Brendan Eich wrote:
Herby Vojčík wrote:
Hello,
to make things more DRY in spec and more consistent, what if
{some_content} in expression context would be defined as syntactic
sugar for (new Object).{some_content}?
No! You have just reintroduced the ES3 bug where a rebound or shadowed
Object can be used to steal data expressed via an object literal
(formerly, pre-built-in-JSON.parse, via eval).
An object literal {...} is *not* equivalent to (new Object).{...}. The
original value of Object.prototype is used. It turns out no observable
constructor effects come from the original value of Object, so there's
no need to construct a plain empty object, just allocate one that
delegates directly to Object.prototype.
I think you are making a mistake treating the spec as code to "desugar"
from more primitive (object literal) to more derived (mustache
expression) forms. Indeed it seems to me the pun on object literals that
mustache does is not helpful, on balance. There is no new object, not
even an observable throwaway.
/be
I meant it so, with the always used "use original value of ...", of
course, just did not write it. Of course, I did not mean to use (new
Object) in dynamic manner.
This was only about defining that what is inside {} is that same what is
inside .{} and the same object would be created if .{} was applied to
_safe_ empty object.
Herby
_______________________________________________
es-discuss mailing list
[email protected]
https://mail.mozilla.org/listinfo/es-discuss
