java.lang.NullPointerException: point must not be null.

[jruston]

I am seeing a small number of crashes coming from this library. I have seen several hundred occurrences over the past few months. The crash looks like the following:

Fatal Exception: java.lang.NullPointerException: point must not be null.
at com.google.android.gms.common.internal.Preconditions.checkNotNull(com.google.android.gms:play-services-basement@@18.9.0:2)
       at com.google.android.gms.maps.model.LatLngBounds.contains(com.google.android.gms:play-services-maps@@19.2.0:1)
       at com.google.maps.android.clustering.view.DefaultClusterRenderer$RenderTask.run(DefaultClusterRenderer.java:495)
       at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1167)
       at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:641)
       at java.lang.Thread.run(Thread.java:929)

I can't seem to replicate it myself, but it seems to affect mostly Samsung, Xiaomi and Vivo devices, 90% of which are on Android 10. I also see some occurrences on Android 11, 12 and 13.

I am using v3.19.1 of the library. It seems to occur in this version and also previous versions.

[reidconner1111]

We are also seeing about 500 of these per day in a high-user-volume application -- same stacktrace, same breakdown as the op, mostly Xiaomi, Samsung, and Vivo devices with some Hauwei, about 90% Android 10 with the rest on Android 11 and 12. We are still on v3.10.0, however.

[kikoso]

Hi @jruston ,

Would you be able to point out to the oldest version where this issue has been reported?

[jruston]

@kikoso The oldest version that I can find that has this issue is v3.11.0, but that's only because Crashlytics only allows me to see crashes that occurred in the past 90 days, and there is close to no usage of my app versions with an SDK older than v3.11.0. I'm guessing the issue goes back further than v3.11.0, as can be seen by @reidconner1111's comment where the issue is also present in v3.10.0.

[reidconner1111]

We have been on v3.10.0 for some time and don't have much visibility further back than that either, unfortunately, so I cannot say whether the issue was present before v3.10.0 -- it is definitely present in v3.10.0 though.