Update copas.settimeout

[AMD-NICK]

I used copas.settimeout after copas.wrap, but the socket was still being read endlessly until I printed the timeout table while copas is running and noticed that there were unwrapped sockets in the table

[Tieske]

Not sure about this. Typically the copas.xxx(skt, ...) functions must be called with a luasocket socket, not a copas one. Not just this one, but all of them.

I think, if anything, we should add assertions, and fail hard if called with the wrong socket type.

if skt is an already wrapped socket, one should call skt:settimeouts(a, b, c)

[AMD-NICK]

really, it's my mistake. I looked at the documentation sock:settimeout instead of copas.settimeout, and it says to execute after copas.wrap. I only just now noticed that I was looking at the wrong thing.

So assert could be useful for cases like mine

[Tieske]

closing this for now.