codesearch: add go.mod

regexp: don't loop forever on >1MB lines

This can happen especially with cgrep, which isn't filtering
for likely-text inputs.

Matches will be chopped into 1MB chunks but oh well.

While we're here, make sure we print \n even for matches
at the end of a file without a trailing newline.

Fixes google#49.

codesearch: excise code.google.com