Closed Bug 779753 Opened 12 years ago Closed 11 years ago

Intermittent test_browserElement_oop_SetVisible.html, test_browserElement_oop_SetVisibleFrames.html, test_browserElement_oop_Stop.html | finished in a non-clean fashion (in /tests/dom/browser-element/mochitest/test_browserElement_oop_SendEvent.html)

Categories

(Core :: DOM: Core & HTML, defect)

Product:
Core
Component:
DOM: Core & HTML
Platform:
x86_64
Linux
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED

People

(Reporter: emorley, Unassigned)

References

(Depends on 1 open bug)

Details

(Keywords: intermittent-failure, Whiteboard: [test disabled on Linux])

Attachments

(2 obsolete files) 

Rev3 Fedora 12x64 mozilla-inbound pgo test mochitests-2/5 on 2012-08-01 20:38:51 PDT for push 20c9acfd33eb

slave: talos-r3-fed64-031

https://tbpl.mozilla.org/php/getParsedLog.php?id=14056403&tree=Mozilla-Inbound

{
1293 INFO TEST-START | /tests/dom/browser-element/mochitest/test_browserElement_oop_SetVisible.html
1294 INFO TEST-PASS | /tests/dom/browser-element/mochitest/test_browserElement_oop_SetVisible.html | Receive a touchstart event.
1295 INFO TEST-PASS | /tests/dom/browser-element/mochitest/test_browserElement_oop_SetVisible.html | Receive a touchmove event.
1296 INFO TEST-PASS | /tests/dom/browser-element/mochitest/test_browserElement_oop_SetVisible.html | Receive a touchend event.
1297 INFO TEST-UNEXPECTED-FAIL | /tests/dom/browser-element/mochitest/test_browserElement_oop_SetVisible.html | finished in a non-clean fashion (in /tests/dom/browser-element/mochitest/test_browserElement_oop_SendEvent.html)
1298 INFO TEST-END | /tests/dom/browser-element/mochitest/test_browserElement_oop_SetVisible.html | finished in 17ms
}
Attached patch Disable test on Linux (obsolete) — — Splinter Review 
These tests are really only B2G only so no need to run on Linux, especially given the failures.
Attachment #653103 - Flags: review?(philringnalda)
Attached patch Disable test on Linux (obsolete) — — Splinter Review 
Without thinko this time.
Attachment #653103 - Attachment is obsolete: true
Attachment #653103 - Flags: review?(philringnalda)
Attachment #653104 - Flags: review?(philringnalda)
Attachment #653104 - Flags: review?(philringnalda) → review?(justin.lebar+bug)
> Attachment #653104 [details] [diff] - Flags: review?(philringnalda@gmail.com) → review?(justin.lebar+bug@gmail.com)

Thanks a lot for re-assigning this review.

(In reply to Ed Morley [:edmorley] from comment #25)
> Created attachment 653103 [details] [diff] [review]
> Disable test on Linux
> 
> These tests are really only B2G only so no need to run on Linux, especially
> given the failures.

We don't have any B2G mochitest coverage; Linux is the closest thing we have.  I'm not comfortable disabling this or any of the browser-element tests on Linux.

I'll see if I can figure out this orange, but I have to prioritize it below B2G blockers, of which there are many.  Sorry, guys.
Attachment #653104 - Flags: review?(justin.lebar+bug) → review-
(In reply to Justin Lebar [:jlebar] from comment #27)
> We don't have any B2G mochitest coverage; Linux is the closest thing we
> have.  I'm not comfortable disabling this or any of the browser-element
> tests on Linux.
> 
> I'll see if I can figure out this orange, but I have to prioritize it below
> B2G blockers, of which there are many.  Sorry, guys.

Ah, I stand corrected :-)

Fair points - though be aware we're pretty much ignoring all dom/browser-element/* failures given the constant stream of them, both new and old - so the sooner we can get some confidence behind the tests, the better :-)
> Fair points - though be aware we're pretty much ignoring all dom/browser-element/* failures given 
> the constant stream of them, both new and old

As far as I can tell from the WOO page, this is the only outstanding browser-element orange.

Are there others I should be keeping track of?  Is this the worst one?
I can't remember off the top of my head, I'll try and have a look over the next day or two. It may just be the Android failures (which have since been disabled) influencing my perception - but either way, there are a set of components/directories in which I'm much more inclined to think "*sigh*, yet another flaky test" (eg anything under content/media/*), and dom/browser-element/* is one that at least for the last week or two, has been on that mental list.
> there are a set of components/directories in which I'm much more inclined to think "*sigh*, yet 
> another flaky test" (eg anything under content/media/*), and dom/browser-element/* is one that at 
> least for the last week or two, has been on that mental list.

