1585239 - Intermittent accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | Uncaught exception - OperationError: The operation failed for an operation-specific reason

Status: RESOLVED INVALID

(Core :: Disability Access APIs, defect, P5)

Description

[Treeherder Bug Filer]

Filed by: aciure [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer.html#?job_id=269191405&repo=autoland
Full log: https://queue.taskcluster.net/v1/task/NmWBB_88T4OccTIfBWqVYg/runs/0/artifacts/public/logs/live_backing.log

---

[task 2019-10-01T03:30:25.179Z] 03:30:25 INFO - TEST-START | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js
[task 2019-10-01T03:30:26.033Z] 03:30:26 INFO - GECKO(1718) | JavaScript error: chrome://mochitests/content/browser/accessible/tests/browser/states/head.js, line 55: TypeError: acc.getAccessibleFor(...) is null
[task 2019-10-01T03:30:26.033Z] 03:30:26 INFO - GECKO(1718) | JavaScript error: , line 0: NS_ERROR_UNEXPECTED
[task 2019-10-01T03:30:26.033Z] 03:30:26 INFO - GECKO(1718) | JavaScript error: , line 0: OperationError: The operation failed for an operation-specific reason
[task 2019-10-01T03:30:26.078Z] 03:30:26 INFO - TEST-INFO | started process screencapture
[task 2019-10-01T03:30:26.226Z] 03:30:26 INFO - TEST-INFO | screencapture: exit 0
[task 2019-10-01T03:30:26.226Z] 03:30:26 INFO - Buffered messages logged at 03:30:25
[task 2019-10-01T03:30:26.228Z] 03:30:26 INFO - Entering test bound
[task 2019-10-01T03:30:26.228Z] 03:30:26 INFO - Console message: OpenGL compositor Initialized Succesfully.
[task 2019-10-01T03:30:26.228Z] 03:30:26 INFO - Version: 2.1 INTEL-12.9.22
[task 2019-10-01T03:30:26.228Z] 03:30:26 INFO - Vendor: Intel Inc.
[task 2019-10-01T03:30:26.228Z] 03:30:26 INFO - Renderer: Intel Iris OpenGL Engine
[task 2019-10-01T03:30:26.228Z] 03:30:26 INFO - FBO Texture Target: TEXTURE_2D
[task 2019-10-01T03:30:26.229Z] 03:30:26 INFO - Buffered messages logged at 03:30:26
[task 2019-10-01T03:30:26.229Z] 03:30:26 INFO - Console message: [JavaScript Error: "OperationError: The operation failed for an operation-specific reason"]
[task 2019-10-01T03:30:26.229Z] 03:30:26 INFO - Console message: [JavaScript Error: "TypeError: acc.getAccessibleFor(...) is null" {file: "chrome://mochitests/content/browser/accessible/tests/browser/states/head.js" line: 55}]
[task 2019-10-01T03:30:26.229Z] 03:30:26 INFO - Console message: [JavaScript Error: "NS_ERROR_UNEXPECTED"]
[task 2019-10-01T03:30:26.230Z] 03:30:26 INFO - Buffered messages finished
[task 2019-10-01T03:30:26.230Z] 03:30:26 INFO - TEST-UNEXPECTED-FAIL | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | Uncaught exception - OperationError: The operation failed for an operation-specific reason
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - Not taking screenshot here: see the one that was previously logged
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - TEST-UNEXPECTED-FAIL | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | A promise chain failed to handle a rejection: The operation failed for an operation-specific reason - stack: (No stack available.)
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - Rejection date: Tue Oct 01 2019 03:30:25 GMT+0000 (Greenwich Mean Time) - false == true - JS frame :: resource://testing-common/PromiseTestUtils.jsm :: assertNoUncaughtRejections :: line 265
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - Stack trace:
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - resource://testing-common/PromiseTestUtils.jsm:assertNoUncaughtRejections:265
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - chrome://mochikit/content/browser-test.js:Tester_execTest/<:1381
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - chrome://mochikit/content/browser-test.js:Tester_execTest:1385
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - chrome://mochikit/content/browser-test.js:nextTest/<:1213
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - chrome://mochikit/content/tests/SimpleTest/SimpleTest.js:SimpleTest.waitForFocus/waitForFocusInner/focusedOrLoaded/<:805
[task 2019-10-01T03:30:26.235Z] 03:30:26 INFO - Leaving test bound
[task 2019-10-01T03:30:26.236Z] 03:30:26 INFO - GECKO(1718) | MEMORY STAT | vsize 7045MB | residentFast 339MB | heapAllocated 155MB
[task 2019-10-01T03:30:26.236Z] 03:30:26 INFO - TEST-OK | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | took 895ms
[task 2019-10-01T03:30:26.236Z] 03:30:26 INFO - checking window state

