Calendar:QA Chat:2006-05-03:Log

From MozillaWiki
Jump to: navigation, search
mschroeder: welcome back clint 
[09:36am] ctalbert: thans
[09:36am] ctalbert: k
[09:37am] ctalbert: I think it is time to start the QA chat
[09:38am] ctalbert: This morning, Andreas, mschroeder , bvdbos and I went through
several of the QA discussion needed bugs.  The only ones of those that are left are
 just in a pending state depending on testing that needs to be done.
[09:38am] ulf|away joined the chat room.
[09:39am] ulf|away is now known as ulf.
[09:39am] ctalbert: We had a small test day on Tuesday.  I don't know exactly how 
many people participated since I was not involved, but it looks like we ran 62 
tests and filed 5 bugs
[09:39am] ctalbert: Omar won the test day
[09:40am] ulf: great! congrats!
[09:40am] ctalbert: We are about to release a pre-release 0.5.  I don't think it 
is the full RC just yet though.
[09:40am] ctalbert: As far as I know the 10.3 Mac issue still blocks
[09:41am] ulf: and Andreas found another critical one
[09:41am] ctalbert: Which one was that?
[09:42am] • ctalbert hopes to finally catch up on email today
[09:42am] mschroeder: bug 379462
[09:42am] Andreas: bug 379462
[09:42am] firebot: mschroeder: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=379462 
cri, --, Lightning 0.5, michael.buettner@sun.com, ASSI, Changing timezone of event in 
various timezones isn't possible
[09:42am] ssitter: I think we should enable branding for the pre-RC, this way we can 
get test coverage on this topic now
[09:42am] ulf: thx.
[09:42am] ctalbert: ssitter: test coverage on branding, or test coverage on the issue 
in 379462?
[09:43am] ssitter: branding topics
[09:43am] ctalbert: I agree.  What do we have to do change that?  iirc, it is a setting
 in the mozconfig on the tinderbox, right?
[09:44am] ssitter: yes, I think you have to change some configs in cvs
[09:44am] • ssitter points to lilmatt
[09:45am] ctalbert: I'll take it up with him.  (I think he's out here on the west coast 
somewhere and may not be awake yet)
[09:46am] ctalbert: Nice find, Andreas, by the way.
[09:46am] Andreas: thanks
[09:46am] ctalbert: ssitter, are we waiting for the full RC to do all the rest of the
 things on your list?
[09:46am] ctalbert: (l10n, turn off fx support, etc)
[09:47am] ssitter: as I understand the meeting notes from yesterday it will be as small
 as possible, so probably no l10n
[09:47am] ctalbert: ok.
[09:48am] mschroeder: Why isn't fx support turned off by default?
[09:48am] ssitter: afaik for development testing in Firefox
[09:48am] ctalbert: A very good question.  I think it's because jminta was sort of working
 on an experimental version of Lightning for Firefox
[09:49am] ctalbert: I think there is also a shortcut to loading lightning in firefox so you
 don't have to load it in Thunderbird, but I've never used it.
[09:49am] ssitter: for the last releases it was temporary disabled during release building
[09:49am] ctalbert: I think we should continue disabling the firefox support for release
[09:50am] ctalbert: So, should we move on to have a look at the QA Wanted list?
[09:51am] ssitter: just a note on 379462 again, I think we need a decision if this 
is a blocker asap
[09:51am] ctalbert: The list, sorted by severity: http://tinyurl.com/yrfbzk
[09:52am] ctalbert: Bug 367822 is first
[09:52am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=367822 
cri, --, ---, nobody@mozilla.org, NEW, Sunbird 0.3 may corrupt remote calendar if 
closed while popup alarm box is active
[09:52am] ssitter: for me it looks like the same issue is in 0.3.1, at least in 
Sunbird, maybe even 0.3
[09:54am] ctalbert: I would believe it.  Has anyone tried this since bug 355755 was 
fixed?
[09:54am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=355755
 nor, --, Sunbird 0.5, daniel.boelzle@sun.com, RESO FIXED, Alarms are dismissed when
  Sunbird is closed
[09:55am] ctalbert: sounds like we should try it with a current nightly and if it still
 happens, note it and clear the QAWanted status.
