Crash in Lightstreamer iOS Client (3.0.1)

[eric280]

We recently upgraded our LS iOS client from 2.1.2 to 3.0.1. We are now seeing this crash report:

EXC_BAD_ACCESS KERN_INVALID_ADDRESS 0x0000000000000000

Crashed: Thread
0 CoreFoundation 0x1972574c8 __CFStringCreateImmutableFunnel3 + 1104
1 CoreFoundation 0x197257d48 CFStringCreateWithBytes + 48
2 Lightstreamer_iOS_Client 0x106213a84 -[LSHTTPConnection dispatchOperation:didReceiveData:] + 328
3 Lightstreamer_iOS_Client 0x10621f60c __45-[LSURLDispatchOperation taskDidReceiveData:]_block_invoke + 60
4 Lightstreamer_iOS_Client 0x1061d6f3c -[LSThreadPoolThread main] + 836
5 Foundation 0x1976b3cd0 __NSThread__start__ + 864
6 libsystem_pthread.dylib 0x196fc18fc _pthread_start + 168
7 libsystem_pthread.dylib 0x196fc99d4 thread_start + 8

Any insight or suggestions would be appreciated!

[DarioCrivelli]

We had no report of a similar issue on that version, so the case should be investigated in the details.
However, to do this, you should upgrade further to the latest version of the library.
You can see in the changelog that you can expect very few compatibility issues.
However, this requires Server version 7.0 at least. If you are using Server version 6.1, note that it is no longer supported as well.
Is it possible for you to follow this path?

[eric280]

Thank you, DarioCrivelli. Yes, we plan to upgrade the Server to 7.0, but I have no influence on the scheduling for that upgrade. Until that time, I am stuck on the 3.0.1 version of the iOS SDK.

If the crash above was occurring in our own code, like in one of our LSClientDelegate methods, I would expect to see symbols from our app on the stack. Since they aren't present, I assume the crash is occurring within the Lightstreamer_iOS_Client binary.

[DarioCrivelli]

After a second check, we are now confident that the issue is the same that was fixed in version 4.1.2 of the library.
Hence, the upgrade to at least 4.1.2 is the way to proceed. Unfortunately, I confirm that this version is not compatible with Server 6.1.
On the other hand, a backporting of the fix to 3.0.1 is not planned, since the version is no longer supported.

Sorry for the inconvenience. Do you see the issue frequently? We expect it to occur rarely.

[eric280]

Thanks! At least it saves us time trying to track down an issue in our own code. I have passed this information on to our server team, and we are attempting to escalate our upgrade schedule.

The issue is infrequent, however, it is our top crash. I do not have specific metrics at the moment, but they do not appear urgent enough to warrant a client downgrade, yet.