my todo list
[18:48] <ctalbert> cool
[18:48] <mschroeder> What will be landed until next Tuesday?
[18:48] <mschroeder> The new event dialog?
[18:48] <ctalbert> I'm not sure.  I tend to doubt that will make it.
[18:49] <ctalbert> Stuff is coming in quite fast right now. 
[18:49] <mschroeder> The today pane landed today
[18:49] <ctalbert> Cool
[18:49] <ctalbert> Maybe we should just have a test day on the new UI
[18:50] <mschroeder> I think our Alarms also need some hardcore testing. But I think, that would produce a lot of new bug reports. :(
[18:51] <ctalbert> That would be OK.  The alarms do seem to have digressed a bit.  They have gotten so bad on my laptop that I am consistently getting alarms 20 or 50 minutes late.
[18:51] <ctalbert> I just rebuilt lightning yesterday to see if a current build will help (I'm running 0.5)
[18:51] <ctalbert> sounds like I should rebuild again to get the today pane :-)
[18:52] <ctalbert> Let's do alarms, and maybe we can also try to focus on the QA Wanted defects, there are quite a few.
[18:53] <mschroeder> Maybe we should compile a list on the wiki with prerequirements to check the qawanted bug... some are very specific.
[18:53] <ctalbert> Interesting idea.   A list like this:
[18:54] <ctalbert> If you have OpenXchange setup running through a proxy, then please help us test bug <blah> and <blah>
[18:54] <ctalbert> That sounds like a pretty good idea
[18:55] <mschroeder> well... that's even better. I only thought of a plain list.
[18:55] <ctalbert> I am should have more time next week to devote to the calendar project.
[18:56] <mschroeder> wonderful :)
[18:57] <ctalbert> I'll put together the wiki and the blog post later today.  Mshcroeder, do you want to make a first run through the QA Wanted bugs and come up with prerequisites?  
[18:57] <mschroeder> yes
[18:57] <ctalbert> Thanks, that'd really help
[18:58] <ctalbert> Andreas: is there anything from the prototype/WCAP side you want the test day to focus on?
[18:58] * hubert has joined #calendar-qa
[18:59] <ctalbert> Hi hubert
[18:59] <Andreas> Maybe the today pane
[18:59] <hubert> hi
[19:00] <ctalbert> Andreas: cool, I am going to mention that for certain.  I'll also mention the new mail/calendar switching model too.  The ability to see a new UI will help draw people to come to the test day and try out a build (I hope).
[19:01] <ctalbert> We have only one QA Discussion bug: bug 386991
[19:01] <firebot> ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=386991 maj, --, ---, nobody@mozilla.org, UNCO, Agenda created in Linux does not appear in Windows. Reverse is also true.
[19:01] <ssitter> ctalbert: there already was a blog post about the mail/calendar switching mode
[19:01] <mschroeder> i added it some minutes ago ;)
[19:02] <ctalbert> ssitter: Sweet, I haven't checked planet yet.
[19:03] <ctalbert> mschroeder is blogging for us now?
[19:03] <ctalbert> that rocks!
[19:03] <mschroeder> nono
[19:03] <mschroeder> I meant the bug to the qa discussion list
[19:03] <ctalbert> Oh, ok
[19:03] * damian has joined #calendar-qa
[19:03] <ssitter> ctalbert: Simon blogged about three weeks ago
[19:03] <ctalbert> It's stupidly difficult to get a blog account on our blog, hence my excitement
[19:04] <ctalbert> geez, I'm behind the times.
[19:04] <ctalbert> So, bug 386991...
[19:05] <ctalbert> Sounds like Sebo and Ssitter already solved the issue.  Seems like it should go to INVALID or WFM
[19:06] * mschroeder votes for INVALID.
[19:06] <ctalbert> So does ssitter, based on his comment :-)  I will mark it as invalid
[19:08] <ssitter> maybe we should provide a universal universal builds  (win + linux + mac) next time in the distribution section?
[19:08] <ctalbert> I think that would be a good idea.  Most people expect XPI's to be platform agnostic because so many of them are javascript only
[19:09] <ctalbert> Of course, the universal one would be a bigger download size, but more and more I don't think that really matters to people.
