New issue
Advanced search Search tips

Issue 891419 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Oct 24
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Bug
Flaky-Test: org.chromium.chrome.browser.feed.FeedAppLifecycleTest#activity_state_changes_increment_state_counters



Sign in to add a comment

org.chromium.chrome.browser.feed.FeedAppLifecycleTest#activity_state_changes_increment_state_counters is flaky

Project Member Reported by Findit, Oct 2

Issue description

Components: UI>Browser>ContentSuggestions>Feed
Owner: pnoland@chromium.org
Status: Assigned (was: Untriaged)
Patrick - PTAL. It seems that all of the FeedAppLifecycleTest tests are very flaky across the various android builders:

https://test-results.appspot.com/dashboards/flakiness_dashboard.html#testType=chrome_public_test_apk&tests=FeedAppLifecycleTest

In particular, this is blocking the CQ/tryjobs for android-kitkat-arm-rel:

https://test-results.appspot.com/dashboards/flakiness_dashboard.html#testType=chrome_public_test_apk%20(with%20patch)&tests=FeedAppLifecycleTest

They appear to be flaky-fail:

java.lang.AssertionError: The view is dirty.
  	at org.junit.Assert.fail(Assert.java:88)
  	at org.junit.Assert.assertTrue(Assert.java:41)
  	at org.chromium.content_public.browser.test.util.CriteriaHelper.pollInstrumentationThread(CriteriaHelper.java:92)
  	at org.chromium.content_public.browser.test.util.CriteriaHelper.pollUiThread(CriteriaHelper.java:171)
  	at org.chromium.content_public.browser.test.util.CriteriaHelper.pollUiThread(CriteriaHelper.java:193)
  	at org.chromium.chrome.test.util.ViewUtils.waitForStableView(ViewUtils.java:122)

I'm trying to figure out the impact of this flake atm, I may have to disable the test if it appears too high. Thanks!
I think the stack trace in  #1 is actually for a different flaky test; the activity_state_changes_increment_state_counters failures I saw were native crashes. I will continue to investigate, but if this is having a major impact on the CQ, go ahead and disable it.

Comment 3 Deleted

Project Member

Comment 4 by bugdroid1@chromium.org, Oct 3

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d64662024b6f3aad2b66ec0d0e9658bcb1414485

commit d64662024b6f3aad2b66ec0d0e9658bcb1414485
Author: Stephen McGruer <smcgruer@chromium.org>
Date: Wed Oct 03 19:28:43 2018

[Sheriff] Mark FeedAppLifecycleTest tests as flaky

These tests are failing flaky with crashes.

TBR=pnoland@chromium.org

Bug:  891419 
Change-Id: I99d699e29e3be1f0d167531fe702011bbf2c3c9e
Reviewed-on: https://chromium-review.googlesource.com/c/1259422
Commit-Queue: Stephen McGruer <smcgruer@chromium.org>
Reviewed-by: Stephen McGruer <smcgruer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#596320}
[modify] https://crrev.com/d64662024b6f3aad2b66ec0d0e9658bcb1414485/chrome/android/javatests/src/org/chromium/chrome/browser/feed/FeedAppLifecycleTest.java

Labels: -Sheriff-Chromium
Taking off the sheriff queue now that the test is marked flaky.
I was able to repro once I turned dcheck_always_on to true in my local release build. The crash seems to be triggered by a DLOG(FATAL) check when failing to create the history database. 

Labels: -sheriff-android
Labels: O-Jardin-M3 zine-triaged
Project Member

Comment 9 by bugdroid1@chromium.org, Oct 10

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/3c9859c2d86977f7f69224018d0a1cd8f6fe523c

commit 3c9859c2d86977f7f69224018d0a1cd8f6fe523c
Author: Patrick Noland <pnoland@google.com>
Date: Wed Oct 10 19:05:24 2018

Fix FeedAppLifecycleTest flakiness

(I speculate that) having both a ChromeBrowserTestRule and a
ChromeTabbedActivityTestRule creates a race in native initialization.
In our case this race can result in I/O related failures, which first
cause a fatal failure when initializing various history dbs.
We avoid this by removing the need for ChromeBrowserTestRule, taking
care to set up FeedProcessScope before native is initialized.

