The goal of this project is to provide a fast, stable and functional wrapper for a fast C++ regular expression library re2.
Erlang-re2 is implemented as a NIF module.
Erlang-re2 is at the very early stage of development, and is more like a proof of concent at the moment.
It's not recommended for general use, but the early results look very promising.
When compared to a standard Erlang re library, re2 is a bit faster on simple regular expressions like:
"^h.*o$"
2-6 times faster on more complex regular expressions like:
"^((a*)|(b*))((c*)|(d*))$"
And 35-640 times faster on monster regular expressions like this (by the way, it's a regular expression for urls):
"^(http|https|ftp)\://([a-zA-Z0-9\.\-]+(\:[a-zA-Z0-9\.&%\$\-"
"]+)*@)*((25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|"
"[1-9])\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|"
"[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|"
"[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|"
"[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[0-9])|localhost|"
"([a-zA-Z0-9\-]+\.)*[a-zA-Z0-9\-]+\.(com|edu|gov|int|mil|net|org"
"|biz|arpa|info|name|pro|aero|coop|museum|[a-zA-Z]{2}))"
"(\:[0-9]+)*(/($|[a-zA-Z0-9\.\,\?\'\\\+&%\$#\=~_\-]+))*$"
It is especially much faster if a regular expression doesn't match an input string.
Erlang-re2 has been tested on Erlang R15B03-1, but I suppose it should run wihtout problems on R16 as well.
re2 uses rebar for compilation and test running:
$ ./rebar compile $ ./rebar ct
To run performance tests:
$ ./rebar compile shell 1> application:start(re2). 2> re2_perf:run_tests().