[09:56am] ssitter: sebo I think, see comment 4
[09:57am] ctalbert: heh, I didn't see the date on that comment
[09:57am] ssitter: now the problem don't appear 'automatically', only if you force
 it 'manually'
[09:58am] ctalbert: ok.  Then I'll decrease the severity on it to major, and remove 
QAWanted sound good?
[09:58am] ssitter: according to Bug 328810 Sunbird should wait for write to finish
[09:58am] firebot: ssitter: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=328810
 maj, --, ---, jminta@gmail.com, RESO FIXED, shutting down calendar doesn't wait for
  in-progress writes to finish
[09:59am] ssitter: but in this case the write starts after the shutdown was started
 if I understand correctly
[10:00am] ssitter: so the mechanism above doesn't seem to intercept this
[10:00am] ssitter: ctalbert: yeah, I think so too
[10:01am] ctalbert: cool.  I added your comments there as well.
[10:01am] ctalbert: bug 372287 is next...oh wait.  we know that one
[10:01am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=372287
 cri, --, ---, lilmatt@flock.com, NEW, Sunbird crashes on open in Mac OS 10.3, works 
 in Mac OS 10.4
[10:02am] ctalbert: I'll dig around up here in the QA lab and see if there are any 
10.3 boxes
[10:02am] ctalbert: next up is bug 320362
[10:02am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=320362 
maj, --, ---, nobody@mozilla.org, UNCO, Unable to subscribe to remote calendar
[10:03am] mschroeder: daniel attached a patch to the MacOSX 10.3 bug today
[10:03am] • ctalbert looks
[10:04am] ctalbert: Hmm...that does look promising based on how lilmatt described the
 problem
[10:05am] ctalbert: On to bug 320362
[10:06am] ctalbert: mschroeder:  This is another one of those "sharing profile between
 linux and windows" issues
[10:06am] ctalbert: that we discussed earlier
[10:06am] ctalbert: That means that 320362, 357493, 354674 may all be somewhat related.
[10:07am] ctalbert: My idea is that one of us with a dual boot system should figure 
out if/how to share the profile between windows and linux and then publish that as 
the standard way to do this.  That might help resolve some of these bugs.
[10:08am] ctalbert: I don't think this is a common configuration for most people,
 but we do seem to have several folks out there that are trying it.
[10:10am] ctalbert: I think this should be easily solved if people can share a set 
of ICS files between the two OS's, rather than attempting to share profiles between 
the OS's.
[10:10am] ctalbert: If we have a published mechanism for doing this (sharing ICS's 
etc) then we can WONTFIX the rest of these. (Or make them post-1.0 bugs)
[10:12am] ctalbert: In this bug, this guy is essentially trying to create a local calendar over the network.
[10:12am] ssitter: is this a calendar bug or a mozStorage bug?
[10:13am] ctalbert: good question I'm not sure
[10:14am] ctalbert: It was orignally reported when we were just using ICS as the backend store
[10:15am] ctalbert: Then in 2006 it the guy trying to create a local calendar over a NFS mount comes into the picture
[10:15am] ctalbert: What do ya'll think we should do with this?
[10:16am] ssitter: comment #0 mentions storage.sdb, so it's already using mozstorage
[10:16am] ctalbert: oh duh, you're right
[10:16am] ssitter: I'd say mark it for retest as you suggested
[10:16am] ctalbert: ok.  I'll do that
[10:16am] mschroeder: +1
[10:17am] ssitter: if it can be confirmed decide where the bug belongs to
[10:17am] ctalbert: right.  Ok.  bug 356269 is next, see what ya'll think I'm going
 to quickly update 320362
[10:17am] ssitter: might be worth to also test with firefox that also used mozstorage
 but for different files
[10:18am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=356269
 maj, --, ---, nobody@mozilla.org, NEW, localization will not work Lightning are 
 always english
[10:19am] ssitter: this is wfm for me but reproduced by others -> retest with 0.5 
l10n Lightning as written in comment 17
[10:22am] ctalbert: ah yes, indeed
[10:22am] ctalbert: bug 359796 is next
[10:22am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=359796
 maj, --, ---, nobody@mozilla.org, NEW, Alarms fire too early, ignoring timezone settings
[10:25am] firebot: New Calendar - Lightning Only bug 379622 filed by lists@fairsky.us.