[19:09] <ssitter> the only problem would be the size > 4MByte
[19:09] <ctalbert> And if it does they can grab the OS specific one
[19:11] <ctalbert> damian: hubert: Since you joined a bit late, I wanted to let you know that we will have a test day on Tuesday covering alarms and the new calendar switching/today pane User Interface.  We will also invite people to help us test QA Wanted bugs.
[19:11] <ctalbert> Damian: Do we have test cases yet for the new UI in lightning?
[19:11] <damian> no
[19:12] <damian> you mean prototype?
[19:12] <ctalbert> No, the new mail/calendar switching and the today pane (which just landed today)
[19:12] <ctalbert> I'll help write some this weekend.
[19:12] <damian> oh, I havent' seen it, so for sure we don't have test cases
[19:13] <ctalbert> heh, me either :-)
[19:13] <ctalbert> Is there anything else folks want to discuss for the QA Chat we have about 15 minutes remaining.
[19:13] <damian> ok, so this is the best time to write test cases, then we can start implement it :)
[19:15] <damian> I have one, Clinet I wrote to you about problem with night builds, there's no place where you can get information which build should be used for testing
[19:15] <damian> Clint, sorry
[19:15] <ssitter> the test cases should probably be based on the detailed specification on wiki.mozilla.org
[19:15] <ctalbert> ssitter, I agree
[19:16] <ctalbert> damian, yes, I responded to that email didn't I? But, I didn't make a wiki page for explaining the trunk/branch/ strategy
[19:17] <damian> yes, you did of course, but is there any progress? I would write this manual/faq but my knowledge about this is, well...
[19:18] <damian> I'm not sure if we need separate page for this, but let's just include such information somewhere
[19:18] <ssitter> I read about the "in-litmus" flag today [http://quality.mozilla.org/node/305]. Maybe this would make sense for calendar too?
[19:18] * ctalbert will add the what builds to test information to the QA _Home page.
[19:18] <mschroeder> ... we should do this also for non-calendar devs checking in patches without cross-commit
[19:19] <ctalbert> ssitter, thanks, I was going to suggest this as well.
[19:19] <ssitter> damian: for Lightning this info is available from http://www.mozilla.org/projects/calendar/lightning/download.html and all the testday pages
[19:20] <mschroeder> as happened today with removed-files.in and packages-static which are now out of sync between Trunk & Branch
[19:20] <ctalbert> I think he was looking for something that detailed the differences between the trunk and the branch builds.
[19:20] <ssitter> damian: for Sunbird only the moz1.8 builds are linked from http://www.mozilla.org/projects/calendar/sunbird/download.html
[19:20] <damian> ssitter: that's not what I would like to have: I prefer format like this: ulr_to_[nightly|release]build - what is inside when use this build for testing and that's all, just short info
[19:21] <damian> ok, but will be nice to publish information why we keep others? someone may want to use national build, so there is some other option
[19:21] <ctalbert> mschroeder: I think that should be handled in a one-on-one conversation.  We don't have that many people with CVS write access that it should be a problem.
[19:21] <damian> there is also prototype: so what is the purpose of this and why/when it should be use for testing
[19:22] <ctalbert> And the person who vouched for that person should have already talked to them about it.
[19:23] <mschroeder> ctalbert: or we should simply fork our files used in the build & packaging process
[19:24] <damian> simply scenario when i would be helpful: you fioiund the bug in release version: which build you need to download to confirm this issue?
[19:24] <ctalbert> mschroeder: if the trunk continues to change, then we may have to.
[19:24] <ctalbert> continues to change its build methodology...
[19:25] <mschroeder> I think there was a statement by mvl about this in a bug.
[19:26] <damian> ctalbert: sorry that I bother, but I was looking for that information after 0.5 and I had to ask here
[19:26] <ctalbert> damian, no, it's ok.  I totally forgot to create the wiki page after I emailed you.
