Calendar:QA Chat:2006-04-26:Log

From MozillaWiki
Jump to: navigation, search
[11:30]         ctalbert        Looks like it's about time to start the QA Chat meeting
[11:31] 	ctalbert current agenda, so we'll just reuse the one from
 last week for the links:
[11:32] 	ctalbert	We only have one remaining blocker before 0.5 RC is ready 
to go.  I am reviewing that blocker right now.
[11:32] 	ctalbert	er..reviewing a patch for that blocker
[11:33] 	bvdbos	is there a patch already, does the save before send work?
[11:33] 	ctalbert	there is a patch
[11:33] 	ctalbert	and give me one sec...
[11:34] 	ctalbert	yes, the event is formatted correctly.
[11:34] 	ctalbert	Daniel: Looks like the patch works :-)
[11:35] 	daniel	ok
[11:35] 	firebot granted review for attachment 262888 
on bug 377641.
[11:36] 	firebot	Bug maj,
 --, ---,, NEW, Outlook 2003 does not read iTIP/iMIP invitation
  because of missing UID in VEVENT
[11:36] 	ctalbert	So, as I was saying, the next nightly will be the 0.5 RC1
[11:36] 	bvdbos	:-)
[11:36] 	daniel	so hopefully no regressions...
[11:36] 	ctalbert	yeah, hopefully not.  
[11:37] 	daniel	you check in?
[11:37] 	ctalbert	Sure I can check it in after the QA Chat
[11:38] 	ctalbert	So, we should start testing the nightlies tomorrow, looking
 for any regressions, since this will essentially start the release process for 0.5.
[11:38] 	ssitter	(the next nightly might be the 0.5 RC1 if you switch all the 
tinderbox settings today and other stuff like disabling Firefox support, ...)
[11:38] 	ctalbert	yes, that's true.
[11:38] 	ctalbert	We would also need to start working on the localized builds.
[11:39] 	ctalbert	But, I think it makes sense to have a test day on Tuesday
 since 0.5 is essentially "code-complete" at this point.  This was the last blocking bug.
[11:40] 	ctalbert	The only problem with having a test day on Tuesday is that
 I will be travelling again.  That is the day I go out to california and start at
  MoCo so I don't know how availble I will be on that day.
[11:41] 	ctalbert	But, I think we should have one anyway, and start getting
 the "release candidate candidate" in front of people.  
[11:41] 	bvdbos	not to be a pain in the ass, but doesn't bug 377761 also block?
[11:41] 	firebot	bvdbos: Bug
 maj, --, ---,, NEW, Outlook 200x does not recognize iTIP/iMIP 
 invitation because of MIME type issues
[11:43] 	ctalbert	No, it doesn't.  With the UID issue fixed, when the user 
double clicks on the "attached" ICS in Outlook, Outlook opens it as a calendar event
 and allows the user to add it to the calendar.
[11:44] 	bvdbos	Then it's a WFM? I hope this is also solved now but not sure. 
We'll have to leave a comment then for the reporter
[11:44] 	Andreas	I think the last one is bug 372287, the Mac OS 10.3 bug
[11:44] 	firebot	Andreas: Bug
 cri, --, ---,, NEW, Sunbird crashes on open in Mac OS 10.3, works
  in Mac OS 10.4
[11:45] 	ctalbert	It is something we want to fix, but not something we need to
 fix for 0.5.  While Outlook doesn't recognize our ICS a a calendar attachment innately,
  it does add it to the calendar properly once the UID issue is fixed. 
[11:46] 	bvdbos	ok... Andreas, sipaq wrote abouw this on the weblog, though it's 
sincere, it's not blocking...
[11:46] 	ssitter	talkback finds four reports of this Mac OS X crash for Sunbird,
 for Thunderbird I wasn't able to find anything
[11:46] 	Andreas	ok
[11:46] 	ctalbert	ssitter: That's interesting.  The code ought to break on 
either app.  Maybe no one with 10.3 runs Thunderbird?? 
[11:46] 	daniel	lilmatt already has a system set up for that, will investigate
[11:48] 	ctalbert	It seems unlikely that no one with 10.3 is running lightning
[11:48] 	bvdbos	Did someone try the idea of mvl? It seems unlikely to me this 
will fix it, but you never know...
