Crash in com.apollographql.websocket queue

[aaronbarsky]

Summary

Previously discussed in #3390
It was hoped that the 1.16.1 patch would solve the issue, but the crash is still happening.

There are several flavours of the crash:

 Crashed: com.apollographql.websocket
0  CoreFoundation                 0xc70c CF_IS_OBJC + 76
1  CoreFoundation                 0x8ac8 CFRelease + 60
2  CoreFoundation                 0xdb368 _signalEventSync + 252
3  CoreFoundation                 0xd9568 ___signalEventQueue_block_invoke + 28
4  libdispatch.dylib              0x2248 _dispatch_call_block_and_release + 32
5  libdispatch.dylib              0x3fa8 _dispatch_client_callout + 20
6  libdispatch.dylib              0xb5cc _dispatch_lane_serial_drain + 768
7  libdispatch.dylib              0xc124 _dispatch_lane_invoke + 380
8  libdispatch.dylib              0x1738c _dispatch_root_queue_drain_deferred_wlh + 288
9  libdispatch.dylib              0x16bd8 _dispatch_workloop_worker_thread + 540
10 libsystem_pthread.dylib        0x3680 _pthread_wqthread + 288
11 libsystem_pthread.dylib        0x1474 start_wqthread + 8

          Crashed: com.apollographql.websocket
0  libobjc.A.dylib                0x3c20 objc_msgSend + 32
1  CoreFoundation                 0xdbd14 _outputStreamCallbackFunc + 36
2  CoreFoundation                 0xdb344 _signalEventSync + 216
3  CoreFoundation                 0xd9568 ___signalEventQueue_block_invoke + 28
4  libdispatch.dylib              0x2248 _dispatch_call_block_and_release + 32
5  libdispatch.dylib              0x3fa8 _dispatch_client_callout + 20
6  libdispatch.dylib              0xb5cc _dispatch_lane_serial_drain + 768
7  libdispatch.dylib              0xc124 _dispatch_lane_invoke + 380
8  libdispatch.dylib              0x1738c _dispatch_root_queue_drain_deferred_wlh + 288
9  libdispatch.dylib              0x16bd8 _dispatch_workloop_worker_thread + 540
10 libsystem_pthread.dylib        0x3680 _pthread_wqthread + 288
11 libsystem_pthread.dylib        0x1474 start_wqthread + 8

          Crashed: com.apollographql.websocket
0  libobjc.A.dylib                0x1c5c objc_retain_x0 + 16
1  libobjc.A.dylib                0x1c5c objc_retain + 16
2  CoreFoundation                 0xdb2c8 _signalEventSync + 92
3  CoreFoundation                 0xd9568 ___signalEventQueue_block_invoke + 28
4  libdispatch.dylib              0x2248 _dispatch_call_block_and_release + 32
5  libdispatch.dylib              0x3fa8 _dispatch_client_callout + 20
6  libdispatch.dylib              0xb5cc _dispatch_lane_serial_drain + 768
7  libdispatch.dylib              0xc124 _dispatch_lane_invoke + 380
8  libdispatch.dylib              0x1738c _dispatch_root_queue_drain_deferred_wlh + 288
9  libdispatch.dylib              0x16bd8 _dispatch_workloop_worker_thread + 540
10 libsystem_pthread.dylib        0x3680 _pthread_wqthread + 288
11 libsystem_pthread.dylib        0x1474 start_wqthread + 8

 Crashed: com.apollographql.websocket
0  libobjc.A.dylib                0x2354 objc_release_x0 + 16
1  libobjc.A.dylib                0x2354 objc_release + 16
2  CoreFoundation                 0xdb368 _signalEventSync + 252
3  CoreFoundation                 0xd9568 ___signalEventQueue_block_invoke + 28
4  libdispatch.dylib              0x2248 _dispatch_call_block_and_release + 32
5  libdispatch.dylib              0x3fa8 _dispatch_client_callout + 20
6  libdispatch.dylib              0xb5cc _dispatch_lane_serial_drain + 768
7  libdispatch.dylib              0xc124 _dispatch_lane_invoke + 380
8  libdispatch.dylib              0x1738c _dispatch_root_queue_drain_deferred_wlh + 288
9  libdispatch.dylib              0x16bd8 _dispatch_workloop_worker_thread + 540
10 libsystem_pthread.dylib        0x3680 _pthread_wqthread + 288
11 libsystem_pthread.dylib        0x1474 start_wqthread + 8       

          Crashed: com.apollographql.websocket
0  libobjc.A.dylib                0xdce8 objc_loadWeakRetained + 156
1  CoreFoundation                 0xdbd14 _outputStreamCallbackFunc + 36
2  CoreFoundation                 0xdb344 _signalEventSync + 216
3  CoreFoundation                 0xd9568 ___signalEventQueue_block_invoke + 28
4  libdispatch.dylib              0x2248 _dispatch_call_block_and_release + 32
5  libdispatch.dylib              0x3fa8 _dispatch_client_callout + 20
6  libdispatch.dylib              0xb5cc _dispatch_lane_serial_drain + 768
7  libdispatch.dylib              0xc124 _dispatch_lane_invoke + 380
8  libdispatch.dylib              0x1738c _dispatch_root_queue_drain_deferred_wlh + 288
9  libdispatch.dylib              0x16bd8 _dispatch_workloop_worker_thread + 540
10 libsystem_pthread.dylib        0x3680 _pthread_wqthread + 288
11 libsystem_pthread.dylib        0x1474 start_wqthread + 8 

Sadly they all start from the system level and have no Apollo specific code.

I am aware of the challenges in fixing this non reproducible bug. I'm just re-opening the issue in case it motivates time towards WebSockets 2.0.

Version

1.16.1

Steps to reproduce the behavior

Alas, no reproduction steps.

Logs

Anything else?

No response

[calvincestari]

@aaronbarsky - Release 1.18.0 also contains websocket fixes, it's definitely worth trying that.

[aaronbarsky]

That's great news! I'll integrate 1.18.0 into the next release and let you know how it goes.

[vladdorfman]

We experience the same crashes 1.18.0 doesn't really improve the situation.

[AnthonyMDev]

We've recently released 1.21.0 with a few more fixes for web sockets. Can users who are experiencing this please verify if this is still an issue with that version?

We're going to be re-writing the entire web socket layer for 2.0 in the coming months. We've spent a lot of time chasing down race conditions in the existing web socket layer in the past few months, but in order to completely remove these race conditions and dead locks, a rewrite is really necessary.

If this isn't resolved in 1.21.0, we likely won't be investing much more time into tracking these down and instead will focus our efforts on the rewrite. But if you'd like to take a crack at a PR to address this, we're happy to take a look!

[github-actions]

Do you have any feedback for the maintainers? Please tell us by taking a one-minute survey. Your responses will help us understand Apollo iOS usage and allow us to serve you better.