py - Pierre Yves Ricau

> next start position was: 40 + 16 + 9 (something else with 9 bytes) The thing about your case, is that it sounds like your heap dump has an...

Oh, actually, you just found a documentation bug. The version is meant to be interpolated in the doc. I believe we should still use single quotes but replace `${leakCanaryVersion}` with...

thanks!

Interesting, looks like we're failing to close? I wonder what could possibly have changed in 2.9

Seems like a fair ask. PRs welcome.

Interesting! So it looks like `mPreSortedChildren` is a field on ViewGroup. Can you provide the versions and manufacturers you've reproduced this on? Looking at the latest [ViewGroup sources](https://cs.android.com/android/platform/superproject/+/master:frameworks/base/core/java/android/view/ViewGroup.java), I can...

Note: I'm deleting comments that are not moving the conversation forward or helping. This bug is a bug in the Android OS, it was marked fixed by Google on April...

Thanks for the report! Actually what's going on here is that leak detection has a side effect (which you can see in that source link you shared): calling `retainedObjectsChecker.clearObjectsWatchedBeforeHeapDump()` which...

Sounds like a bug in either the Android SDK or a manufacturer impl, you should file an issue with a repro case, which means you need to figure out how...

The trace indicates that `ContentObserver$Transport.mContentObserver` isn't null, and it is set to null when `ContentObserver$Transport#releaseContentObserver()` is called, so that wasn't called, which means `ContentResolver.unregisterContentObserver()` wasn't called which means `TextClock.unregisterObserver()` wasn't...