Catch ThreadInterruptedException in Server.cs when closing stream/client in SendLoop

[darrencperry]

These were getting through and logging as errors when client disconnects due to socket exceptions, trying to match functionality with ThreadAbortException.

It's possible this isn't the only scenario where a ThreadInterruptedException is thrown though? Maybe we do still want to log it?

[miwarnec]

@darrencperry is there an easy way to reproduce it?

[miwarnec]

and what do you mean with 'it still gets through to here when aborting'? - aborting what exactly?

[darrencperry]

It's not easy to reproduce without a temperamental network.

That line you quoted is just copied from the above exception comment, I'm just trying to replicate the functionality of that error but I'm not 100% on whether a thread interrupted exception should be treated the same way as thread aborted exception.

I can't do much more on this in the near future as our project moved over to Ignorance, sorry!

[miwarnec]

sorry, didn't have time for telepathy in a while.
can you confirm that the issue still exists in the latest version?
did a major rewrite in 2021 :)

[darrencperry]

sorry, I'm not working with Telepathy or Mirror any more so I won't be able to do any more work on this, this pr solved an issue I was having at the time but feel free to close it, thanks!

[evandroabukamel]

Hey, guys. I'd like to revive this topic. In my game we are just updated and old version of Mirror to v93.0.1 and we still face this problem. The image bellow is the count of those logs in a separate production environment. It's not much but it happens in very specific case and we're testing only with ~6,000 players, we have around 100,000 in total, so this log tend to increase if we release this update globally.

Today I applied this change in our code and started to test in a separate environment. Let's wait.

[GarnicaJR]

Hey, guys. I'd like to revive this topic. In my game we are just updated and old version of Mirror to v93.0.1 and we still face this problem. The image bellow is the count of those logs in a separate production environment. It's not much but it happens in very specific case and we're testing only with ~6,000 players, we have around 100,000 in total, so this log tend to increase if we release this update globally.

Today I applied this change in our code and started to test in a separate environment. Let's wait.

Did you get any improvement?

[evandroabukamel]

No, we ended up switching to KCP, it was performing better and was not causing those exceptions.