[11:50] 	ctalbert	I don't think so.  I think lilmatt has been concentrating
 on setting up the system.  He was able to successfully reproduce the crash on Tuesday.
[11:51] 	ctalbert	I looked at mvl's suggestion, I don't see that it would 
necessarily change anything, but without a 10.3 box, I can't really try it out.
[11:52] 	ctalbert	Speaking of the F2F meeting, are there any questions?  (I
 will be uploading notes for the last day onto the wiki page after this meeting)
[11:52] 	ctalbert	brb
[11:52] 	daniel	when will be the next one? :-)
[11:53] 	ctalbert	heh, I don't know.
[11:53] 	ssitter	today + x
[11:54] 	ctalbert	I'm not getting back on a plane today! :-)
[11:54] 	daniel	so maybe next week... ;-)
[11:55] 	daniel	you are going by truck to california? or by plane?
[11:55] 	ctalbert	plane
[11:55] 	daniel	and moving your stuff will do your wife then? ;-)
[11:56] 	ctalbert	I'll be back in Austin on May 10 for a couple of months
[11:56] 	ctalbert	Then we'll drive to CA in early July
[11:56] 	ctalbert	But, anyhow....we have 14 bugs on QA Discussion list.
[11:57] 	ctalbert	the first one is bug 247486
[11:57] 	firebot	ctalbert: Bug
 maj, --, ---,, NEW, can't load several calendars with different 
 passwords on same server
[11:57] 	bvdbos	this one has gotten nasty
[11:58] 	bvdbos	see my comment 12
[11:59] 	ctalbert	that's what I'm currently reading
[12:02] 	ctalbert	It looks to me like something odd is going on between these
 URLs and the password manager
[12:02] 	daniel	at least wcap has no problem since it is doing the password 
manager lookup/saving by itself
[12:03] 	daniel	needs to do so, because the password is sent encoded into url
[12:03] 	ctalbert	Is it really a valid use-case that a server would have the
 same name for two different authentication realms.
[12:04] 	bvdbos	yes it is like Pete claims... there are several providers which
 serve this. This morning I found out that icalx has a workaround though
[12:04] 	ctalbert	What's the workaround for them? 
[12:04] 	bvdbos	icalx serves the username in the realm
[12:05] 	bvdbos	so it gets stored as (iCalX Username calname.ics) || username || password
[12:06] 	bvdbos	but also the problem exists for simple webdav, you can't subsribe 
with user@domain/cal.ics
[12:06] 	ctalbert	Ok.  I think I understand.  Essentially the bug is that we
 need to handle this sort of access in our authorization.
[12:07] 	bvdbos	yep, relnote would be enough for now though... user:pass@domain/cal.ics
 works as does does domain/cal.ics as long as anonymous access is blocked by the server
[12:08] 	ctalbert	It sounds like a valid bug to me.  It also sounds like we
 have a good test case for it too. 
[12:08] 	ctalbert	I'll add the relnote settting to the whiteboard then.  Is
 there anything else we should change on this bug?
[12:09] 	bvdbos	If you agree my comment is clear enough, just ommit qa-discussion
[12:09] 	ctalbert	I think it's fine.
[12:09] 	ctalbert	thx for the research on that one.
[12:09] 	ctalbert	Next up is bug 266102
[12:09] 	firebot	ctalbert: Bug
 tri, --, ---,, NEW, Cancelling a calendar add opens calendar window
[12:12] 	firebot set the Resolution field on bug 266102 to
[12:12] 	mschroeder	agreed ;)
[12:12] 	ctalbert	That one was pretty easy.  Next is bug 291717
[12:12] 	ctalbert	heh
[12:12] 	firebot	ctalbert: Bug
 nor, --, ---,, NEW, CalendarEvent.ics - UNICODE [email event 
 needs to use UTF-8]
[12:13] 	bvdbos	I think we can use the writing code elsewhere (importing non-utf8)
 but this bug is invalid I think
[12:14] 	ctalbert	yes, that code is no longer in the product, and I believe we
 use UTF-8 just fine now when working with invitations.
[12:15] 	ssitter	has anyone tested invitations with some non-ascii content, e.g. 
