874429 - Intermittent test_form_autocomplete.html | Test timed out | test_form_autocomplete_with_list.html | undefined Checking menu entry #0 - got PASS1, expected bc dæ | undefined Checking menu entry #1 - got final, expected abaæl | Autocomplete popup not...

Reporter

Description

•

12 years ago

Attached image screenshot (deleted) — Details

https://tbpl.mozilla.org/php/getParsedLog.php?id=23168038&tree=Mozilla-Inbound Windows XP 32-bit mozilla-inbound debug test mochitest-5 on 2013-05-20 12:34:58 PDT for push 905e081060c5 slave: t-xp32-ix-082 13:50:48 INFO - 175118 ERROR TEST-UNEXPECTED-FAIL | /tests/toolkit/components/satchel/test/test_form_autocomplete.html | Test timed out. 13:50:50 INFO - 175203 ERROR TEST-UNEXPECTED-FAIL | /tests/toolkit/components/satchel/test/test_form_autocomplete_with_list.html | undefined Checking menu entry #0 - got PASS1, expected bc dæ 13:50:50 INFO - 175204 ERROR TEST-UNEXPECTED-FAIL | /tests/toolkit/components/satchel/test/test_form_autocomplete_with_list.html | undefined Checking menu entry #1 - got final, expected abaæ 13:50:50 INFO - 175208 ERROR TEST-UNEXPECTED-FAIL | /tests/toolkit/components/satchel/test/test_form_autocomplete_with_list.html | Autocomplete popup not expected during test 200 13:50:50 INFO - 175210 ERROR TEST-UNEXPECTED-FAIL | /tests/toolkit/components/satchel/test/test_form_autocomplete_with_list.html | Checking form3 input - got d, expected Foo 13:56:17 INFO - 175212 ERROR TEST-UNEXPECTED-FAIL | /tests/toolkit/components/satchel/test/test_form_autocomplete_with_list.html | Test timed out.

Comment hidden (Legacy TBPL/Treeherder Robot)

Ryan VanderMeulen [:RyanVM]

Reporter

Comment 113

•

11 years ago

Justin, could you please look into this or suggest someone who can?

Flags: needinfo?(dolske)

Comment hidden (Legacy TBPL/Treeherder Robot)

Justin Dolske [:Dolske]

Comment 119

•

11 years ago

From a quick skim of 3 recent failures, they all start to timeout here: ... 05:15:06 INFO - 176231 INFO TEST-PASS | /tests/toolkit/components/satchel/test/test_form_autocomplete.html | Starting test #252 05:15:06 INFO - 176232 INFO TEST-PASS | /tests/toolkit/components/satchel/test/test_form_autocomplete.html | 252 Checking length of expected menu 05:15:06 INFO - 176233 INFO TEST-PASS | /tests/toolkit/components/satchel/test/test_form_autocomplete.html | 252 Checking menu entry #0 05:15:07 INFO - 176234 INFO TEST-PASS | /tests/toolkit/components/satchel/test/test_form_autocomplete.html | Starting test #253 05:15:07 INFO - 176235 INFO TEST-PASS | /tests/toolkit/components/satchel/test/test_form_autocomplete.html | 253 Checking length of expected menu 05:15:07 INFO - 176236 INFO TEST-PASS | /tests/toolkit/components/satchel/test/test_form_autocomplete.html | 253 Checking menu entry #0 ...then nothing but a bunch of --DOCSHELL and --DOMWINDOW (GC/CC) until the test finally times out. It's slightly suspicious that the last modification to this test (bug 875274) was just a couple days after this bug was filed. But it landed _after_ this was filed, so can't directly be responsible. :) Makes me wonder if something landed to prepare the way and made this start to fail, though. More likely is that test #252 is using a setTimeout() (!!!) to launch test #253, and something now happens after the timeout now to make things go screwy? It's a little weird because 252-->253 is checking to see that nothing happens after typing a space, so in theory the setTimeout isn't really evil here. Unless something now does happen, as so this fails intermittently. The screen shot shows the expected values in the <input>. Beyond that will need some debugging to figure out what the autocomplete stuff is doing. The Hg history is a little muddy, but afaict this setTimeout was added by Enn in bug 697377 as part of the async form history work. Tag, you're it.

Assignee: nobody → enndeakin

Flags: needinfo?(dolske)

Justin Dolske [:Dolske]

Comment 120

•

11 years ago

I guess I should also mention that I'm assuming the test_form_autocomplete.html timeout is the important failure, and that the following test_form_autocomplete_with_list.html failure is just because the test ended uncleanly.

Justin Dolske [:Dolske]

Updated

•

11 years ago

Component: Autocomplete → Form Manager

Comment hidden (Legacy TBPL/Treeherder Robot)

Ryan VanderMeulen [:RyanVM]

Reporter

Comment 192

•

11 years ago

Neil, have you had a chance to look at this yet, by chance? :)

status-firefox24: --- → affected

status-firefox25: --- → affected

status-firefox26: --- → affected

Flags: needinfo?(enndeakin)

Comment hidden (Legacy TBPL/Treeherder Robot)

Neil Deakin

Comment 203

•

11 years ago

I have looked into, but haven't got a failing case with any useful logging yet.