[10:25am] firebot: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=379622 nor,
 --, ---, nobody@mozilla.org, UNCO, in calendar view, impossible to get back to 
 email with keyboard (tab)
[10:26am] ctalbert: As far as I can tell, 359796 doesn't look like it needs more
 QA.  It seems to be an issue calculating the alarm time when you have recurring 
 all day events.  Which is a point where RFC 2445 is a little vague.
[10:26am] ctalbert: I think we should remove QAWanted.  I think there are more
 than enough examples here.
[10:27am] mschroeder: Archaeopteryx comments at the end sound like the alarm 
timer/hibernation problem
[10:28am] ctalbert: Which one?
[10:28am] mschroeder: that's alarms appearing later, not earlier (comment#26)
[10:29am] ctalbert: true
[10:29am] mschroeder: Where did he get "TZID=/Mozilla.org/BasicTimezones/GMT" 
in his ics files from?
[10:30am] ctalbert: I dunno
[10:30am] ctalbert: He's using 0.5pre he says
[10:31am] ctalbert: Archaeopteryx: ping
[10:31am] ctalbert: I'll talk to him about this
[10:31am] mschroeder: okay
[10:31am] ctalbert: and we'll leave the bug as it stands for now.
[10:32am] ctalbert: Looks like we're about out of time.
[10:32am] ctalbert: Is there anything else folks want to bring up?
[10:33am] mschroeder: I already installed stable webcalendar 1.0.5 for two of 
the qa discussion needed bugs and will install the development version 1.1.2 
(which has full timezone support)

[10:33am] ctalbert: Awesome.  And, so the version these folks were probably 
using did *not* have full timezone support.
[10:33am] ctalbert: Thanks, mschroeder
[10:33am] mschroeder: only offset
[10:34am] ctalbert: yikes
[10:34am] mschroeder: you can set an offset or take it from the server
[10:34am] ctalbert: I see
[10:35am] ctalbert: Ok, well this ends the QA Chat.  Keep up the good work.  Need
 to make sure 0.5 is a good release.  I'll probably more and more available as I
 start to get a handle on my new job here.
[10:36am] ctalbert: Thanks a lot for all ya'll do.mschroeder: welcome back clint 
[09:36am] ctalbert: thans
[09:36am] ctalbert: k
[09:37am] ctalbert: I think it is time to start the QA chat
[09:38am] ctalbert: This morning, Andreas, mschroeder , bvdbos and I went through
several of the QA discussion needed bugs.  The only ones of those that are left are
 just in a pending state depending on testing that needs to be done.
[09:38am] ulf|away joined the chat room.
[09:39am] ulf|away is now known as ulf.
[09:39am] ctalbert: We had a small test day on Tuesday.  I don't know exactly how 
many people participated since I was not involved, but it looks like we ran 62 
tests and filed 5 bugs
[09:39am] ctalbert: Omar won the test day
[09:40am] ulf: great! congrats!
[09:40am] ctalbert: We are about to release a pre-release 0.5.  I don't think it 
is the full RC just yet though.
[09:40am] ctalbert: As far as I know the 10.3 Mac issue still blocks
[09:41am] ulf: and Andreas found another critical one
[09:41am] ctalbert: Which one was that?
[09:42am] • ctalbert hopes to finally catch up on email today
[09:42am] mschroeder: bug 379462
[09:42am] Andreas: bug 379462
[09:42am] firebot: mschroeder: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=379462 
cri, --, Lightning 0.5, michael.buettner@sun.com, ASSI, Changing timezone of event in 
various timezones isn't possible
[09:42am] ssitter: I think we should enable branding for the pre-RC, this way we can 
get test coverage on this topic now
[09:42am] ulf: thx.
[09:42am] ctalbert: ssitter: test coverage on branding, or test coverage on the issue 
in 379462?
[09:43am] ssitter: branding topics
[09:43am] ctalbert: I agree.  What do we have to do change that?  iirc, it is a setting
 in the mozconfig on the tinderbox, right?
[09:44am] ssitter: yes, I think you have to change some configs in cvs
[09:44am] • ssitter points to lilmatt
[09:45am] ctalbert: I'll take it up with him.  (I think he's out here on the west coast 
somewhere and may not be awake yet)
[09:46am] ctalbert: Nice find, Andreas, by the way.