Chinese characters?
[12:17] 	ctalbert	I did some time ago, and it worked.  Although, I have one bug
 open about non-ANSI characters coming in from Outlook.  Essentially if you use 
 Outlook to send an invitation with accents on letters (i.e. German or French or 
 whatnot) then those get munged when we read the invitation.  We should add UTF-8
  testing to our testcases.
[12:18] 	bvdbos	There's two or three bugs for importing and reading non-utf8
[12:19] 	bvdbos	bug 354951 was the last afaik
[12:19] 	firebot	bvdbos: Bug
 nor, --, ---,, NEW, ICS import fails in ip_parseFromStream [in 
[12:21] 	daniel	hmm, some time ago I have also submitted a patch because Google
 has produced invalid UTf-8 at that time.
[12:21] 	daniel	I suggested relaxes decoding, replacing characters that are invalid
[12:22] 	daniel	but mvl complained that this might cause dataloss during roundtripping,
 which is right.
[12:22] 	ctalbert	interesting.  So, we're not completely out of the UTF-8 woods 
[12:22] 	bvdbos	is you google for importExport.js the first hit is a forum-post 
where we need this for synching, perhaps this should be readded?
[12:23] 	daniel	I think we may get it right doing relaxed reading (replacing the 
invalids), but making those calendars read-only then.
[12:23] 	daniel	if we actually figure out an encoding error.
[12:24] 	ctalbert	We still (or again) have problems with UTF-8 invitaitons too, 
it seems.  
[12:24] 	ctalbert	Just tried it.
[12:24] 	*	JasnaPaka ( has joined #calendar-qa
[12:24] 	ctalbert	Part of it comes through, part doesn't
[12:24] 	ctalbert	I'll update the information in the bug.
[12:26] 	*	daniel has quit (Quit: have appointment with my wife)
[12:28] 	ctalbert	Ok.  updated 291717 with the current behavior.
[12:29] 	ctalbert	bug 330463 is next
[12:29] 	firebot	ctalbert: Bug 
nor, --, ---,, UNCO, Cannot create events in Lightning, regardless
 of which view is displayed [works with new profile]
[12:31] 	bvdbos	I think it's WFM, haven't heard from them and ssitter pinpointed 
the first problem
[12:31] 	ctalbert	I agree
[12:32] 	firebot set the Resolution field on bug 330463 to 
[12:33] 	ctalbert	Ok.  I think we're about out of time.  Is there anything 
else folks would like to bring up?
[12:34] 	bvdbos	How is press-coverage of sunbird 0.5 done? 
[12:35] 	ctalbert	I'm going to do some work with QAWanted tomorrow morning
 (my time -- afternoon for all ya'll)
[12:35] 	*	Andreas has quit (Ping timeout)
[12:36] 	bvdbos	is there one person who makes a press-statement and the 
localisers gather emailadresses and translate the press-release?
[12:36] 	ctalbert	It's done by a sort of ad-hoc process.  Simon and I usually
 collaborate on it a bit.  I was going to also ask my reporter friend if she would 
 write something on us
[12:36] 	ctalbert	No, it's not that organized
[12:36] 	ctalbert	But, it would be nice if it were.... ;-)
[12:37] 	bvdbos	I'll file a bug:-)
[12:37] 	ctalbert	ha!  Actually, it might be better to file an email.
[12:37] 	ctalbert	Or well....maybe a bug.  
[12:38] 	*	ctalbert checks to see if he has an up to date email address for Sipaq
[12:38] 	bvdbos	perhaps we can open an emailadress where all listst of emailadresses
 for different countries should be send, or just mail to sipaq?
[12:40] 	ctalbert	The more I think about it, the more I like the idea of 
opening a bug.  Then we can work on the press-release collaboratively without emailing 
the silly thing back and forth (we'll just attach it instead).  And we can CC all the 
localizers on the bug.  Can you file it against our Website component?  
[12:40] 	bvdbos	will do... 
[12:40] 	ctalbert	thanks
[12:41] 	bvdbos	see you tomorrow.... people coming to buy my house, have to 
clean up....
[12:41] 	ctalbert	ok.  I think that's the end of the QA Chat.  Thanks for your