On Dec 26, 2010, at 12:08 PM, Garrett Smith wrote:
> On 12/26/10, Angus Croll <[email protected]> wrote:
>> Actually closures *do* reference every outer variable even when not
>> used - if they did not, JavaScript's entire lexical scoping principal
>> would break. Evey invocation of a function establishes and enters a
>> new execution context. Each execution context defines a
>> VariableEnvironment which is a repository for variables declared by
>> that context. In addition, each VariableEnvironment will reference the
>> VariableEnvironment of its enclosing context. In this way every
>> function invocation references - in the form of a hierarchy - the
>> entire universe of variables to which it has access.
>>
>> What effect do closures have on memory leaking? Almost impossible to
>> answer
>
> The question needs to be answered on a case-by case basis. The simpler
> functions, the easier this is.
>
> since every JavaScript function in effect forms a closure -
>> that is how functions reference outer variables including global
>> variables. Better to not get hung up on closures and memory leaks per
>> se and instead make the following checks:
>>
>> 1) is my code designed to work in IE7 or older?
>
> The discussed example leaks explicitly in every browser.
>
>> 2) Does my JavaScript make direct reference to a DOM object?
> DOM objects take more memory than user-defined objects and primitives,
> so the consequence of DOM object leaks are more consequential.
>
>> 3) Does that DOM object have any attributes that reference back to my
>> JavaScript (typically expando properties)
>>
>
> That is a step that can help find a different leak pattern.
>
> That pattern uses scope, too, and I believe that a more common case
> for that is not expando properties but is instead properties that
> refer to the native object via event handlers, e.g.
>
> myObj.onclick = function() {
>
> };
>
> where the `myObj` has a reference to the function and the function has
> a reference to `myObj`, via scope.
>
> That's a dfferent leak pattern.
>
> I believe that your TOFU style top post reflects your style of
> reading. It is hindering communications and helping to spread
> misinformation. Please learn how to post so that we can avoid that!
> TIA.
>
>> If the answer to any of the above is no then you probably won't suffer
>> memory leaks - so sit back, relax and take full advantage of
>> closures ;-)
>>
>
> That contradicts your opening line: "Actually closures *do* reference
> every outer variable even when not used". That line is correct.
is there a way we can try this out and probe it? i don't think that by
having access to the variable we get a leak
its by having a reference isn't it? ES garbage collector is a reference
count style collector. maybe they leak if you return
the function reference but not if you return the function result?
if someone has a way to probe this will be very informative.
>
> [snip a fullquote under with signatures etc]
> --
> Garrett
>
> --
> To view archived discussions from the original JSMentors Mailman list:
> http://www.mail-archive.com/[email protected]/
>
> To search via a non-Google archive, visit here:
> http://www.mail-archive.com/[email protected]/
>
> To unsubscribe from this group, send email to
> [email protected]
--
To view archived discussions from the original JSMentors Mailman list:
http://www.mail-archive.com/[email protected]/
To search via a non-Google archive, visit here:
http://www.mail-archive.com/[email protected]/
To unsubscribe from this group, send email to
[email protected]
