Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>

This will create easily solved breakage on updates but
I figure it's best to pull the band aid off as early as possible

Far simpler now. Sorry this took a while.

Closes #427
Closes #429
Closes #434
Closes #436
Closes #437

Closes #330

Closes #330

Closes #405

You should always be reading from the connection with CloseRead so this shouldn't
have affected anyone using the library correctly.

Closes #298
Closes #394

The close frame was being received from the peer before we were able to reset our
write timeout and so we thought the write kept failing but it never was...

Thanks @univerio and @bhallionOhbibi

Updates e9d08816

Closes #354

Also ran gofmt on everything. Thanks again @paralin. #334

Co-authored-by: Christian Stewart <christian@paral.in>

Closes #248

Luckily, due to the 5s timeout on the close handshake, this would have
had very minimal effects on anyone in production.

Closes #220 and #230

Closes #231

Closes #215

Closes #213

Will prevent deadlock if a writer is used after close.

Closes #191

Much simpler config. Github Actions just isn't there yet.

See https://github.com/nhooyr/websocket/issues/182

Was noop before.

Closes #182