Bug:  891419 
Change-Id: I65b09a60339be7475abbb0550e2ed162f9cb4075
Reviewed-on: https://chromium-review.googlesource.com/c/1263430
Commit-Queue: Patrick Noland <pnoland@chromium.org>
Reviewed-by: Filip Gorski <fgorski@chromium.org>
Cr-Commit-Position: refs/heads/master@{#598430}
[modify] https://crrev.com/3c9859c2d86977f7f69224018d0a1cd8f6fe523c/chrome/android/feed/core/java/src/org/chromium/chrome/browser/feed/FeedAppLifecycle.java
[modify] https://crrev.com/3c9859c2d86977f7f69224018d0a1cd8f6fe523c/chrome/android/feed/core/java/src/org/chromium/chrome/browser/feed/FeedProcessScopeFactory.java
[modify] https://crrev.com/3c9859c2d86977f7f69224018d0a1cd8f6fe523c/chrome/android/javatests/src/org/chromium/chrome/browser/feed/FeedAppLifecycleTest.java

Project Member

Comment 10 by bugdroid1@chromium.org, Oct 10

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/9966f68f3d92d728f4a3397d6f22e1c5281c4b82

commit 9966f68f3d92d728f4a3397d6f22e1c5281c4b82
Author: Filip Gorski <fgorski@chromium.org>
Date: Wed Oct 10 23:51:46 2018

Revert "Fix FeedAppLifecycleTest flakiness"

This reverts commit 3c9859c2d86977f7f69224018d0a1cd8f6fe523c.

Reason for revert: <INSERT REASONING HERE>

Original change's description:
> Fix FeedAppLifecycleTest flakiness
> 
> (I speculate that) having both a ChromeBrowserTestRule and a
> ChromeTabbedActivityTestRule creates a race in native initialization.
> In our case this race can result in I/O related failures, which first
> cause a fatal failure when initializing various history dbs.
> We avoid this by removing the need for ChromeBrowserTestRule, taking
> care to set up FeedProcessScope before native is initialized.
> 
> Bug:  891419 
> Change-Id: I65b09a60339be7475abbb0550e2ed162f9cb4075
> Reviewed-on: https://chromium-review.googlesource.com/c/1263430
> Commit-Queue: Patrick Noland <pnoland@chromium.org>
> Reviewed-by: Filip Gorski <fgorski@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#598430}

TBR=fgorski@chromium.org,twellington@chromium.org,pnoland@chromium.org

Change-Id: Id092242604232d9785d48fd1c7352be0a232ef4e
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug:  891419 
Reviewed-on: https://chromium-review.googlesource.com/c/1274898
Reviewed-by: Filip Gorski <fgorski@chromium.org>
Commit-Queue: Filip Gorski <fgorski@chromium.org>
Cr-Commit-Position: refs/heads/master@{#598576}
[modify] https://crrev.com/9966f68f3d92d728f4a3397d6f22e1c5281c4b82/chrome/android/feed/core/java/src/org/chromium/chrome/browser/feed/FeedAppLifecycle.java
[modify] https://crrev.com/9966f68f3d92d728f4a3397d6f22e1c5281c4b82/chrome/android/feed/core/java/src/org/chromium/chrome/browser/feed/FeedProcessScopeFactory.java
[modify] https://crrev.com/9966f68f3d92d728f4a3397d6f22e1c5281c4b82/chrome/android/javatests/src/org/chromium/chrome/browser/feed/FeedAppLifecycleTest.java

Labels: -Pri-1 Pri-3
Is this fixed?

Marking as P3 since we wouldn't block launch on this.
Status: Fixed (was: Assigned)
I believe this is fixed. It hasn't flaked since the change in #11 was submitted.
Cc: s...@chromium.org
Looks like this missed the last Dev cut 

https://chromiumdash.appspot.com/commit/750261de4e50c04b2def3d6a37dabe9a98c3dded

Sign in to add a comment