Calendar:QA Chat Log:2006-10-12

From MozillaWiki
Jump to: navigation, search
<ctalbert>	Alright, let's get started with the Calendar QA Chat.
<ctalbert>	The agenda is pretty full:
<ctalbert>	Anyone interested in putting the meeting log up on that page 
<ctalbert>	ok I'll do it.
<ctalbert>	First, Congrats to everyone. We have shipped 0.3!
<jminta>	lilmatt reported that we had > 10000 downloads in the first 24hrs
<ctalbert>	Wow
<Sebo>	Where has it been anounced?
<Sebo>	apart from Newsgroup, I mean.
<ctalbert>	Our mozillaZine blog, the calendar newsgroup...
<ctalbert>	I'm not sure where else. I think there was some press surrounding
 it that Cedric put together in France
<Sebo>	Thats a pretty amazing number, then!
<ctalbert>	Yes, it is.
<Sebo>	But there was nothing at MozillaZine news
<ctalbert>	Their "submit news" functionality is broken. I attempted to 
contact the webmaster but still haven't heard anything from him/her.
<ssitter>	I saw articles on some of the bigger IT news sizes in Germany
<ssitter>	sites
<ssitter>	might be interesting to collect all the 0.3 related articles in
<ssitter>	and use for feedback later
<ctalbert>	It would. Then we'd know who to ping when we need publicity for 
events and what not.
<ctalbert>	If you can send me the links, I can put them on our wiki 
<ssitter>	create ?
<ctalbert>	Sounds like a good place.
<ctalbert>	I'll search around the American media, see if I can find any mention of us.
<ctalbert>	By the way, there will be a post-mortem discussion on the release
 during the next *Wednesday* Calendar Conference Call. You're welcome to attend 
that, if you want.
<ctalbert>	But, in case people can't make that, we'll be talking about the 
post-mortem ideas in our QA Chat next week.
<ctalbert>	So, I encourage people to make notes of things we can do better, 
things we did well etc.
<ctalbert>	Our next Test day will be this coming Tuesday, October 17.
<ctalbert>	I want this test day to be targetted at specific areas of functionality. What should those areas be?
<jminta>	what's the current imip/itip state? have we done much testing 
<ctalbert>	We have not.
<jminta>	some checks of invites from the major other apps might be useful
<ctalbert>	I think so.
<ctalbert>	We've seen several issues lately with importing different ICS 
files. Perhaps we can try import/export testing from other apps as well?
<jminta>	maybe frame this as a genearl inter-op testday?
<ctalbert>	Sounds good.
<ctalbert>	What about the test day hours? Did the midnight to midnight 
timeframe work out for people at the last one?
<ctalbert>	It felt very awkward to me. But, I don't know if that was because
 I was in California, and having to deal with an unfamiliar GMT offset, or if it 
was awkward in general.
<Sebo>	Well, for Europe the time is ok, I think.
<Andreas>	Yes, that's right.
<ctalbert>	Cool. Then I think we'll keep it.
<ctalbert>	I might extend the end time to 1AM or 2AM UTC. That would give 
people in California a little more time in their afternoon to help out.
<ctalbert>	Anyone want to update the Test Day wiki for a new day? celina63 
has made it pretty simple now. I think it's mostly just cut and paste and 
changing the dates.
<ctalbert>	It's very similiar to the way the wiki for these meetings is 
<Sebo>	I can play around with it.
<ctalbert>	thanks
<ctalbert>	I was looking at the calendar last night and saw that we will 
have a test day on Halloween Oct. 31st.
<ctalbert>	I was thinking about changing things for that one.
<ctalbert>	What do you think about trying a team based test day?
<ctalbert>	There would be two teams. We would set up the core members of the
 team in our meeting before the test day. And when other people show up in the 
channel, they'd be assigned to a team.
<ctalbert>	The teams would compete to see who can find the most bugs, run 
the most test cases, etc.
<ctalbert>	And since it was Halloween, I was thinking of making a monster 
based theme: For example: It could be team Frankenstein versus team Godzilla. Or 
Ghosts versus Goblins.
<ctalbert>	My goal is to do a little recruiting by doing something 
completely different and fun.
<ctalbert>	What do ya'll think? Am I nuts?
<ctalbert>	OK, I'll take the silence to mean: "Clint is out of his mind."
<ctalbert>	So, on to the next serious point.
<ctalbert>	We need a regression strategy.
<ctalbert>	This is something that I believe we should get into place before 
<ctalbert>	I put some ideas onto the agenda about that. I have jminta's QA 
extension code. Is anyone interested in working on this problem?
<Sebo>	I dont really understand what this would do
<ctalbert>	jminta: Do you want to explain?
<jminta>	my extension? sure
<jminta>	the idea is that a lot of bugs are easily identifiable by simple 
<jminta>	for example, at one point ics files with ORGANIZER properties 
were losing that property when we cloned the items
<firebot> set the Resolution field on bug 356383 to 
DUPLICATE of bug 215971.
<firebot>	Bug enh, --, 
---,, RESO DUPLICATE, Reload remote calendars should be 
automatically executed before changes are made
<jminta>	my extension is a framework for running tests like "Take this 
file with an ORGANIZER, clone it, and see if it still has an organizer"
<firebot>	Bug enh, --, 
---,, ASSI, auto refresh remote calendar every x minutes
<jminta>	firebot: shut up
<firebot>	jminta: I wasn't talking to you.
<jminta>	also, for cases where someone writes bad code and an exception is
<jminta>	those are easy tests to add into my framework
<jminta>	most bugs with "Event X doesn't appear in view" can also be done 
this way
<jminta>	it's just a matter of polishing up the framework and then writing
 the tests
<ctalbert>	Sebo, does that answer your question?
<jminta>	I have an example test at
<jminta>	(the test checks to make sure we store event titles correctly)
<Sebo>	I got it, thanks
<ctalbert>	I was thinking that we can use some of this code to enhance the 
existing XPC-based tests and create an automatic regression test as well.
<ctalbert>	This would be something that we could run after a build finishes 
and we would know immediately if the recent checkins have caused any regressions.
<lilmatt>	automatic == good
* jminta	notes that his extension sacrifices 'automatic' in exchange for 
being able to test the front-end
<ctalbert>	I realize that ^^^^ and I think that's why we need both
* jminta	agrees
<lilmatt>	We were talking to that guy at Calconnect about using something 
like Applescript to be able to test the front-end
<lilmatt>	Like with an Automator test suite or something
<Sebo>	So, in general: take all the litmus basic functional tests and rewrite 
them for the extension?
<ctalbert>	True, but I have no idea when or if that is going to happen. Do 
you have his contact info?
<ctalbert>	Sebo, not all. I think this extension can be used through litmus 
for people to run more complex steps. I don't intend to replace Litmus testcases 
with it.
<ctalbert>	For example, Litmus would say: 1. Install extension, install test
 48, use the calendar-qa GUI to run test 48.
<jminta>	Litmus is much better at dealing with what the user sees
<jminta>	while the extension deals better with internal states/variables
<ctalbert>	Exactly.
<ctalbert>	It gives us a way to allow folks testing with Litmus to run the 
kinds of tests those of us with more experience can run.
<ctalbert>	For example: Timezone tests.
<ctalbert>	Is anyone interested in working on this problem? I'm going to 
devote some of my time to it during the next few weeks, but I have several other 
committments to meet as well.
<ctalbert>	And so, I'll defnitely need help.
<jminta>	starting next weekend i'll have a bit of time i can devote
<lilmatt>	Nice!
<jminta>	(yay for fall break)
<ctalbert>	That's also when my schedule opens up a bit.
<ctalbert>	Great.
<lilmatt>	jminta: Are you staying in CT?
<jminta>	lilmatt: no, some time in South Bend and some time in Florida
<ctalbert>	We are almost out of time, but I'm willing to go over if we have 
to. Does anyone have anything to add, discuss? Or any suggestions for the 
calendar ToDo list?
<Sebo>	Will there be a meeting to decide on 0.5 items?
<ctalbert>	Good question. I should have covered that. We will be discussing 
that in next Wednesday's conference call.
<ctalbert>	I can also add that to the calendar-qa chat agenda for next week 
<Sebo>	ok
<ctalbert>	The conference call is open to everyone, so don't be shy about 
<lilmatt>	Over 1300 downloads of lightning from amo since launch
<ctalbert>	Wow, that's awesome.
<lilmatt>	...and it only got on amo around midday
<ctalbert>	:-)
<ctalbert>	I am going to remove the "upgrade" items from the calendar QA 
ToDo list, unless there are objections
<lilmatt>	Just be ready to add them back if/when we land jminta's migratro
<lilmatt>	migrator
<ctalbert>	Definitely.
<ctalbert>	In the interest of time, I don't want to go through a 
brainstorming session here about the ToDo list like we usually do.
<ctalbert>	So, I encourage everyone to edit it, adding things you believe 
need to be added, removing things if needed.
<ctalbert>	Thanks very much for all that you have done to get 0.3 out the door.
<ctalbert>	Does anyone have anything else?
<ctalbert>	To quickly wrap up:
<ctalbert>	ssitter and I will put together press links from 0.3 onto the 
<ctalbert>	sebo will play with switching the test day wiki to reflect the 
new day
<ctalbert>	jminta and I will work on the regression test strategy after next
<ctalbert>	we have a test day on the 17th and we'll need to mobilise for 
<lilmatt>	We should announce that too, and us devs should try and get some 
fixes in for the stuff we are seeing in bug reports before then
<lilmatt>	I've documented some of what I'm seeing at
<ctalbert>	Indeed.
<ctalbert>	lilmatt: did all the items you listed in yesterday's meeting get 
onto this page?
<ctalbert>	oh, there it is at the bottom
<lilmatt>	ya.
<lilmatt>	This was a brain dump.
<lilmatt>	It can get organized later
<ctalbert>	cool
<ctalbert>	okay. I think this concludes the meeting. Thanks, everyone.
<ctalbert>	Sorry for going over time.