Comment 1

[Intermittent Failures Robot]

7 failures in 4292 pushes (0.002 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 4
• mozilla-central: 3

Platform breakdown:

• macosx1014-64-shippable: 3
• windows10-aarch64: 1
• linux64-shippable: 2
• linux64: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-09-30&endday=2019-10-06&tree=all

Comment 2

[Intermittent Failures Robot]

14 failures in 3801 pushes (0.004 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• mozilla-central: 7
• autoland: 4
• try: 3

Platform breakdown:

• linux64-shippable: 2
• windows7-32-shippable: 2
• windows10-aarch64: 6
• linux64: 3
• macosx1014-64-shippable: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-10-07&endday=2019-10-13&tree=all

Comment 4

[:Gijs (he/him)]

Hiro or Surkov, can you take a look? From the logs, it seems the getAccessibleFor call at https://searchfox.org/mozilla-central/rev/17756e2a5c180d980a4b08d99f8cc0c97290ae8d/accessible/tests/browser/states/head.js#41 can return null.

Comment 5

[Hiroyuki Ikezoe (:hiro)]

Sure, I can take a look. It seems that we need to make sure the accessibility service has been initialized in the cross-process iframe.

Comment 6

[alexander :surkov (:asurkov)]

it's a bit weird that there's no any passing logs [1], the logs starts right from the "acc.getAccessibleFor(...) is null" failure. I suppose though that the first one fails [2], which means waitForIFrameUpdates doesn't always work [3]. Also see discussion at [4], where waitForIFrameUpdates was flaky on osx.

So, in order to make a11y feel ok, I think waitForIFrameUpdates method should also check whether iframe accessible document is in good shape yet (checking BUSY state should be ok, similar to [5]).

[1] https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=269191405&repo=autoland&lineNumber=1095-1096
[2] https://searchfox.org/mozilla-central/source/accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js#61
[3] https://searchfox.org/mozilla-central/source/accessible/tests/browser/states/head.js#25
[4] https://phabricator.services.mozilla.com/D44951#inline-274332
[5] https://searchfox.org/mozilla-central/source/accessible/tests/mochitest/common.js#152

Comment 7

[Intermittent Failures Robot]

8 failures in 3964 pushes (0.002 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 6
• mozilla-central: 2

Platform breakdown:

• linux64: 5
• linux64-shippable: 1
• macosx1014-64-shippable: 2

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-10-14&endday=2019-10-20&tree=all

Comment 8

[Intermittent Failures Robot]

11 failures in 4182 pushes (0.003 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 6
• mozilla-central: 3
• try: 2

Platform breakdown:

• macosx1014-64-shippable: 2
• linux64: 5
• linux64-shippable: 4

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-10-21&endday=2019-10-27&tree=all

Comment 9

[Intermittent Failures Robot]

14 failures in 4314 pushes (0.003 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• try: 6
• autoland: 7
• mozilla-central: 1

Platform breakdown:

