Bridger Maxwell

Yeah, I wasn’t sure if this is better at the NIO level or the library level that uses NIO. I’ll make another issue to ask there. Sent with GitHawk

@tanner0101 The scheduleTask API is what I used to address this issue in https://github.com/vapor/websocket-kit/pull/68 Sent with GitHawk

@hactar I'd recommend using https://github.com/ianpartridge/swift-backtrace to get a better crash trace. Fwiw, my app has been running in production for a few months without crashing so I don't think these...

That theory makes sense to me! I just got another crash that seems related. All I got from the logs is this: `Fatal error: Trying to remove task, but it's...

Honestly, I'm not sure... I guess the dynamic framework is nice if you don't want to repeat the binary size in each extension, like an Apple Watch binary that also...

Thanks for digging into this! Sorry it is more complicated than I thought it would be. It would be awesome if CocoaPods allowed the app to choose whether it is...

I ran into trouble trying to add a PAT too. I didn't guess the right permissions, so adding the PAT caused the app to get errors every time it tried...

I'm also seeing this issue with Xcode 12 and multiple threads. I'm not sure if this is due to an Xcode change or a SQLite.swift change.

Thanks @ianpartridge. I'll follow up again next week. There may still be a concurrency issue, but I am seeing I was wrong about the buffer with incomplete data. WSSocketProcessor has...

I tried building with the GCD_ASYNCH compiler flag to used GCD instead of epoll. That seemed to help because I am not able to repro the issue anymore. The downside...