[dojo-contributors] Dojo 1.7 Goals: "use strict"

I suggest that if it is used, the "use strict" is inside the AMD
factory function. There are issues with combining code that does a
"use strict" at the top with code that is not "use strict" friendly. I
have hit this with requirejs, so just something to be aware of. I have
gone to a model of not using "use strict" but using JSLint to lint the
code, I feel like it gives a decent tradeoff on getting the form in a
more strict form. Although I'm curious where the gaps are with that
approach vs. explicitly setting "use strict". I am sure there are
some, the default value of "this" in unbound function calls probably
being one (which we should not be relying on anyway).

James

fyi - yesterday ff4 was released with support for "use strict" and this
morning chrome 11 was released in the beta channel with support for "use
strict". this gives us 2 reasonable platforms to test on.

ben...

i should have said that chrome 11 has *improved* support for "use
strict" - some support has existed since chrome 8. code i was using
that was previously working in chrome 10 failed in chrome 11 due to an
improper usage of "use strict". i assume this to be due to extra
support for "use strict" introduced in chrome 11.

ben...

i'm all in favor of trying to use strict mode but i wonder if we are
severely limited in how we can apply it without breaking things. try
this in ff4 - http://jsfiddle.net/xRYHq/

it seems that any class created with dojo.declare while in strict mode
will be unable to call this.inherited because of inherited using
arguments.callee to find the caller. wouldn't this heavily reduce the
number of places we can use strict mode?

ben...

Last time I checked it was possible to update dojo.declare() to avoid
using arguments.* If the strict mode is something we want for 1.7, I can
look into that.

Cheers,

Eugene Lazutkin
http://lazutkin.com/