• macosx1014-64-shippable: 4
• linux64: 4
• linux64-qr: 2
• linux64-shippable: 4

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-10-28&endday=2019-11-03&tree=all

Comment 10

[Intermittent Failures Robot]

10 failures in 4850 pushes (0.002 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• mozilla-central: 3
• autoland: 6
• try: 1

Platform breakdown:

• linux64-shippable: 3
• linux64: 4
• macosx1014-64-shippable: 3

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-11-04&endday=2019-11-10&tree=all

Comment 11

[Intermittent Failures Robot]

16 failures in 4207 pushes (0.004 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 12
• mozilla-central: 3
• try: 1

Platform breakdown:

• macosx1014-64-shippable: 9
• linux64-shippable: 4
• linux64: 3

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-11-11&endday=2019-11-17&tree=all

Comment 12

[Intermittent Failures Robot]

20 failures in 5444 pushes (0.004 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 10
• mozilla-central: 5
• try: 5

Platform breakdown:

• linux64-shippable: 8
• macosx1014-64-shippable: 4
• linux64: 8

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-11-18&endday=2019-11-24&tree=all

Comment 13

[Andreea Pavel [:apavel]]

There are 22 total failures in the last 7 days on linux64 opt and asan, linux64-shippable opt and macosx1014-64-shippable opt.

Recent failure log: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=278353409&repo=autoland&lineNumber=1880

[task 2019-11-27T02:29:31.309Z] 02:29:31 INFO - TEST-START | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js
[task 2019-11-27T02:29:31.914Z] 02:29:31 INFO - GECKO(1909) | JavaScript error: chrome://mochitests/content/browser/accessible/tests/browser/states/head.js, line 42: TypeError: acc.getAccessibleFor(...) is null
[task 2019-11-27T02:29:31.914Z] 02:29:31 INFO - GECKO(1909) | JavaScript error: , line 0: NS_ERROR_UNEXPECTED
[task 2019-11-27T02:29:31.915Z] 02:29:31 INFO - GECKO(1909) | JavaScript error: , line 0: OperationError: The operation failed for an operation-specific reason
[task 2019-11-27T02:29:31.935Z] 02:29:31 INFO - TEST-INFO | started process screentopng
[task 2019-11-27T02:29:32.539Z] 02:29:32 INFO - TEST-INFO | screentopng: exit 0
[task 2019-11-27T02:29:32.542Z] 02:29:32 INFO - Buffered messages logged at 02:29:31
[task 2019-11-27T02:29:32.542Z] 02:29:32 INFO - Entering test bound
[task 2019-11-27T02:29:32.542Z] 02:29:32 INFO - Console message: [JavaScript Error: "TypeError: acc.getAccessibleFor(...) is null" {file: "chrome://mochitests/content/browser/accessible/tests/browser/states/head.js" line: 42}]
[task 2019-11-27T02:29:32.542Z] 02:29:32 INFO - Console message: [JavaScript Error: "NS_ERROR_UNEXPECTED"]
[task 2019-11-27T02:29:32.542Z] 02:29:32 INFO - Console message: [JavaScript Error: "OperationError: The operation failed for an operation-specific reason"]
[task 2019-11-27T02:29:32.542Z] 02:29:32 INFO - Buffered messages finished
[task 2019-11-27T02:29:32.545Z] 02:29:32 INFO - TEST-UNEXPECTED-FAIL | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | Uncaught exception - OperationError: The operation failed for an operation-specific reason
[task 2019-11-27T02:29:32.545Z] 02:29:32 INFO - Not taking screenshot here: see the one that was previously logged
[task 2019-11-27T02:29:32.545Z] 02:29:32 INFO - TEST-UNEXPECTED-FAIL | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | A promise chain failed to handle a rejection: The operation failed for an operation-specific reason - stack: (No stack available.)
[task 2019-11-27T02:29:32.545Z] 02:29:32 INFO - Rejection date: Wed Nov 27 2019 02:29:31 GMT+0000 (Coordinated Universal Time) - false == true - JS frame :: resource://testing-common/PromiseTestUtils.jsm :: assertNoUncaughtRejections :: line 265
[task 2019-11-27T02:29:32.546Z] 02:29:32 INFO - Stack trace:
[task 2019-11-27T02:29:32.546Z] 02:29:32 INFO - resource://testing-common/PromiseTestUtils.jsm:assertNoUncaughtRejections:265
[task 2019-11-27T02:29:32.546Z] 02:29:32 INFO - chrome://mochikit/content/browser-test.js:Tester_execTest/<:1100
[task 2019-11-27T02:29:32.546Z] 02:29:32 INFO - chrome://mochikit/content/browser-test.js:Tester_execTest:1104
[task 2019-11-27T02:29:32.547Z] 02:29:32 INFO - chrome://mochikit/content/browser-test.js:nextTest/<:932
[task 2019-11-27T02:29:32.547Z] 02:29:32 INFO - chrome://mochikit/content/tests/SimpleTest/SimpleTest.js:SimpleTest.waitForFocus/waitForFocusInner/focusedOrLoaded/<:805
[task 2019-11-27T02:29:32.548Z] 02:29:32 INFO - Leaving test bound
[task 2019-11-27T02:29:32.548Z] 02:29:32 INFO - GECKO(1909) | MEMORY STAT | vsize 2844MB | residentFast 357MB | heapAllocated 158MB
[task 2019-11-27T02:29:32.549Z] 02:29:32 INFO - TEST-OK | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | took 636ms

