coffee-register


require coffeescript files "on-the-fly" without precompilation up to 2000% faster than the official coffee-script/register package.

Usage

index.js:
``javascript
require('coffee-register');

// That's it! After this call require coffee files as you would JS files.
require('./somefile.coffee')
require('./another') // ext is optional
require('./dir') // loads './dir/index.coffee'
`




Benchmarks


How is it so much faster?

There are two primary reasons:

a) This module attaches a hook in node's module system to be invoked only for files ending with .coffee as opposed to the official coffee-script/register which hooks onto .coffee, .litcoffee, .coffee.md. Each additional hook imposes overhead on module loading times and since the latter 2 extensions are rarely used they have been ommited by default (although they can be manually registered by using require('coffee-register').register(['.coffee', '.litcoffee', ...]) and any other extension you wish)

b) By leverging dynamic caching. When coffee-register encounters a coffee file it compiles it and then saves it to disk, mapping its content's hash to the saved compiled file so that the next time it encounters this coffee file it inspects its content's hash and attempts to load it from cache. The cache never has to be purged as the process is done automatically for you.

What about child_process forks?


Forks created by child_process will also work after this module is loaded.

Source maps

Due to how the official coffee-script package works source maps will only work by default for non-cached files (i.e. only on the first time they are loaded). coffee-register provides an optional workaround which can be enabled by setting the SOURCE_MAPS env variable to true.

Example:
`bash
SOURCE_MAPS=1 node index.js
`

Coffeescript v1 vs v2

This module can use both v1 and v2 for compiling.

For the lastest v1 compiler:
`bash
$ npm install coffee-register@1
`

For the lastest v2 compiler:
`bash
$ npm install coffee-register
``

License

MIT © Daniel Kalen