If you want your production JavaScript looking sharp, then use js-beautify. Are you more concerned with going fast than looking pretty? Then jsanic is the tool for you.
Check out this overly contrived example:
> for i in {1..100000}; do echo "a = {"; done | /usr/bin/time -f 'seconds: %e mem: %M kb' ./jsanic -i >/dev/null
seconds: 0.91 mem: 3164 kb
> for i in {1..100000}; do echo "a = {"; done | /usr/bin/time -f 'seconds: %e mem: %M kb' js-beautify >/dev/null
Command terminated by signal 9
seconds: 13.70 mem: 2986796 kbWhile testing a website, I ran into a 5.6 MB JavaScript file. My system is not weak, it has an Intel i7 and 16G of ram and I run a minimal desktop. Yet every beautifier I threw at this file crashed after prolonged execution. Eventually js-beautify (a great tool) got it after I closed all other resource hogs (Firefox, Java, Chrome). It took a long time and there were still many lines that were thousands of characters long.
I want something better and I want to learn about compilers anyway. May as well write my own parser.
Once something is used, it is freed. The scanner (cache.c) will only read
from the file once, and won't store any more of the file in memory than is
necessary. The lexer (tokenizer.c) will build a thread safe double linked
list of tokens. A separate thread can then pop from the tokens as they become
available.
Eventually I want to create something of an AST so that I can replace minified variables with unique, memorable nouns according to scope.