Hiroyuki can you assign someone to take a look?

Comment 14

[Hiroyuki Ikezoe (:hiro)]

The "someone" is me. :)

I started looking into this again, and tried the way what :surkov suggested in comment 6. A key factor is waitForFocus in addA11Loadevent, without waiting for focus, the test fails locally even if we check the busy state for the iframe document. )Note that I can't reproduce the failure locally with the current m-c. )

Comment 15

[Hiroyuki Ikezoe (:hiro)]

It didn't fix the failure completely. See https://treeherder.mozilla.org/#/jobs?repo=try&revision=90806830abfdadead49ef9b274dcac5c4cf615e4&selectedJob=278549669

The failure told us that even if we can get the accessible document for the cross-process iframe, we still fail to get the accessible element in question in the document. I will add code to try the accessible element until it succeeds.

Comment 16

[Hiroyuki Ikezoe (:hiro)]

(In reply to Hiroyuki Ikezoe (:hiro) from comment #15)

It didn't fix the failure completely. See https://treeherder.mozilla.org/#/jobs?repo=try&revision=90806830abfdadead49ef9b274dcac5c4cf615e4&selectedJob=278549669

The failure told us that even if we can get the accessible document for the cross-process iframe, we still fail to get the accessible element in question in the document. I will add code to try the accessible element until it succeeds.

This worked on a try run.
https://treeherder.mozilla.org/#/jobs?repo=try&revision=c6b4acfbe261cb3685a6f9c870d4a1fbb2956e4e

Comment 17

[Hiroyuki Ikezoe (:hiro)]

Attachment: Bug 1585239 - Wait for accessibility document ready in iframe. r?surkov

Comment 18

[Hiroyuki Ikezoe (:hiro)]

Attachment: Bug 1585239 - Try to get the accessible element in out-of-process iframe until it succeeds. r?surkov

We sometime fail to nsIAccessibilityService.getAccessibleFor to get the
accessible element in out-of-process iframe even if the accessible document of
the iframe is ready to be used, so we try to get the accessible element until we
succeed to work around this issue.

Depends on D55067

Comment 19

[Intermittent Failures Robot]

13 failures in 4331 pushes (0.003 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• try: 2
• autoland: 9
• mozilla-central: 2

Platform breakdown:

• linux64: 4
• linux64-shippable: 4
• macosx1014-64-shippable: 5

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-11-25&endday=2019-12-01&tree=all

Comment 20

[Hiroyuki Ikezoe (:hiro)]

Note that D55067 doesn't fix the failure completely but I hope it will mitigate it. So I am going to land it now.