[09:46am] Andreas: thanks
[09:46am] ctalbert: ssitter, are we waiting for the full RC to do all the rest of the
 things on your list?
[09:46am] ctalbert: (l10n, turn off fx support, etc)
[09:47am] ssitter: as I understand the meeting notes from yesterday it will be as small
 as possible, so probably no l10n
[09:47am] ctalbert: ok.
[09:48am] mschroeder: Why isn't fx support turned off by default?
[09:48am] ssitter: afaik for development testing in Firefox
[09:48am] ctalbert: A very good question.  I think it's because jminta was sort of working
 on an experimental version of Lightning for Firefox
[09:49am] ctalbert: I think there is also a shortcut to loading lightning in firefox so you
 don't have to load it in Thunderbird, but I've never used it.
[09:49am] ssitter: for the last releases it was temporary disabled during release building
[09:49am] ctalbert: I think we should continue disabling the firefox support for release
[09:50am] ctalbert: So, should we move on to have a look at the QA Wanted list?
[09:51am] ssitter: just a note on 379462 again, I think we need a decision if this 
is a blocker asap
[09:51am] ctalbert: The list, sorted by severity: http://tinyurl.com/yrfbzk
[09:52am] ctalbert: Bug 367822 is first
[09:52am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=367822 
cri, --, ---, nobody@mozilla.org, NEW, Sunbird 0.3 may corrupt remote calendar if 
closed while popup alarm box is active
[09:52am] ssitter: for me it looks like the same issue is in 0.3.1, at least in 
Sunbird, maybe even 0.3
[09:54am] ctalbert: I would believe it.  Has anyone tried this since bug 355755 was 
fixed?
[09:54am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=355755
 nor, --, Sunbird 0.5, daniel.boelzle@sun.com, RESO FIXED, Alarms are dismissed when
  Sunbird is closed
[09:55am] ctalbert: sounds like we should try it with a current nightly and if it still
 happens, note it and clear the QAWanted status.
[09:56am] ssitter: sebo I think, see comment 4
[09:57am] ctalbert: heh, I didn't see the date on that comment
[09:57am] ssitter: now the problem don't appear 'automatically', only if you force
 it 'manually'
[09:58am] ctalbert: ok.  Then I'll decrease the severity on it to major, and remove 
QAWanted sound good?
[09:58am] ssitter: according to Bug 328810 Sunbird should wait for write to finish
[09:58am] firebot: ssitter: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=328810
 maj, --, ---, jminta@gmail.com, RESO FIXED, shutting down calendar doesn't wait for
  in-progress writes to finish
[09:59am] ssitter: but in this case the write starts after the shutdown was started
 if I understand correctly
[10:00am] ssitter: so the mechanism above doesn't seem to intercept this
[10:00am] ssitter: ctalbert: yeah, I think so too
[10:01am] ctalbert: cool.  I added your comments there as well.
[10:01am] ctalbert: bug 372287 is next...oh wait.  we know that one
[10:01am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=372287
 cri, --, ---, lilmatt@flock.com, NEW, Sunbird crashes on open in Mac OS 10.3, works 
 in Mac OS 10.4
[10:02am] ctalbert: I'll dig around up here in the QA lab and see if there are any 
10.3 boxes
[10:02am] ctalbert: next up is bug 320362
[10:02am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=320362 
maj, --, ---, nobody@mozilla.org, UNCO, Unable to subscribe to remote calendar
[10:03am] mschroeder: daniel attached a patch to the MacOSX 10.3 bug today
[10:03am] • ctalbert looks
[10:04am] ctalbert: Hmm...that does look promising based on how lilmatt described the
 problem
[10:05am] ctalbert: On to bug 320362
[10:06am] ctalbert: mschroeder:  This is another one of those "sharing profile between
 linux and windows" issues
[10:06am] ctalbert: that we discussed earlier
[10:06am] ctalbert: That means that 320362, 357493, 354674 may all be somewhat related.
[10:07am] ctalbert: My idea is that one of us with a dual boot system should figure 
out if/how to share the profile between windows and linux and then publish that as 
the standard way to do this.  That might help resolve some of these bugs.
[10:08am] ctalbert: I don't think this is a common configuration for most people,
 but we do seem to have several folks out there that are trying it.
