Comment

Peter Bengtsson

Actually, a quick test:

$ uglifyjs < maxlength_countdown.js > /tmp/uglify.js
$ java -jar ../compiler.jar < maxlength_countdown.js > /tmp/closure.js
$ cd /tmp
$ emacs uglify.js # remove the preamble comment
$ gzip < uglify.js > uglify.js.gz
$ gzip < closure.js > closure.js.gz
$ ls -ltr
-rw-r--r-- 1 peterbe wheel 853 11 Jul 11:23 closure.js
-rw-r--r-- 1 peterbe wheel 857 11 Jul 11:23 uglify.js
-rw-r--r-- 1 peterbe wheel 408 11 Jul 11:24 closure.js.gz
-rw-r--r-- 1 peterbe wheel 405 11 Jul 11:24 uglify.js.gz

So, in this scenario UglifyJS is beaten by 4 bytes but when gzipped it beats Closure by 3 bytes.

Shit this is getting really nitpicky and complex at the same time.

Parent comment

Anthony Ricaud

Have you tried comparing the minified sizes after gzipping? This is the real metric because we send our code gzipped (right?). That might gives a different result