Closed Bug 1174872 Opened 9 years ago Closed 9 years ago

Intermittent test_fetch_basic_http_sw_reroute.html | Test timed out

Categories

(Core :: DOM: Service Workers, defect)

Product:
Core
Component:
DOM: Service Workers
Platform:
x86_64
macOS
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla43
Tracking Flags:
Tracking Status
firefox41 --- affected
firefox43 --- fixed

People

(Reporter: RyanVM, Unassigned)

References

Details

(Keywords: intermittent-failure)

Attachments

(1 file)

Call FailWithNetworkError if nsIThreadRetargetableRequest::RetargetDeliveryTo() fails
1.04 KB, patch
bkelly
: feedback-
Details | Diff | Splinter Review 
09:19:50 INFO - === hello pass
09:19:50 INFO - : response should match
09:19:50 INFO - 3892 INFO TEST-PASS | dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | Worker: 200 === 200: status should match
09:19:50 INFO - 3893 INFO TEST-PASS | dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | Worker: {"title":"aBook","author":"john"} === {"title":"aBook","author":"john"}: json response should match
09:19:50 INFO - 3894 INFO TEST-PASS | dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | Worker: 200 === 200: wrong status
09:19:50 INFO - 3895 INFO TEST-PASS | dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | Worker: true: expected json parse failure
09:19:50 INFO - 3896 INFO TEST-PASS | dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | Worker: 200: 200
09:19:50 INFO - 3897 INFO TEST-PASS | dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | Worker: 65536: 65536
09:19:50 INFO - 3898 INFO TEST-PASS | dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | Worker: 65536 === 65536: wrong value at offset 65536
09:19:50 INFO - 3899 INFO TEST-UNEXPECTED-FAIL | dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | Test timed out. - expected PASS
09:19:50 INFO - MEMORY STAT vsize after test: 4081725440
09:19:50 INFO - MEMORY STAT residentFast after test: 582930432
09:19:50 INFO - MEMORY STAT heapAllocated after test: 89964136
09:19:50 INFO - 3900 INFO TEST-OK | dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | took 316269ms
Ben, can you please take a look at this please? :)
Component: DOM → DOM: Service Workers
Flags: needinfo?(bkelly)
Probably won't get to this for a week or so.
Assignee: nobody → bkelly
Status: NEW → ASSIGNED
Flags: needinfo?(bkelly)
This needs to be fixed, but I'm not actually working on it currently.
Assignee: bkelly → nobody
Blocks: ServiceWorkers-v1
Status: ASSIGNED → NEW
This seems to only be effecting mulet now.
I think this may be a manifestation of bug 1185150.
I think we should disable this test on Mulet and move on with our lives.
Assignee: nobody → ehsan
Status: NEW → ASSIGNED
Keywords: leave-open
Whiteboard: [test disabled on Mulet]
Keywords: leave-open
(In reply to Treeherder Robot from comment #640)
> log:
> https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-
> inbound&job_id=14045425
> repository: mozilla-inbound
> start_time: 2015-09-14T04:44:18
> who: tomcat[at]mozilla[dot]com
> machine: t-snow-r4-0084
> buildname: Rev4 MacOSX Snow Leopard 10.6 mozilla-inbound debug test
> mochitest-3
> revision: a0ce99dca834
> 
> 3878 INFO TEST-UNEXPECTED-FAIL |
> dom/tests/mochitest/fetch/test_fetch_basic_http_sw_reroute.html | Test timed
> out.
> Assertion failure: mResponseAvailableCalled, at
> /builds/slave/m-in-m64-d-0000000000000000000/build/src/dom/fetch/FetchDriver.
> cpp:62

Hmm, this one seems to be hitting an assertion that you added, Nikhil.  I think this means that we are not calling FailWithNetworkError() where we should.  Looking at the code, there is for example <http://hg.mozilla.org/mozilla-central/annotate/9ed17db42e3e/dom/fetch/FetchDriver.cpp#l854> where we don't do this, and indeed, you can find the following in the log:

05:12:05     INFO -  [Parent 1052] WARNING: 'NS_FAILED(rr->RetargetDeliveryTo(sts))', file /builds/slave/m-in-m64-d-0000000000000000000/build/src/dom/fetch/FetchDriver.cpp, line 855
05:12:05     INFO -  [Parent 1052] WARNING: 'NS_FAILED(rr->RetargetDeliveryTo(sts))', file /builds/slave/m-in-m64-d-0000000000000000000/build/src/dom/fetch/FetchDriver.cpp, line 855
Comment on attachment 8660725 [details] [diff] [review]
Call FailWithNetworkError if nsIThreadRetargetableRequest::RetargetDeliveryTo() fails

Review of attachment 8660725 [details] [diff] [review]:
-----------------------------------------------------------------

Failing retargeting is not a fatal error.  It will run just fine on the main thread.  (I asked about doing this very patch before.)
Attachment #8660725 - Flags: feedback-
(In reply to Ben Kelly [:bkelly] from comment #643)
> Comment on attachment 8660725 [details] [diff] [review]
> Call FailWithNetworkError if
> nsIThreadRetargetableRequest::RetargetDeliveryTo() fails
> 
> Review of attachment 8660725 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> Failing retargeting is not a fatal error.  It will run just fine on the main
> thread.  (I asked about doing this very patch before.)

Does that mean we need to adjust the assertion in ~FetchDriver()?
I take it back.  It doesn't look like bug 1198544.

But this is unrelated to the retargeting because retargeting is optional.

I wonder if we're not holding the channel alive during some async operation and its getting released too soon under certain race conditions.
No longer depends on: 1198544
The channel is getting released by the cycle collector.
Stealing.

This is a regression from bug 119049.  nsCorsListenerProxy is not calling OnStartRequest() when it fails here:

  https://dxr.mozilla.org/mozilla-central/source/netwerk/protocol/http/nsCORSListenerProxy.cpp#519
Assignee: ehsan → bkelly
Err... not sure its this bug.  This bug is for test_fetch_basic_http_sw_reroute.html.  The failures mentioned in comment 641 and in bug 1203680 are for the cors tests.  I'm going to leave this bug open and fix the CORS issues in bug 1203680.
Assignee: bkelly → nobody
Status: ASSIGNED → NEW
This try suggests the mulet failure has been fixed:

  https://treeherder.mozilla.org/#/jobs?repo=try&revision=ed185694d5cf

The orange is all in test_sharedWorker.html which is unrelated to this bug.

I also checked b2g in that try, but it appears these tests are still busted there.  Leaving disabled.
Whiteboard: [test disabled on Mulet]
https://hg.mozilla.org/mozilla-central/rev/ddb002f98a4a
Status: NEW → RESOLVED
Closed: 9 years ago
status-firefox43: --- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla43
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: