Calendar:QA Chat:2007-07-12:Log

Meeting time?
[11:33am] Andreas joined the chat room.
[11:34am] ctalbert: yes indeed
[11:35am] ctalbert: ** QA Chat meeting **
[11:36am] ctalbert: We had a decent small test day on Tuesday.  We found that 
the some of the reason for the ICS/WebDav issues seem to be related to the icsParser 
check-in back in January. This is true for the HTTPS ICS/Webdav issues, and might be true for the other issues as well.
[11:37am] ctalbert: It also appears that the reporters of bug 373439 (caldav) have 
discovered that it is proxy handling that is breaking caldav support in LIghtning
[11:37am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=373439
 maj, --, ---, nobody@mozilla.org, UNCO, Contents caldav calendars disappear after 
 restarting Thunderbird. Sunbird works fine!
[11:37am] ctalbert: So, I think progress has been made.
[11:39am] ctalbert: I think we're going to declare Alan Schwartz as the winner of
 the test day since he did quite a bit to help us get on the right track with figuring
  out the webdav issue.  Also, many thanks to ssitter who figured out the reproduciblity
   and  regression range for the https bug.
[11:40am] ctalbert: I think that is all my news.
[11:41am] ssitter: and many thanks to Fallen|away for providing the test server
[11:41am] ctalbert: mschroeder: Last week you started working on the calendar QA ToDo
 list.  What is the status of the changes you wanted to make to it? 
[11:41am] ctalbert: ssitter: +1
[11:42am] ssitter: now we just need to pick one bug for this issue and resolve all other
 as duplicate ...
[11:42am] ctalbert: ssitter: that's also true.
[11:43am] mschroeder: ctalbert: I wanted to add specific components to test and maybe 
collecting ideas/bugs for automated testing
[11:44am] ctalbert: mschroeder: ok, sounds good.  Do you mind if I add stuff to it as 
well while you're working on it?
[11:46am] ctalbert: mschroeder: Also, in the news, this testing framework just landed:
 http://developer.mozilla.org/en/docs/Browser_chrome_tests
[11:46am] ctalbert: I'm going to be using it for some firefox testing, and I should be 
able to determine if this is a viable option for calendar as well once I get familiar with it.
[11:47am] ctalbert: And there was a nifty discussion yesterday of the Mozilla Project 
on http://air.mozilla.com, if you missed it they are going to post a recording of it.
[11:47am] mschroeder: ctalbert: You can add anything you want.
[11:48am] ctalbert: cool.
[11:49am] mschroeder: ctalbert: I'll have a look at the new framework.
[11:50am] ctalbert: mschroeder: awesome.  I should start looking at it late next week
 w.r.t.. firefox specific tests, so it'd be great if you could look at it sooner for
  use in calendar.  Let me know if I can help with anything w/ it.
[11:50am] ctalbert: Before we do QA Discussion bugs, let's see if we can dupe some of 
these caldav/webdav bugs.
[11:50am] ctalbert: ssitter: do you have a good list of them all?
[11:50am] • ctalbert tries to make one, even though he knows Stefan is faster
[11:51am] ssitter: no, not yet
[11:51am] ctalbert: ssitter: there is the list on http://wiki.mozilla.org/Calendar:Status
[11:55am] ctalbert: 386345 is WFM
[11:56am] ctalbert: bug 386326, comment 0 states: with proxy --> problem, without 
proxy-->no problem. (note this is a proxy problem with ICS, so it affects both ICS and caldav)
[11:56am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=386326 maj,
 --, ---, nobody@mozilla.org, UNCO, Remote calendars no more load
[11:57am] ctalbert: bug 386316 is also behind a proxy
[11:57am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=386316 nor,
 --, ---, nobody@mozilla.org, UNCO, Events not shown in month view
[11:59am] ctalbert: In several of these, I keep seeing that the old build (usually 0.3) 
worked ok, and then when they upgraded it broke.  In the caldav bug, they mention that they
 had issues on 0.3.1 and 0.5, but that on a 0.4pre build it worked.  I think we need to also
  figure out when the proxy case broke down.
[12:00pm] ssitter: was the events-not-shown issue traced back to proxy issues? caldav only 
or all providers?
[12:01pm] ctalbert: The disappearing events on caldav was traced back to proxy issues.
[12:02pm] ctalbert: ^^ that is bug 373439
[12:02pm] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=373439 maj, --,
 ---, nobody@mozilla.org, UNCO, Contents caldav calendars disappear after restarting Thunderbird.
  Sunbird works fine!
[12:03pm] ctalbert: Since I've got a bunch of these open now in my browser, why don't I tag them
 on the whiteboard with something, then we can more easily search for them.
[12:03pm] zwierz joined the chat room.
[12:03pm] ctalbert: [0.5 disappearing events]
[12:03pm] zwierz is now known as hubert|away.
[12:03pm] ctalbert: ^^ how does that sound?
[12:04pm] ssitter: sounds good
[12:05pm] ssitter: something like [0.5 publish fails] for the others?
[12:06pm] ctalbert: for which others?
[12:07pm] ctalbert: I was thinking to tag all of them that might be related with the same tag
 so that we can go through and figure out which are dupes.
[12:07pm] ssitter: the ones about failing to write .ics when using Limit/LimitExcept server 
rule
[12:07pm] ctalbert: ah, ok, yes that makes sense
[12:09pm] ctalbert: ssitter or mschroeder: Do one of you know how to answer Damian's questin 
in comment 5 on bug 292173?  That's our only QA Discussion bug atm
[12:10pm] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=292173 cri, --, 
---, nobody@mozilla.org, VERI FIXED, no error message when .ics is corrupted
[12:10pm] ctalbert: I thought there would be a popup that says the ICS could not be imported...
[12:13pm] mschroeder: If he doesn't get a error message maybe the bug isn't fixed.
[12:13pm] ctalbert: possible...
[12:15pm] ssitter: probably something changed in the 18 months between comment #4 and comment #5
[12:15pm] ctalbert: doh! I didn't notice that.
[12:16pm] ssitter: I'd rather suggest to open a new bug instead of reopening bugs closed long 
long ago
[12:17pm] mschroeder: agreed
[12:17pm] ssitter: I just tested it using other invalid data and get the error popup
[12:19pm] ssitter: maybe we just report corrupt ics content but ignore weird stuff the never 
could be an ics file 
[12:21pm] mschroeder: Has someone a link to mxr/lxr with the underlying code?
[12:21pm] ctalbert: maybe so
[12:22pm] ssitter: look for icsprovider, icsparser and icsimporter
[12:23pm] ctalbert: ssitter, mschroeder: Here is the list of bugs that I have tagged: http://tinyurl.com/2osx3y
[12:24pm] ctalbert: that should have both the [0.5 publish fails] and [0.5 disappear events] 
tags on that query
[12:25pm] ctalbert: ssitter: mschroeder here is a better query: http://tinyurl.com/2pvfmr
[12:25pm] ctalbert: I don't think everything is tagged yet, so please add more bugs to the tags
[12:26pm] mschroeder: okay
[12:29pm] ctalbert: I added a couple more to it.
[12:29pm] ctalbert: total of 10 now
[12:30pm] ctalbert: I think we're about out of time today (sorry I took so long to make the tag). 
 Does anyone have anything else to discuss?
[12:30pm] mschroeder: no.
[12:31pm] ssitter: ctalbert: after the write failure all events will disappear too. I think that's 
different from the use a proxy see nothing issue
[12:32pm] ctalbert: You're right.  I'll bet we're missing a try/catch somewhere and when the events 
fail to write, javascript throws and the views stop their update process.  Therefore, all events are gone.
[12:32pm] ctalbert: So, to summarize: we have three issues:
[12:32pm] ctalbert: 1. Fail to write, causing all events to disappear from views (including those
 of local calendar)
[12:32pm] ctalbert: 2. Proxy issues
[12:33pm] ctalbert: 3. icalParser issue that disrupts networking ability
[12:33pm] ctalbert: ssitter: Does that look correct?
[12:33pm] • ctalbert wants to blog about this
[12:34pm] ssitter: the 3rd issue is not clear to me
[12:34pm] ctalbert: The third issue is what you found on the testday, the problem with HTTPs 
that seems to be related to the icsParser check-in.  (I didn't state it very well)
[12:35pm] ssitter: that's your number 1
[12:35pm] ctalbert: You think 1 and 3 are the same?
[12:36pm] ssitter: the HTTPs was tracked back to the server config used by Fallen
[12:36pm] ctalbert: So, it has nothing to do with the icalParser code?
[12:36pm] ssitter: after changing the server config it worked
[12:37pm] ctalbert: oh.  I missed that.
[12:37pm] ssitter: it must have something to do with the icsparser
[12:37pm] ctalbert: right.
[12:37pm] ssitter: icsparser -> icsparser/icsserializer
[12:38pm] ssitter: that's what I want to know from mvl in bug 387559 comment 2
[12:38pm] ctalbert: So, in summary we only have two issues: the write issue and the proxy issue.
[12:38pm] firebot: ssitter: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=387559 nor, --,
 ---, nobody@mozilla.org, NEW, Creating event on webdav server using https url fails
[12:39pm] ssitter: I should update the summary on my bug or resolve it against one of the others
[12:39pm] ctalbert: yes, so do I.  That Alan Schwartz guy said that he wasn't using https iirc,
 but I don't know what his limit/limitexcept settings are.  We should ask him.
[12:39pm] ssitter: we have number one:
[12:39pm] ctalbert: ssitter: I would tend to resolve the other bugs to yours, since yours are
 usually clearer to follow 
[12:39pm] ctalbert: s/usually/always
[12:40pm] ssitter: server config allows anonymous read nut requires auth for write
[12:40pm] ssitter: write fails, events are gone, process hangs
[12:41pm] ssitter: ctalbert: ok, what would be a better bug summary?
[12:42pm] • ssitter thinks he should steal from bug 380746
[12:42pm] firebot: ssitter: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=380746 nor, --,
 ---, nobody@mozilla.org, UNCO, webdav write failure if HTTP AUTH needed for PUT but not GET
[12:43pm] ctalbert: I agree, that sums it up.
[12:46pm] ssitter: ctalbert: Event creation/change fails if server requires authentication for
 write (PUT) but not for read (GET)
[12:46pm] ctalbert: That's a much more user friendly summary.  +1!
[12:47pm] ssitter: that's a special summary for mvl 
[12:47pm] ctalbert: heh
[12:49pm] ctalbert: Alright, I think I'm going to call the end of the meeting here.  I will 
post the summary info and the query from above (with all the related bugs) to the blog tonight
 in the form of our "test day results".  Ping me if there's anything you want me to include 
 in the blog that you think of later.