[19:26] <ctalbert> Ya'll have to keep me honest. :-)
[19:27] <damian> :)
[19:28] <mschroeder> ctalbert has to fight with a minotaur in parallel ;)
[19:28] <ctalbert> damian, I think we should start using the "in-litmus" flag that ssitter mentioned.  It's exactly the same as what you've been doing, except now there will be a real bugzilla keyword for it instead of a whiteboard flag.  I think we will need to file a bug to get that flag turned on for the calendar product.
[19:29] <ctalbert> mschroeder: lol
[19:29] <ctalbert> And I'm lost in a maze with only a ball of string.....
[19:30] <damian> well, let's do it!
[19:30] * ctalbert will check with ispiked about whether they turned that on for everything or just for Firefox.  (ispiked sits across from me here)
[19:31] <damian> one more question: what we are goning to do with test cases that have been already created and they are marked in whiteboard? should be updated to flag when new flag is created, right?
[19:32] <ctalbert> damian: yes, that would be ideal. 
[19:32] <ctalbert> That's what ispiked did for all the firefox ones
[19:32] <damian> otherelse we might have a problem to find then
[19:32] <damian> yeah
[19:32] <ctalbert> I think there is a way to change multiple bugs at once in bugzilla, but I've never tried to do it.
[19:33] <damian> it is but i'm not sure what it works with whiteboard if there are more than one description that we wnat to remove, problably not 
[19:33] <ctalbert> yeah, you're probably right
[19:34] <damian> and, another issue is: how to manage test cases in litmus? soon or later we can get to the point where we realize that we have many duplicates or we cover the same area twice
[19:36] <damian> even now: I found test cases that are exacly the same but created by two persons who call the same component A or B and that's why it's hard to find them amd eliminate
[19:36] <ctalbert> Yeah, I agree.  I don't have any good ideas about that.  I believe that the only way to really do that in the Litmus product now is to tediously go through each test case and check it.
[19:37] * ctalbert recalls that Marcia and Tracy did that for FFx 3
[19:38] <ctalbert> Litmus itself also needs an "admin's guide" to detail how best to organize test cases.
[19:38] * ctalbert has merely tried to copy the existing set ups.  I never really thought about whether they were the best way to organize tests or not
[19:38] <damian> to be honest: I think we lost control some time ago... do you what is covered by test cases? I don't 
[19:38] <ctalbert> No I don't
[19:39] <damian> exacly :(
[19:39] <ctalbert> I wonder if we should have devote a couple of hours a week to reviewing and reorganizing the test cases
[19:40] <mschroeder> ctalbert: http://quality.mozilla.org/howto/litmus/admin-tutorial
[19:40] <mschroeder> ;)
[19:40] <damian> first step which I would do is to split test cases for litmus and sunbird, remove this simply and complex test case - I don't see any benefit of this
[19:40] <ctalbert> Maybe we can call for help on that one
[19:41] <ctalbert> Yeah, but that doesn't get into the issues of what should be in a BFT or what should be in an FFT
[19:41] <ctalbert> And like Damian, I've thought that for *our* product it doesn't make sense to really split the two.
[19:41] <ctalbert> For Ffx and Tbird, I can see the usefulness for those.
[19:42] <ctalbert> Guys, I would absolutely love to continue this discussion, but I do have another meeting in twenty minutes.
[19:43] <damian> ctalbert: can we talk about it in week? maybe someone will find some good solution
[19:43] <ctalbert> Damian, mschroeder, whoever else is interested, can we discuss Litmus Reorganization at the next calendar QA Chat?
[19:43] <ctalbert> heh, damian beat me to it
[19:43] <damian> lovely
[19:43] <ctalbert> sounds like a plan.  
[19:44] <ctalbert> Damian, for this weekend, let's just add test cases to what we have.  I don't want to try to do a major reorganization right before a test day
[19:44] <ctalbert> that would be too hectic
[19:44] <damian> all good ideas can to put into wiki, so will be discussed in week
[19:44] <damian> right
[19:44] <ctalbert> damian: excellent
[19:44] * ssitter goes taking a look at the today pane
[19:44] <ctalbert> ok, thanks guys.