Why does Google's Closure Compiler leave a few unnecessary spaces or line breaks?

Question

I've noticed that every time I use Google's Closure Compiler Service, it leaves a few unnecessary spaces in the compiled code presented on the right-hand side of the page. These correspond to line breaks in the hosted version of the compiled code.

For example (note the line breaks, each of which seems unnecessary):

http://troy.onespot.com/static/stack_overflow/closure_spaces.js

To date, I've just been removing them manually, but I'm curious why they're there. Is it to limit the line length of the hosted version of the code to make it more readable? Could the compiler be smart enough to leave or insert those intentionally to maximize GZIP compression efforts?

I know that they have a trivial effect on the file size, but with so much effort going into minifying every last byte in the source script, it's counterintuitive why they're there.

Answer 1

Quoting the Closure Compiler FAQ:

Why are there random line feeds in compiled scripts?

The Closure Compiler intentionally adds line breaks every 500 characters or so. Firewalls and proxies sometimes corrupt or ignore large JavaScript files with very long lines. Adding line breaks every 500 characters prevents this problem. Removing the line breaks has no effect on a script's semantics. The impact on code size is small, and the Compiler optimizes line break placement so that the code size penalty is even smaller when files are gzipped.

You knew it was smart! :)