[10:10am] ctalbert: I think this should be easily solved if people can share a set 
of ICS files between the two OS's, rather than attempting to share profiles between 
the OS's.
[10:10am] ctalbert: If we have a published mechanism for doing this (sharing ICS's 
etc) then we can WONTFIX the rest of these. (Or make them post-1.0 bugs)
[10:12am] ctalbert: In this bug, this guy is essentially trying to create a local calendar over the network.
[10:12am] ssitter: is this a calendar bug or a mozStorage bug?
[10:13am] ctalbert: good question I'm not sure
[10:14am] ctalbert: It was orignally reported when we were just using ICS as the backend store
[10:15am] ctalbert: Then in 2006 it the guy trying to create a local calendar over a NFS mount comes into the picture
[10:15am] ctalbert: What do ya'll think we should do with this?
[10:16am] ssitter: comment #0 mentions storage.sdb, so it's already using mozstorage
[10:16am] ctalbert: oh duh, you're right
[10:16am] ssitter: I'd say mark it for retest as you suggested
[10:16am] ctalbert: ok.  I'll do that
[10:16am] mschroeder: +1
[10:17am] ssitter: if it can be confirmed decide where the bug belongs to
[10:17am] ctalbert: right.  Ok.  bug 356269 is next, see what ya'll think I'm going
 to quickly update 320362
[10:17am] ssitter: might be worth to also test with firefox that also used mozstorage
 but for different files
[10:18am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=356269
 maj, --, ---, nobody@mozilla.org, NEW, localization will not work Lightning are 
 always english
[10:19am] ssitter: this is wfm for me but reproduced by others -> retest with 0.5 
l10n Lightning as written in comment 17
[10:22am] ctalbert: ah yes, indeed
[10:22am] ctalbert: bug 359796 is next
[10:22am] firebot: ctalbert: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=359796
 maj, --, ---, nobody@mozilla.org, NEW, Alarms fire too early, ignoring timezone settings
[10:25am] firebot: New Calendar - Lightning Only bug 379622 filed by lists@fairsky.us.
[10:25am] firebot: Bug https://bugzilla.mozilla.org/show_bug.cgi?id=379622 nor,
 --, ---, nobody@mozilla.org, UNCO, in calendar view, impossible to get back to 
 email with keyboard (tab)
[10:26am] ctalbert: As far as I can tell, 359796 doesn't look like it needs more
 QA.  It seems to be an issue calculating the alarm time when you have recurring 
 all day events.  Which is a point where RFC 2445 is a little vague.
[10:26am] ctalbert: I think we should remove QAWanted.  I think there are more
 than enough examples here.
[10:27am] mschroeder: Archaeopteryx comments at the end sound like the alarm 
timer/hibernation problem
[10:28am] ctalbert: Which one?
[10:28am] mschroeder: that's alarms appearing later, not earlier (comment#26)
[10:29am] ctalbert: true
[10:29am] mschroeder: Where did he get "TZID=/Mozilla.org/BasicTimezones/GMT" 
in his ics files from?
[10:30am] ctalbert: I dunno
[10:30am] ctalbert: He's using 0.5pre he says
[10:31am] ctalbert: Archaeopteryx: ping
[10:31am] ctalbert: I'll talk to him about this
[10:31am] mschroeder: okay
[10:31am] ctalbert: and we'll leave the bug as it stands for now.
[10:32am] ctalbert: Looks like we're about out of time.
[10:32am] ctalbert: Is there anything else folks want to bring up?
[10:33am] mschroeder: I already installed stable webcalendar 1.0.5 for two of 
the qa discussion needed bugs and will install the development version 1.1.2 
(which has full timezone support)

[10:33am] ctalbert: Awesome.  And, so the version these folks were probably 
using did *not* have full timezone support.
[10:33am] ctalbert: Thanks, mschroeder
[10:33am] mschroeder: only offset
[10:34am] ctalbert: yikes
[10:34am] mschroeder: you can set an offset or take it from the server
[10:34am] ctalbert: I see
[10:35am] ctalbert: Ok, well this ends the QA Chat.  Keep up the good work.  Need
 to make sure 0.5 is a good release.  I'll probably more and more available as I
 start to get a handle on my new job here.
[10:36am] ctalbert: Thanks a lot for all ya'll do.