Why did you need to comment out this line? When I run it, it seems to run just fine with the assert present

I get the following error during the test discovery and when a test finishes. I believe it happens anytime the client would be created.

Personally, I don't understand why it happens, especially if it isn't happening on a non windows machine. For me, it happens every time even though the tests run to completion

Error executing callback:
...ta/lazy/neotest-vstest/lua/neotest-vstest/mtp/client.lua:36: ECONNRESET
stack traceback:
	[C]: in function 'assert'
	...ta/lazy/neotest-vstest/lua/neotest-vstest/mtp/client.lua:36: in function <...ta/lazy/neotest-vstest/lua/neotest-vstest/mtp/client.lua:35>
	[C]: in function 'resume'
	.../AppData/Local/nvim-data/lazy/nvim-nio/lua/nio/tasks.lua:118: in function 'cb'
	.../AppData/Local/nvim-data/lazy/nvim-nio/lua/nio/tasks.lua:183: in function <.../AppData/Local/nvim-data/lazy/nvim-nio/lua/nio/tasks.lua:182>
Error executing callback:
...ta/lazy/neotest-vstest/lua/neotest-vstest/mtp/client.lua:36: ECONNRESET
stack traceback:
	[C]: in function 'assert'
	...ta/lazy/neotest-vstest/lua/neotest-vstest/mtp/client.lua:36: in function <...ta/lazy/neotest-vstest/lua/neotest-vstest/mtp/client.lua:35>

Hmm, there might be some timing issue going on here, where the LSP kills the connection before the server does.
Might have to look into a different kind of error handling if the assert puts the plugin into a zombie state

I'll mess around until I no longer get that assert error unless you had another idea for the error handling

I've thought a little about this and I think removing the assert is fine – the error handling is not really working as intended for the MTP server at the moment regardless.

Ideally, any assertion errors in the MTP "proxy" process should trigger a shutdown of the LSP client but at this stage it just crashes the proxy process.

If you want to play around with the error handling it would be very welcomed but I am also fine with getting this merged if the assert is just replaced with a log statement so we can track when errors are thrown.

Looks like the error was important. The client is being killed prematurely and not all of the test discovery results are being captured

Looks like client:request_sync has a default timeout of 1s so even when ignoring the error I noticed all of my tests weren't always there. I changed this to 30s for now

I also got that assert changed to a logger.warn