/*macros: modules.jsm*/
Whenever I see stuff intended for non-human consumption (doc-comments don't count, because they are just a standardization of a format so they can be more easily consumed by humans, and things like vim modelines don't count because they are not intended to change the meaning of the rest of the file) inside comment markers, I'm reminded of this old JWZ rant about the CDDB format http://www.jwz.org/doc/cddb.htmlI mean that the module implementation will be a leaky abstraction because I can still do something like "Object.prototype.boom = true" inside a module and cause widespread brokenness. This encapsulation/robustness problem isn't something a simple macro can fix very easily, unfortunately, so both the module proposal that you linked to and your macro proposal sound like using a hammer as a screwdriver for that particular problem.
>> The `s' in `defs' stands for "statement"
Ah, that clears things up. I'd prefer something like a "defexpression" though, which would allow
var a = module A {}
On a side note, var a = if (blah) would be cool to see too (well actually, I'd be more interested in var a = for(blah) ).defexpression makes sense, too (but I'd call it `defe' - the more often you use something, the shorter it should be). For some reason, I had the intuition that it would be easier to add statements than expressions, but I don't see why it can't be as simple, as long as they start with some kind of unique token that can be considered a reserved keyword when parsing.
I'm not sure what you mean. I'm thinking you would annotate the beginning of foo.xjs (extended JavaScript) with
Then run the preprocessor, which will generate foo.js - a plain old JavaScript file.The `s' in `defs' stands for "statement", so your example would throw the same error as "var a = if (x==4) ..."
error: expected expression