/That/ is certainly fair, and frequent orange in my tests has been going on for much longer than the last week or two.

But I really hope we don't get to the point where we regularly ignore new oranges in these tests.  I'm putting a lot of effort into writing (and fixing) the tests!

(Looking at WOO again, the only other browser-element orange I see is bug 774676, which appears to be a bug in gfx code that I'm pretty clueless about.)

Anyway, I appreciate your patience.  Multi-process tests are hard.
Attachment #653104 - Attachment is obsolete: true
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → WORKSFORME
Whiteboard: [orange]
Dumbly reopening.  Let me know where this log should have been posted

https://tbpl.mozilla.org/php/getParsedLog.php?id=17674312&tree=Mozilla-Aurora
Status: RESOLVED → REOPENED
Resolution: WORKSFORME → ---
Justin, this seems to have regressed in the last week - could you take a look? :-)
Flags: needinfo?(justin.lebar+bug)
(In reply to Ed Morley [UTC+0; email:edmorley@moco] from comment #56)
> Justin, this seems to have regressed in the last week - could you take a
> look? :-)

It's pretty hard to justify working on this when I have B2G blockers onhand.  If I get some cycles, I'll see what I can do.  This does look pretty bad...
Flags: needinfo?(justin.lebar+bug)
Test disabled on Linux for too many intermittent failures:
https://hg.mozilla.org/integration/mozilla-inbound/rev/b42bab7f7295
Whiteboard: [test disabled on Linux][leave open]
Now morphed to test_browserElement_oop_SetVisibleFrames.html
Summary: Intermittent test_browserElement_oop_SetVisible.html | finished in a non-clean fashion (in /tests/dom/browser-element/mochitest/test_browserElement_oop_SendEvent.html) → Intermittent test_browserElement_oop_SetVisible.html, test_browserElement_oop_SetVisibleFrames.html | finished in a non-clean fashion (in /tests/dom/browser-element/mochitest/test_browserElement_oop_SendEvent.html)
Disable test_browserElement_oop_SetVisibleFrames.html & test_browserElement_oop_SetVisibleFrames2.html on Linux as well:
https://hg.mozilla.org/integration/mozilla-inbound/rev/b0fc7f12a86c
And now test_browserElement_oop_Stop.html, sigh.
https://tbpl.mozilla.org/php/getParsedLog.php?id=18051186&tree=Mozilla-Inbound
Summary: Intermittent test_browserElement_oop_SetVisible.html, test_browserElement_oop_SetVisibleFrames.html | finished in a non-clean fashion (in /tests/dom/browser-element/mochitest/test_browserElement_oop_SendEvent.html) → Intermittent test_browserElement_oop_SetVisible.html, test_browserElement_oop_SetVisibleFrames.html, test_browserElement_oop_Stop.html | finished in a non-clean fashion (in /tests/dom/browser-element/mochitest/test_browserElement_oop_SendEvent.html)
Justin, can you take a look at this or else find an owner please :-)
Flags: needinfo?(justin.lebar+bug)
(In reply to Ed Morley [:edmorley UTC+0] from comment #217)
> Justin, can you take a look at this or else find an owner please :-)

Ed, I think I'd get hung by the B2G management if I took anyone's time away from shipping the v1 device to work on an intermittent orange.

We don't have someone capable who's not on the critical path, as far as I'm aware.  I'm sorry.
Flags: needinfo?(justin.lebar+bug)
Yeah it's an unfortunate predicament (and one I don't envy), but likewise that doesn't leave me with many options for now, sorry. Thank you for replying anyway :-)

test_browserElement_oop_SendEvent.html disabled on Linux (along with the previously disabled):
https://hg.mozilla.org/integration/mozilla-inbound/rev/60313ab86d67
Let's call this fixed and I'll open a new bug for re-enabling the test, if that works for you guys.
Status: REOPENED → RESOLVED
Closed: 12 years ago → 11 years ago
Resolution: --- → FIXED
I filed bug 848083 on figuring out what's going wrong here.
(In reply to Justin Lebar [:jlebar] from comment #222)
> Let's call this fixed and I'll open a new bug for re-enabling the test, if
> that works for you guys.

There are a couple of other tests that fail, just much less frequently - so we may need to reopen; but I'll leave closed for now.
Depends on: 848083
Whiteboard: [test disabled on Linux][leave open] → [test disabled on Linux]
Component: DOM: Mozilla Extensions → DOM
Component: DOM → DOM: Core & HTML
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: