For the record, i've nothing against requirejs. I like their module work and it seems to be used in a lot of places (even dojo's 1.7 loader is able to load requirejs modules and i think they've rewritten their modules so that they can be loaded by requirejs and other AMD loaders).
Anyway, some more comments inline: On Fri, Jun 24, 2011 at 20:01, Howard Lewis Ship <[email protected]> wrote: > True, perhaps there's a slimmed-down version of requreJS (or something > similar) that does the loading, and the callback-on-load, but not the > modules? Actually, I doubt the modules code is that big a deal. > > In any case, > > I hadn't thought about working in terms of requireJS modules, and > given how URLs for JavaScript library assets are organized, it may not > be possible to use the module name (i.e., require('core')) but always > the path (require('/assets/123456/tapestry/core.js'). > That's not the case. In requirejs you can define a base url (in this case "/assets/123456" and then require("tapestry/core") will work. You can also define different urls per submodule so that require("contrib/extra") loads the code from "/contrib-assets/7890" or whatever See http://requirejs.org/docs/api.html#config > However, requireJS modules may not be a bad idea, its where I'm > already headed with T5.extend(). It will help with modularity and > namespace management. > > It will be interesting to experiment ... for example, perhaps > requireJS would be the only JS in full page render, and would be > responsible for bringing in core.js and other main modules. Or, > perhaps, requireJS would simply be another JS rolled into the core > stack. And of course the other interesting experiment is js/css optimization through requirejs. The difference with the current work in Tapestry would be that in the requirejs case you will be able to spread your js code in multiple files without needing to include all of them in the stack - you'd just add the one that requires the rest, e.t.c. > > Intuitively, bootstrapping using requireJS (the first option) would > lead to fast page load times. We will need to update the solution > that deals with premature clicks/form submits, etc. -- Andreas Andreou - [email protected] - http://blog.andyhot.gr Apache Tapestry PMC / http://chesstu.be owner Open Source / JEE Consulting --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
