Closed
Bug 961137
Opened 12 years ago
Closed 12 years ago
[Search] Places marionette test disabled
Categories
(addons.mozilla.org Graveyard :: Search, defect)
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: kgrandon, Assigned: daleharvey)
Details
(Whiteboard: [systemsfe])
Attachments
(1 file, 1 obsolete file)
Intermittent on travis, going to disable the test for now.
| Reporter | ||
Comment 1•12 years ago
|
||
| Assignee | ||
Comment 2•12 years ago
|
||
To be honest was a bit lazy of me, but the change notification is sent before the change is guaranteed to have been written, its either that or the 'press home in between searching' quite stable enough, will get this passing 100%, apologies.
Assignee: nobody → dale
| Reporter | ||
Comment 3•12 years ago
|
||
Disbaled other test, but lost the log in my clipboard, sorry =/
https://github.com/mozilla-b2g/gaia/commit/df3b929f3e979e5e7eef5d8394a4b800c5c94ed8
| Assignee | ||
Comment 4•12 years ago
|
||
> change notification is sent before the change is guaranteed to have been written
Is fixed by https://github.com/mozilla-b2g/gaia/commit/a1e985d3a045daa796812712dfb81486becc5a9b
I also started using a new approach to tests with the favicon thing, I now repeatedly search until we get the results I want, we had problems with things not being completed (icon fetched, place saved, store synced), this approach solved that, so I think we should be able to get these 100% now
| Assignee | ||
Comment 5•12 years ago
|
||
I am gonna split this into 2 bugs, I think the original tests are fixed, looking green to me but going to do a 100 run before I do a PR
The favicon bug and issues Kevin has been seeing are a fairly serious issue with our places implementation
| Assignee | ||
Comment 6•12 years ago
|
||
As mentioned the fix for this was in https://github.com/mozilla-b2g/gaia/commit/a1e985d3a045daa796812712dfb81486becc5a9b
Its going on 50 green runs between https://travis-ci.org/daleharvey/gaia/builds/17406566 and https://travis-ci.org/daleharvey/gaia/builds/17408429
Attachment #8363645 -
Flags: review?(kgrandon)
| Assignee | ||
Comment 7•12 years ago
|
||
The problem with the favicon bug is filed @ https://bugzilla.mozilla.org/show_bug.cgi?id=962554
| Reporter | ||
Comment 8•12 years ago
|
||
Interesting. Test run looks good, but there was one failure.
1) Places tests Search for previously visited URL:
+ expected - actual
+http://localhost:51835/sample.html
-Sample page
I suppose what happens is that we are storing the place for the URL, before the page loads? If that's the case, it's probably not a problem with the test - but perhaps we are storing more than we want to? Let me know what you think..
Flags: needinfo?(dale)
| Assignee | ||
Comment 9•12 years ago
|
||
It could possibly be racey due to https://bugzilla.mozilla.org/show_bug.cgi?id=962554, another test run with both fixes going now https://travis-ci.org/daleharvey/gaia/builds/17447160
Flags: needinfo?(dale)
Updated•12 years ago
|
Whiteboard: [systemsfe]
| Reporter | ||
Comment 10•12 years ago
|
||
Comment on attachment 8363645 [details] [review]
https://github.com/mozilla-b2g/gaia/pull/15601
Two failures on that last run, so going to clear the review flag for now. I have some ideas of what might be happening, and will try to help out once I get the rocketbar stuff into master.
Attachment #8363645 -
Flags: review?(kgrandon)
| Assignee | ||
Comment 11•12 years ago
|
||
Found the issue, code assumed the place had been written before attempting to set icons / titles, its related to https://bugzilla.mozilla.org/show_bug.cgi?id=962554
the 100 runs had one failure, https://travis-ci.org/daleharvey/gaia/builds/17477958 but it was a crash unrelated to this patch
Attachment #8363645 -
Attachment is obsolete: true
Attachment #8364883 -
Flags: review?(kgrandon)
| Reporter | ||
Comment 12•12 years ago
|
||
Comment on attachment 8364883 [details]
https://github.com/daleharvey/gaia/commit/d003522d5c659417881183c62a5424adad37df2c
I am a little bit worried about the failure I saw in travis, but doesn't appear to be caused by this patch. I'd say we should land this and we can investigate the other test if we see it fail again.
1) navigation "before each" hook:
Error: timeout of 20000ms exceeded
at null.<anonymous> (/home/travis/build/daleharvey/gaia/node_modules/mocha/lib/runnable.js:165:14)
at Timer.listOnTimeout [as ontimeout] (timers.js:110:15)
2) navigation "after each" hook:
Error: Not connected. To write data you must call connect first.
Attachment #8364883 -
Flags: review?(kgrandon) → review+
| Assignee | ||
Comment 13•12 years ago
|
||
Yeh no failures are ok, but given the 600 or so runs I have done, having one ghost seems likely, the changes in the test are needed so at the least they can land and can redisable the tests if theres a problem, it doesnt look like there should be though
https://github.com/mozilla-b2g/gaia/commit/fa90ddf34db8057f9d95b23864902f16e049edd1
| Assignee | ||
Updated•12 years ago
|
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Updated•10 years ago
|
Product: addons.mozilla.org → addons.mozilla.org Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•