-
Notifications
You must be signed in to change notification settings - Fork 0
Optimizing concatenating files for curl.js
Even though curl.js is ultra-fast at asynchronously downloading modules and resources in parallel, it still works best when those modules and resources are bundled together into a small set of files. (Whether "small set" means "one file" or "half a dozen" depends entirely on your application and your environment.)
curl.js has a companion project, cram.js. At its core, cram.js is a simple dependency walker. You feed it a "root module" (which could be your app's "main" module) and it walks the dependency tree, collects all modules and resources, and concatenates them into one file.
cram.js is still under development. We use it very successfully in a few projects already, but it is still at least one rev away from being production-worthy, imho. In the mean time, please consider using James Burke's r.js (see below).
cram.js only works with Rhino at the moment. However, because of its DI design and use of feature testing, it will be quite easy to make it work with Node, Ringo, and Narwhal, as well as with modern browsers (perform build from a url), wscript.exe on Windows, and JavaScriptCore on Mac OS X (in conjunction with a shell script).
If you want to start experimenting with cram.js now, please use the dev branch here.
curl.js can load optimized files built with r.js, the optimizer for RequireJS. Please note, however, that the AMD spec is still not solidified with regards to optimizing plugin-based resources. cram.js and r.js use different approaches when working with plugins. This means that you must use RequireJS's plugins -- not curl.js's plugins -- when using r.js.