Closed Bug 860421 Opened 12 years ago Closed 11 years ago

Intermittent browser_social_chatwindow.js | correct number of chats visible - Got 2, expected 3

Categories

(Firefox Graveyard :: SocialAPI, defect)

Product:
Firefox Graveyard
Component:
SocialAPI
Platform:
x86
Linux
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(firefox24 fixed, firefox25 fixed)

Status:
RESOLVED FIXED
Milestone:
Firefox 25
Tracking Flags:
Tracking Status
firefox24 --- fixed
firefox25 --- fixed

People

(Reporter: RyanVM, Assigned: mixedpuppy)

References

Details

(Keywords: intermittent-failure)

Attachments

(1 file, 1 obsolete file)

fix intermittent failure
24.51 KB, patch
markh
: review+
bajaj
: approval-mozilla-aurora+
Details | Diff | Splinter Review
https://tbpl.mozilla.org/php/getParsedLog.php?id=21649111&tree=Mozilla-Inbound Rev3 Fedora 12 mozilla-inbound opt test mochitest-browser-chrome on 2013-04-10 02:04:58 PDT for push 78e13e815d62 slave: talos-r3-fed-046 09:10:59 INFO - TEST-PASS | chrome://mochitests/content/browser/browser/base/content/test/social/browser_social_chatwindow.js | collapsed chat has a menu item 09:10:59 INFO - TEST-PASS | chrome://mochitests/content/browser/browser/base/content/test/social/browser_social_chatwindow.js | open chat has no menu item 09:10:59 INFO - TEST-PASS | chrome://mochitests/content/browser/browser/base/content/test/social/browser_social_chatwindow.js | popup matches child collapsed state 09:10:59 INFO - TEST-PASS | chrome://mochitests/content/browser/browser/base/content/test/social/browser_social_chatwindow.js | popup has correct count of children 09:10:59 WARNING - TEST-UNEXPECTED-FAIL | chrome://mochitests/content/browser/browser/base/content/test/social/browser_social_chatwindow.js | correct number of chats visible - Got 2, expected 3 09:10:59 INFO - Stack trace: 09:10:59 INFO - JS frame :: chrome://mochikit/content/browser-test.js :: test_is :: line 502 09:10:59 INFO - JS frame :: chrome://mochitests/content/browser/browser/base/content/test/social/browser_social_chatwindow.js :: <TOP_LEVEL> :: line 691 09:10:59 INFO - JS frame :: chrome://mochitests/content/browser/browser/base/content/test/social/browser_social_chatwindow.js :: resize_handler :: line 674 09:10:59 INFO - native frame :: <unknown filename> :: <TOP_LEVEL> :: line 0 09:10:59 INFO - TEST-INFO | chrome://mochitests/content/browser/browser/base/content/test/social/browser_social_chatwindow.js | Check: 2 pixels less and the first is again collapsed (and the nub re-appears)
Attached patch fix intermittent failure (obsolete) — — Splinter Review
I've been able to repro on a recently updated ubuntu (where almost all the test failures are happening). Looking at other resize based tests, I followed what is happening here: https://mxr.mozilla.org/mozilla-central/source/browser/components/tabview/test/browser_tabview_bug625269.js#63 The key here is to wait for resize to actually give use the size we're looking for, probably some transition happening, certainly multiple resize events. This seems to have fixed on my local ubuntu vm. https://tbpl.mozilla.org/?tree=Try&rev=736ca27716fb
Assignee: nobody → mixedpuppy
Attachment #753988 - Flags: review?(felipc)
Comment on attachment 753988 [details] [diff] [review] fix intermittent failure Review of attachment 753988 [details] [diff] [review]: ----------------------------------------------------------------- This looks good in general and would explain some of the failures, however note that: info("skipping this as we can't resize chat area to " + desired + " - got " + newSize); was actually being hit in some cases on try, where the screen-size of the test machine wasn't quite large enough to accommodate the test. This might not be a problem any more - a try run will tell you!
(In reply to Mark Hammond (:markh) from comment #48) > Comment on attachment 753988 [details] [diff] [review] > fix intermittent failure > > Review of attachment 753988 [details] [diff] [review]: > ----------------------------------------------------------------- > > This looks good in general and would explain some of the failures, however > note that: > > info("skipping this as we can't resize chat area to " + desired + " - got " > + newSize); > > was actually being hit in some cases on try, where the screen-size of the > test machine wasn't quite large enough to accommodate the test. This might > not be a problem any more - a try run will tell you! I don't seem to be running into that with the try builds, I think we can let it go in and fix later if that is still an issue somewhere.
Attachment #753988 - Flags: review?(felipc) → review+
https://hg.mozilla.org/mozilla-central/rev/7e189fc274a7
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Target Milestone: --- → Firefox 24
Depends on: 886056
a try build with a minor fix and an extra info statement. once it builds, I'll run a bunch of the tests to see if it indicates any further clues. https://tbpl.mozilla.org/?tree=Try&rev=61bf2f871794
(first come first serve on review) The fix for this test is much simpler than the patch looks. To simplify debugging I moved the failing test to its own test file, to ensure the other tests were not responsible. I moved support functions to head.js, added a bunch of info output that I am leaving for future debugging if something else crops up here. The actual fix is in head.js in the resize_handler function, where I now check the phase of the resize event. I also check the target in socialchat.xml. We were handling resize twice in both locations (*intermittently* in the test, *always* from xbl). I dont quite understand why resize_handler was called twice for different phases under debug builds, but we were. I was finally able to 100% reproduce on a linux debug build, matching the problem on try. This try is pre-cleanup, but is showing the failure doesn't happen (when it was 100% on previous try attempts). https://tbpl.mozilla.org/?tree=Try&rev=2e3f2005c25f This try is with the cleaned up patch and on more platforms (where we were very intermittent) https://tbpl.mozilla.org/?tree=Try&rev=e668bf68a9e7
Attachment #753988 - Attachment is obsolete: true
Attachment #771441 - Flags: review?(mhammond)
Attachment #771441 - Flags: review?(jaws)
Attachment #771441 - Flags: review?(felipc)
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
Comment on attachment 771441 [details] [diff] [review] fix intermittent failure Review of attachment 771441 [details] [diff] [review]: ----------------------------------------------------------------- Seems like black-magic, but I'll take your word for it (and I do know the symptoms were indeed seeing some of those resize events fired twice)...
Attachment #771441 - Flags: review?(mhammond)
Attachment #771441 - Flags: review?(jaws)
Attachment #771441 - Flags: review?(felipc)
Attachment #771441 - Flags: review+
https://hg.mozilla.org/mozilla-central/rev/769c90e47963
Status: REOPENED → RESOLVED
Closed: 12 years ago → 11 years ago
Flags: in-testsuite+
Resolution: --- → FIXED
Comment on attachment 771441 [details] [diff] [review] fix intermittent failure [Approval Request Comment] Bug caused by (feature/regressing bug #): 891225 User impact if declined: none, this is an intermittent orange fix Testing completed (on m-c, etc.): m-c and inbound have not given me an orange since this landed, but aurora and even beta are still producing them Risk to taking this patch (and alternatives if risky): primarily a test fix String or IDL/UUID changes made by this patch: none
Attachment #771441 - Flags: approval-mozilla-aurora?
Attachment #771441 - Flags: approval-mozilla-aurora? → approval-mozilla-aurora+
lol, this never got uplifted due to the target milestone being set incorrectly
Target Milestone: Firefox 24 → Firefox 25
Depends on: 908651
Product: Firefox → Firefox Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: