-
Notifications
You must be signed in to change notification settings - Fork 4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[2661] Fix BroadcastReceiver ANR at ScreenOffTrigger #2732
base: main
Are you sure you want to change the base?
[2661] Fix BroadcastReceiver ANR at ScreenOffTrigger #2732
Conversation
analysisClient.newJob(JobContext(ScreenOffTrigger::class)) | ||
currentJob = job | ||
analysisExecutor.execute { | ||
analysisExecutor.execute { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If analysisExecutor.execute
runs the task quickly enough, there's a chance that we might dump the heap & freeze the VM before the process has had a chance to report back to system_server
, right?
Looks to me like if we want to avoid that, we need to execute with a delay instead"?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good point. I just updated it with a delayed executor. So far seems to be working as well and ANR is no longer reproducible internally
Add a delayedScheduledExecutorService
Adding delay Add a delayedScheduledExecutorService Fix BroadcastReceiver ANR
…om/FranAguilera/leakcanary into franjam/fix_broadcast_receiver_anr
Introducing AnalysisJobHandler to share Job execution logic between triggers
leakcanary/leakcanary-android-release/src/main/java/leakcanary/ScreenOffTrigger.kt
Outdated
Show resolved
Hide resolved
leakcanary/leakcanary-android-release/src/main/java/leakcanary/ScreenOffTrigger.kt
Outdated
Show resolved
Hide resolved
leakcanary/leakcanary-android-release/src/main/java/leakcanary/ScreenOffTrigger.kt
Outdated
Show resolved
Hide resolved
currentJob = null | ||
analysisCallback(result) | ||
if(currentJob.compareAndSet(null, job)){ | ||
delayedScheduledExecutorService.schedule { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Couldn't this have been a simple Handler(Looper.getMainLooper()).postDelayed({}, 500)
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
(you can cancel runnables on handler just fine)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good point, simplified to use Handler
leakcanary/leakcanary-android-release/src/main/java/leakcanary/ScreenOffTrigger.kt
Outdated
Show resolved
Hide resolved
* | ||
* If not specified, the initial delay is 500 ms | ||
*/ | ||
private val analysisExecutorDelayMillis: Long = INITIAL_EXECUTOR_DELAY_MILLIS |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
API nit: take in a Duration
instead, rename the field to analysisExecutorDelay
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
good callout, updated now
leakcanary/leakcanary-android-release/src/main/java/leakcanary/ScreenOffTrigger.kt
Outdated
Show resolved
Hide resolved
- INITIAL_EXECUTOR_DELAY_IN_MILLI -> INITIAL_EXECUTOR_DELAY_MILLIS - Remove accidental change on viz of analysisExecutor - Rely on Handler.postDelayed instead of DelayedScheduledExecutorService
- Use Duration type for analysisExecutorDelay - Rename to submitAnalysisToExecutor
Adding delay Add a delayedScheduledExecutorService Fix BroadcastReceiver ANR Fix typo PR feedback Add a delayedScheduledExecutorService AnalysisJobHandler Introducing AnalysisJobHandler to share Job execution logic between triggers Reduce scope of changes and go back to original approach of DelayedScheduledExecutorService Revert unwanted changes at docs Update formatting of leak-canary-for-releases.md Revert unwanted indentation Rename shouldStartAnalysis to shouldScheduleAnalysis Rename cancelScheduledAction to cancelScheduledAnalysis
2668668
to
d55327c
Compare
…om/FranAguilera/leakcanary into franjam/fix_broadcast_receiver_anr
385e110
to
ac0e425
Compare
c2d8541
to
6146c35
Compare
a9147ce
to
ba4b3f6
Compare
ba4b3f6
to
51e25f9
Compare
Resolves #2661
Issue
BroadcastReceiver#onReceive runs by default on the main thread. If onReceive doesn't complete in a timely manner, the system will report an Application Not Responding
We've seen instances of this occurring frequently on internal builds
Solution
NOTE: Kudos to @pyricau for the guidance on adding a delay.
Let the broadcast receiver complete immediately and schedule the executor with an initial delay to avoid issues with the broadcast receiver not completing
Verification
Verified that the the ANR no longer occurs and reporting still completes without issues on one of our main apps (internal release build)
This was heavily tested on internal builds