AdbMonitor may fail to track devices when many devices get online or offline at the same time.

[mythsman]

https://github.com/Genymobile/gnirehtet/blob/f5c0710c751f7ca82ce413058abc57d9af7208c3/relay-java/src/main/java/com/genymobile/gnirehtet/AdbMonitor.java#L152

Here I think an exception should be raised to reset the current state...

[rom1v]

Hmmm, it should not, it's called in a loop:

https://github.com/Genymobile/gnirehtet/blob/f5c0710c751f7ca82ce413058abc57d9af7208c3/relay-java/src/main/java/com/genymobile/gnirehtet/AdbMonitor.java#L90

:confused:

Here I think an exception should be raised to reset the current state...

It should not be necessary to reset the state, if there is a bug in parsing, it should be fixed without reset.

[mythsman]

Hmmm, it should not, it's called in a loop:

https://github.com/Genymobile/gnirehtet/blob/f5c0710c751f7ca82ce413058abc57d9af7208c3/relay-java/src/main/java/com/genymobile/gnirehtet/AdbMonitor.java#L90

😕

Here I think an exception should be raised to reset the current state...

It should not be necessary to reset the state, if there is a bug in parsing, it should be fixed without reset.

But that exception can be caught in an outer loop: https://github.com/Genymobile/gnirehtet/blob/f5c0710c751f7ca82ce413058abc57d9af7208c3/relay-java/src/main/java/com/genymobile/gnirehtet/AdbMonitor.java#L61

and then , the state can be reset correctly.

[mythsman]

But I actually encouter that bug ....

[mythsman]

In my case ,this loop while loops forever , and hit my CPU up to 100%... https://github.com/Genymobile/gnirehtet/blob/f5c0710c751f7ca82ce413058abc57d9af7208c3/relay-java/src/main/java/com/genymobile/gnirehtet/AdbMonitor.java#L90

[rom1v]

But that exception can be caught in an outer loop:

The exception should occur only when a communication with the adb socket fails (so it reconnects and "restart").

In my case ,this loop while loops forever , and hit my CPU up to 100%...

OK, so there's definitely a parsing bug. To be investigated.

Does it also happen with the rust version?

[rom1v]

On which version of gnirehtet (which SHA-1) do you reproduce? master? dev?

[mythsman]

reproduce The neweset v2.5

[mythsman]

But that exception can be caught in an outer loop:

The exception should occur only when a communication with the adb socket fails (so it reconnects and "restart").

In my case ,this loop while loops forever , and hit my CPU up to 100%...

OK, so there's definitely a parsing bug. To be investigated.

Does it also happen with the rust version?

I don't know , I haven't tried..