Comment 21

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/120e92938637
Wait for accessibility document ready in iframe. r=surkov

Comment 22

[Noemi Erli[:noemi_erli]]

https://hg.mozilla.org/mozilla-central/rev/120e92938637

Comment 23

[Intermittent Failures Robot]

48 failures in 5144 pushes (0.009 failures/push) were associated with this bug in the last 7 days.

This is the #14 most frequent failure this week.

** This failure happened more than 30 times this week! Resolving this bug is a high priority. **

** Try to resolve this bug as soon as possible. If unresolved for 2 weeks, the affected test(s) may be disabled. **

Repository breakdown:

• mozilla-central: 14
• try: 11
• autoland: 23

Platform breakdown:

• linux64: 9
• macosx1014-64-devedition: 1
• linux64-shippable: 21
• windows7-32-shippable: 9
• windows7-32: 1
• windows10-64-shippable: 2
• linux64-devedition: 2
• windows7-32-devedition: 1
• windows10-64: 1
• macosx1014-64-shippable: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-12-02&endday=2019-12-08&tree=all

Comment 24

[:Gijs (he/him)]

This actually seems worse now than before... Hiro, are you still working on this?

Comment 25

[Andreea Pavel [:apavel]]

There are 37 total failures in the last 7 days on linux64 opt, linux64-shippable opt, windows7-32-shippable opt

Recent failure log: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=280073463&repo=mozilla-central&lineNumber=1353

