So, I'm writing a web app. Pretty much everything is done client-side, the server is but a RESTful interface. I'm using jQuery as my framework of choice and implementing my code in a Revealing Module Pattern.
The wireframe of my code basically looks like this:
(function($){
$.fn.myplugin = function(method)
{
if (mp[method])
{
return mp[method].apply(this, Array.prototype.slice.call(arguments, 1));
}
else if (typeof method === 'object' || ! method)
{
return mp.init.apply(this, arguments);
}
else
{
$.error('Method ' + method + ' does not exist on $.myplugin');
}
};
var mp =
{
init : function( options )
{
return this.each(function()
{
// stuff
}
},
callbacks : {},
addCallback : function(hook_name, cb_func, priority)
{
// some sanity checking, then push cb_func onto a stack in mp.callbacks[hook_name]
},
doCallbacks : function(hook_name)
{
if (!hook_name) { hook_name = arguments.callee.caller.name; }
// check if any callbacks have been registered for hook_name, if so, execute one after the other
}
};
})(jQuery);
Pretty straightforward, right?
Now, we're able to register (multiple, hierarchical) callbacks from inside as well as from outside the application scope.
What is bugging me: To make the whole thing as extensible as possible, I'd have to resort to something along these lines:
foo : function() {
mp.doCallbacks('foo_before');
// do actual stuff, maybe some hookpoints in between
mp.doCallbacks('foo_after');
}
Every single function inside my app would have to start and end like that. This just doesn't seem right.
So, JS wizards of SO - what do?