Flags: needinfo?(enndeakin)

Comment hidden (Legacy TBPL/Treeherder Robot)

Ed Morley [:emorley]

Comment 270

•

11 years ago

Test disabled until someone looks into the failures: https://hg.mozilla.org/integration/mozilla-inbound/rev/90c25e207b05

Whiteboard: [test disabled][leave open]

Comment hidden (Legacy TBPL/Treeherder Robot)

Ed Morley [:emorley]

Comment 272

•

11 years ago

https://hg.mozilla.org/mozilla-central/rev/90c25e207b05

Comment hidden (Legacy TBPL/Treeherder Robot)

Ryan VanderMeulen [:RyanVM]

Reporter

Updated

•

11 years ago

Flags: needinfo?(ryanvm)

Ryan VanderMeulen [:RyanVM]

Reporter

Comment 276

•

11 years ago

Disabled on the release branches as well. https://hg.mozilla.org/releases/mozilla-aurora/rev/21eb441fa790 https://hg.mozilla.org/releases/mozilla-beta/rev/d1d4f379db20 https://hg.mozilla.org/releases/mozilla-esr24/rev/72eb79235f20

status-firefox24: affected → ---

status-firefox25: affected → disabled

status-firefox26: affected → disabled

status-firefox27: --- → disabled

status-firefox-esr24: --- → disabled

Flags: needinfo?(ryanvm)

Comment hidden (Legacy TBPL/Treeherder Robot)

Matthew N. [:MattN]

Assignee

Comment 278

•

9 years ago

I have a patch which removes a setTimeout and switches a waitForScroll to a waitForMenuChange which should be less fragile. It also makes the test file work again now that <input type=number> behaviour changed by skipping the type=number specific check for now. https://treeherder.mozilla.org/#/jobs?repo=try&revision=69db980e43a0

Assignee: enndeakin → MattN+bmo

Status: NEW → ASSIGNED

Matthew N. [:MattN]

Assignee

Comment 279

•

9 years ago

Attached file MozReview Request: Bug 874429 - Re-enable test_form_autocomplete.html after removing some a setTimeout and skipping <input type=number> testing. r=dolske (deleted) — Details

Bug 874429 - Re-enable test_form_autocomplete.html after removing some a setTimeout and skipping <input type=number> testing. r=dolske

Attachment #8616555 - Flags: review?(dolske)

Justin Dolske [:Dolske]

Updated

•

9 years ago

Attachment #8616555 - Flags: review?(dolske) → review+

Justin Dolske [:Dolske]

Comment 280

•

9 years ago

Comment on attachment 8616555 [details] MozReview Request: Bug 874429 - Re-enable test_form_autocomplete.html after removing some a setTimeout and skipping <input type=number> testing. r=dolske https://reviewboard.mozilla.org/r/10459/#review9233 Ship It!

Matthew N. [:MattN]

Assignee

Updated

•

9 years ago

OS: Windows XP → All

Hardware: x86 → All

Whiteboard: [test disabled][leave open]

Pulsebot

Comment 281

•

9 years ago

https://hg.mozilla.org/integration/fx-team/rev/dd8df14614df

Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)

Comment 282

•

9 years ago

https://hg.mozilla.org/mozilla-central/rev/dd8df14614df

Status: ASSIGNED → RESOLVED

Closed: 9 years ago

status-firefox41: --- → fixed

Resolution: --- → FIXED

Target Milestone: --- → mozilla41

Ryan VanderMeulen [:RyanVM]

Reporter

Comment 283

•

9 years ago

Not going to bother uplifting this unless you feel strongly about it, Matt. ni? me if so.

status-firefox39: --- → wontfix

status-firefox40: --- → wontfix

status-firefox-esr31: --- → wontfix

status-firefox-esr38: --- → wontfix

:Paolo Amadini

Comment 284

•

9 years ago

(In reply to Matthew N. [:MattN] (behind on bugmail) from comment #278) > It also makes the test file > work again now that <input type=number> behaviour changed by skipping the > type=number specific check for now. Did we lose test coverage because of this change? If so, is there a bug on file to restore it? Does the test case cover functionality that we still need to implement for e10s?

Blocks: 1252074

Flags: needinfo?(MattN+bmo)

Matthew N. [:MattN]

Assignee

Comment 285

•

9 years ago

(In reply to :Paolo Amadini from comment #284) > (In reply to Matthew N. [:MattN] (behind on bugmail) from comment #278) > > It also makes the test file > > work again now that <input type=number> behaviour changed by skipping the > > type=number specific check for now. > > Did we lose test coverage because of this change? Not really, the todo() is still there reminding that it's broken. > If so, is there a bug on > file to restore it? Does the test case cover functionality that we still > need to implement for e10s? I don't know of a bug and I'm not sure if type=number support is only broken in e10s or non-e10s too. I don't even remember what we did special (I guess filtering values that aren't numeric). It seems like we don't show autocomplete at all anymore for type=number and I'm undecided on whether that's good or not. One problem is that the down arrow key event is used to decrement the number instead of opening the popup (but even double-clicking doesn't open autocompletion). I guess this is a problem for UX and it's worth looking at what other browsers do. Test case: data:text/html,<input name="q" type="number"/>

Flags: needinfo?(MattN+bmo)

screenshot 12 years ago Ryan VanderMeulen [:RyanVM] (deleted), image/png		Details
MozReview Request: Bug 874429 - Re-enable test_form_autocomplete.html after removing some a setTimeout and skipping <input type=number> testing. r=dolske 9 years ago Matthew N. [:MattN] (deleted), text/x-review-board-request	Dolske : review+	Details