[task 2019-12-06T22:50:07.760Z] 22:50:07 INFO - TEST-START | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js
[task 2019-12-06T22:50:08.181Z] 22:50:08 INFO - GECKO(4988) | JavaScript error: chrome://mochitests/content/browser/accessible/tests/browser/states/head.js, line 71: TypeError: acc.getAccessibleFor(...) is null
[task 2019-12-06T22:50:08.181Z] 22:50:08 INFO - GECKO(4988) | JavaScript error: , line 0: NS_ERROR_UNEXPECTED
[task 2019-12-06T22:50:08.186Z] 22:50:08 INFO - GECKO(4988) | JavaScript error: , line 0: OperationError: The operation failed for an operation-specific reason
[task 2019-12-06T22:50:08.225Z] 22:50:08 INFO - TEST-INFO | started process screenshot
[task 2019-12-06T22:50:08.304Z] 22:50:08 INFO - TEST-INFO | screenshot: exit 0
[task 2019-12-06T22:50:08.304Z] 22:50:08 INFO - Buffered messages logged at 22:50:07
[task 2019-12-06T22:50:08.304Z] 22:50:08 INFO - Entering test bound
[task 2019-12-06T22:50:08.304Z] 22:50:08 INFO - Buffered messages logged at 22:50:08
[task 2019-12-06T22:50:08.304Z] 22:50:08 INFO - must wait for focus
[task 2019-12-06T22:50:08.304Z] 22:50:08 INFO - Console message: [JavaScript Error: "TypeError: acc.getAccessibleFor(...) is null" {file: "chrome://mochitests/content/browser/accessible/tests/browser/states/head.js" line: 71}]
[task 2019-12-06T22:50:08.305Z] 22:50:08 INFO - Console message: [JavaScript Error: "NS_ERROR_UNEXPECTED"]
[task 2019-12-06T22:50:08.305Z] 22:50:08 INFO - Console message: [JavaScript Error: "OperationError: The operation failed for an operation-specific reason"]
[task 2019-12-06T22:50:08.305Z] 22:50:08 INFO - Buffered messages finished
[task 2019-12-06T22:50:08.305Z] 22:50:08 INFO - TEST-UNEXPECTED-FAIL | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | Uncaught exception - OperationError: The operation failed for an operation-specific reason
[task 2019-12-06T22:50:08.305Z] 22:50:08 INFO - Not taking screenshot here: see the one that was previously logged
[task 2019-12-06T22:50:08.305Z] 22:50:08 INFO - TEST-UNEXPECTED-FAIL | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | A promise chain failed to handle a rejection: The operation failed for an operation-specific reason - stack: (No stack available.)
[task 2019-12-06T22:50:08.305Z] 22:50:08 INFO - Rejection date: Fri Dec 06 2019 22:50:08 GMT+0000 (Greenwich Mean Time) - false == true - JS frame :: resource://testing-common/PromiseTestUtils.jsm :: assertNoUncaughtRejections :: line 265
[task 2019-12-06T22:50:08.305Z] 22:50:08 INFO - Stack trace:
[task 2019-12-06T22:50:08.306Z] 22:50:08 INFO - resource://testing-common/PromiseTestUtils.jsm:assertNoUncaughtRejections:265
[task 2019-12-06T22:50:08.306Z] 22:50:08 INFO - chrome://mochikit/content/browser-test.js:Tester_execTest/<:1100
[task 2019-12-06T22:50:08.306Z] 22:50:08 INFO - chrome://mochikit/content/browser-test.js:Tester_execTest:1104
[task 2019-12-06T22:50:08.306Z] 22:50:08 INFO - chrome://mochikit/content/browser-test.js:nextTest/<:932
[task 2019-12-06T22:50:08.306Z] 22:50:08 INFO - chrome://mochikit/content/tests/SimpleTest/SimpleTest.js:SimpleTest.waitForFocus/waitForFocusInner/focusedOrLoaded/<:805
[task 2019-12-06T22:50:08.306Z] 22:50:08 INFO - Leaving test bound
[task 2019-12-06T22:50:08.306Z] 22:50:08 INFO - GECKO(4988) | MEMORY STAT | vsize 2104174MB | vsizeMaxContiguous 66103308MB | residentFast 310MB | heapAllocated 147MB
[task 2019-12-06T22:50:08.306Z] 22:50:08 INFO - TEST-OK | accessible/tests/browser/states/browser_offscreen_element_in_out_of_process_iframe.js | took 496ms

There seems to be a spike since 1st of December, however so far there are no failures since the 8th of December.

Comment 26

[Hiroyuki Ikezoe (:hiro)]

I have no further idea other than D55068. It would be nice someone who is familiar with accessibility code looks into this.

Comment 27

[alexander :surkov (:asurkov)]

(In reply to Hiroyuki Ikezoe (:hiro) from comment #26)

I have no further idea other than D55068. It would be nice someone who is familiar with accessibility code looks into this.

Does suggestion from https://phabricator.services.mozilla.com/D55068#1685788 not work?

If not, then I'd delegate the thing to Jamie, whether the approach form the patch should be taken or should be investigated further (note, it should work, just not optimal).

Comment 28

[Hiroyuki Ikezoe (:hiro)]

It doesn't fix the intermittent failure. It's been already replaced.

Comment 29

[James Teh [:Jamie]]

I think it'd make sense to tweak this test so that it observes a11y events from the parent process, rather than trying to do it from the child. That's what we do in most of our other e10s/Fission tests. That should make the flow simpler, as well as reducing some potential race conditions. Yura, would you mind taking a look at this when you get a chance?

Comment 30

[alexander :surkov (:asurkov)]

(In reply to James Teh [:Jamie] from comment #29)

I think it'd make sense to tweak this test so that it observes a11y events from the parent process, rather than trying to do it from the child. That's what we do in most of our other e10s/Fission tests. That should make the flow simpler, as well as reducing some potential race conditions. Yura, would you mind taking a look at this when you get a chance?

