should not define the module name when doing amd wrapping

[normanzb]

The module name should be left to AMD loader implementation to figure out, predefine the module name will cause inconsistency on actual path and virtual path. also confuse AMD loaders specially when "context" setting is enabled.

http://requirejs.org/docs/api.html#config-context

[ansman]

You are talking about this line, right?

define("validate", [], function () { return validate; });

What would be the consequences of simply removing the first argument?

[normanzb]

yeah, I think we should remove the first argument so that:

1. If validate.js is running in non-amd environment, we use global.validate.
2. If it is running in AMD environment, the developers should knows where the module located and call define(['./path/to/validate'], function(){...}) or require with corresponding path to load it.

By removing the first argument it also benefit below situations:

1. It doesn't confuse AMD module loader about which context it belongs to.
2. It doesn't confuse developers on which module name to use, says we place validate.js in a folder named foo, the correct path to load it should be foo/validate, however, because validate.js hard coded module name, above path doesn't work.
3. After compile it with AMD module optimiser such as r.js, the optimiser can automatically fill in the correct path/module name just after define(. (hard coded module name prevent r.js from filling in the correct module name)

[normanzb]

agreed on removing it?

[ansman]

Would there be any downsides of removing the name?

[ansman]

Released in 0.7.0, validate.js no longer supplies a name when calling define