Array.prototype.join with cycles

# Brandon Benvie (13 years ago)

Given the following:

var x = [1,2,3]; x.push(x); x+"";

What should happen? The behavior demonstrated by V8, Spidermonkey, and Chakra (haven't tested others) is to return the empty string for cycles, thus returning the result '1,2,3,'. However, unless I'm missing something, this behavior is unspecified and an equally valid result would be to infinitely loop or throw a recursion error.

For comparison, JSON.stringify explicitly specifies how to handle recursion.

Given the following:

var x = [1,2,3];
x.push(x);
x+"";

What should happen? The behavior demonstrated by V8, Spidermonkey, and
Chakra (haven't tested others) is to return the empty string for cycles,
thus returning the result '1,2,3,'. However, unless I'm missing something,
this behavior is unspecified and an equally valid result would be to
infinitely loop or throw a recursion error.

For comparison, JSON.stringify explicitly specifies how to handle recursion.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/es-discuss/attachments/20121224/34e8e9a7/attachment.html>

# Brandon Benvie (13 years ago)

I meant JSON.stringify explicitly specifies how to handle cycles.

I meant JSON.stringify explicitly specifies how to handle cycles.

On Monday, December 24, 2012, Brandon Benvie wrote:

> Given the following:
>
> var x = [1,2,3];
> x.push(x);
> x+"";
>
> What should happen? The behavior demonstrated by V8, Spidermonkey, and
> Chakra (haven't tested others) is to return the empty string for cycles,
> thus returning the result '1,2,3,'. However, unless I'm missing something,
> this behavior is unspecified and an equally valid result would be to
> infinitely loop or throw a recursion error.
>
> For comparison, JSON.stringify explicitly specifies how to handle
> recursion.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/es-discuss/attachments/20121224/c2ee1d14/attachment.html>

# Brandon Benvie (13 years ago)

I realized this is more appropriately filed as a bug (I think it's a spec bug anyway...) and have done so ecmascript#1184.

I realized this is more appropriately filed as a bug (I think it's a spec
bug anyway...) and have done so
https://bugs.ecmascript.org/show_bug.cgi?id=1184.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/es-discuss/attachments/20121224/cf56c55d/attachment.html>