989785 - apps/calendar/test/marionette/week_view_test.js failing because of week view regression

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Gaia::Calendar, defect)

Description

[Phil Ringnalda (:philor)]

At least, my guess for why it's suddenly failing today and didn't fail at the end of January or the end of February is that both February 1st and March 1st fell on a Saturday.

https://tbpl.mozilla.org/php/getParsedLog.php?id=36965880&tree=B2g-Inbound

10:35:53     INFO -  TEST-START | week view multiple months (eg. "Dec 2013 Jan 2014")
10:41:23     INFO - Automation Error: mozprocess timed out after 330 seconds running ['make', 'test-integration', 'NPM_REGISTRY=http://npm-mirror.pub.build.mozilla.org', 'REPORTER=mocha-tbpl-reporter', 'TEST_MANIFEST=./shared/test/integration/tbpl-manifest.json']

do {} while () seems like a rather bustage-prone pattern to use in a test, but it's particularly unfortunate in this suite, where bug 953212 means that hangs are not caught in any vaguely useful way.

Comment 1

[Gregor Wagner [:gwagner]]

Attachment: disable test

Comment 2

[Gregor Wagner [:gwagner]]

We should fix this :)

Comment 3

[Gregor Wagner [:gwagner]]

https://github.com/mozilla-b2g/gaia/commit/659562eb1ebd2ec1c09027e45fef9841c06b51f3

Comment 4

[Phil Ringnalda (:philor)]

Curiously enough, it's not failing on mozilla-aurora.

Comment 5

[Miller Medeiros [:millermedeiros]]

I don't think the problem is because of the start of the month.. I think the bug was probably introduced last week when I landed Bug 988079 - which was a patch for the Day View, but seems to affect the Week View in a different way. - and the test was only able to catch the regression since now it required more than a few swipes.

The test implementation is correct. It changes the date until it finds a header that displays 2 months. Pooling is the right implementation for it since the UI doesn't have a way to skip to a certain date. (it won't need swipe for longer than 2 months..)

If I backout the Bug 988079 it will probably break other marionette tests. So not sure how I should approach this.

Comment 6

[Miller Medeiros [:millermedeiros]]

Attachment: Link to Github pull-request: https://github.com/mozilla-b2g/gaia/pull/17831/files

reverted the patch that introduced the bug and wrote more tests to ensure this kind of regression doesn't happen anymore.

Comment 7

[Evan Tseng [:evanxd]]

Hi Miller,

I think we could also migrate week_view_test.js to use new view code.
So Bug 987356 could be duplicated of this bug.
How do you think?

Thanks.

Comment 8

[Evan Tseng [:evanxd]]

Sorry for that.
I didn't see the https://bugzilla.mozilla.org/show_bug.cgi?id=987356#c1 comment.
Thanks for the fix.

Comment 9

[Miller Medeiros [:millermedeiros]]

landed: https://github.com/mozilla-b2g/gaia/commit/89005c32752bd61b3213e8bf4ba4c3927648c11f

PS: I think this is the first time that I see the calendar integration tests catching a real regression. sorry for all the trouble.