Out of curiosity could you elaborate how listening events in parent process could help? My understanding (per D55068) is there's no accessible for an element in a child process when the test tries to access to it, but its document is not busy (so it's constructed) and there's no DOM/layout mutations, which has to guarantee the element is accessible, and - I must be wrong - but it seems there's no direct correlation with parent/child process architecture, and thus it's not clear how moving the logic into parent may help (extra timing perhaps?).

Comment 31

[James Teh [:Jamie]]

(In reply to alexander :surkov (:asurkov) from comment #30)

Out of curiosity could you elaborate how listening events in parent process could help?

You're right: it won't help directly. However, the flow as it stands is somewhat hard to follow, and it's also inconsistent with the way we handle this kind of thing in other tests. That means we need to debug the utility functions used only for this test and makes it hard to diagnose/predict possible race conditions, etc.

That said, looking at this again, I just spotted this:
https://searchfox.org/mozilla-central/rev/6305f6935f496b3a302c7afcc579399a4217729c/accessible/tests/browser/states/head.js#26
The Promise gets created, but not returned or awaited. I think that means that waitForReady will just resolve immediately, so we don't actually wait until the busy state gets cleared. That's perhaps the cause of the problem here.

Comment 32

[alexander :surkov (:asurkov)]

(In reply to James Teh [:Jamie] from comment #31)

(In reply to alexander :surkov (:asurkov) from comment #30)

Out of curiosity could you elaborate how listening events in parent process could help?

You're right: it won't help directly. However, the flow as it stands is somewhat hard to follow, and it's also inconsistent with the way we handle this kind of thing in other tests. That means we need to debug the utility functions used only for this test and makes it hard to diagnose/predict possible race conditions, etc.

fair enough

That said, looking at this again, I just spotted this:
https://searchfox.org/mozilla-central/rev/6305f6935f496b3a302c7afcc579399a4217729c/accessible/tests/browser/states/head.js#26
The Promise gets created, but not returned or awaited. I think that means that waitForReady will just resolve immediately, so we don't actually wait until the busy state gets cleared. That's perhaps the cause of the problem here.

You're right. Hiro, will you be about to check this finding out anytime soon?

Comment 33

[Hiroyuki Ikezoe (:hiro)]

Gosh! Thanks for finding it!

I just pushed a try with an await there.
https://treeherder.mozilla.org/#/jobs?repo=try&revision=84e3da21e477e814ec833e7c1801f71261f6057a

Comment 34

[Hiroyuki Ikezoe (:hiro)]

Though the try result looks great, but I just realized this intermittent failure has stopped since Dec. 7th. I don't know what has changed.

Comment 35

[Hiroyuki Ikezoe (:hiro)]

Attachment: Bug 1585239 - Return the Promise for the accessibility document in the function runs in the iframe context. r?surkov

Comment 36

[James Teh [:Jamie]]

Cancelling NI Yura for now, since hopefully this will be a simple fix.

Comment 37

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/5e8b48c8cd93
Return the Promise for the accessibility document in the function runs in the iframe context. r=surkov

Comment 38

[Oana Pop-Rus]

https://hg.mozilla.org/mozilla-central/rev/5e8b48c8cd93

Comment 39

[Intermittent Failures Robot]

3 failures in 4582 pushes (0.001 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• try: 2
• oak: 1

Platform breakdown:

• linux64: 2
• linux64-shippable: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1585239&startday=2019-12-16&endday=2019-12-22&tree=all

Comment 40

[BugBot [:suhaib / :marco/ :calixte]]

The following patch is waiting for review from an inactive reviewer:

ID	Title	Author	Reviewer Status
D55068	Bug 1585239 - Try to get the accessible element in out-of-process iframe until it succeeds. r?surkov	hiro	surkov: Inactive

:hiro, could you please find another reviewer or abandon the patch if it is no longer relevant?

For more information, please visit auto_nag documentation.

Comment 41

[Hiroyuki Ikezoe (:hiro)]

Looks like the test itself has been removed in bug 1787274.