non-strict direct eval in top level scope

[Allen Wirfs-Brock]

On Jan 22, 2015, at 5:03 PM, Francisco Tolmasky wrote:

> Apologies as I believe this has been discussed before ( https://esdiscuss.org/topic/block-scope-direct-non-strict-eval ), but just trying to get some clarification as to the current state of things, and have not been able to find this information (in a format I can understand). Namely, I’m curious whether eval(“let x = 5”) adds x to the current scope, if not in strict mode, at the “top level” (i.e. not in a function) and 
> called directly. My impression was that with let it should not be adding anything to the current scope, but my tests in Mozilla Firefox (in the console) seem to suggest they do:

eval always adds lexical declarations (let/const/class) in a lexical environment that only exists for the duration of the eval. See http://people.mozilla.org/~jorendorff/es6-draft.html#sec-performeval 

Allen






> 
> > eval(“let x = 5”) 
> > undefined
> > x
> > 5
> 
> In fact, even in functions:
> 
> > (function() { eval("let xyz = 555"); console.log(xyz) })()
> > 555
> 
> In if statements, it seems to be tacked onto the global scope:
> 
> > if (true) { eval(“let x = 5”) } x
> > 5
> 
> If these are just bugs that’s fine, but if not, could someone tell me the expected behavior? Is it always supposed to act as if its creating a new scope for the blocks, or always *except* for top level evals? Or never?
> 
> Thanks,
> 
> Francisco
> 
> -- 
> Francisco Tolmasky
> www.tolmasky.com
> tolmasky at gmail.com
> _______________________________________________
> es-discuss mailing list
> es-discuss at mozilla.org
> https://mail.mozilla.org/listinfo/es-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/es-discuss/attachments/20150122/2abef8d1/attachment.html>