Don't split manually on linebreaks

[incase]

Use splitlines() instead of manually splitting on \r or \n. Use the first line that looks like a PID (all numbers, as re.match only looks at the start of the line and the regex only allows numbers until end of line anchor).

Fixes https://github.com/jupyter/terminado/issues/84

[blink1073]

I'll cut a patch release early next week.