weird var problem

**Gregor Kofler** · Nov 12 '08, 08:45 PM

Re: weird var problem

disappearedng meinte:

Hi everyone,
I have encountered a small problem concerning global variables:
>
var print = function(string ){
var reference = document.getEle mentById("myDiv ");
reference.inner HTML += string + "<br/>";
};
>
This doesn't:
REF = {
div1 : document.getEle mentById("myDiv ")
};
>
var print = function(string ){
REF.div1.innerH TML += string + "<br/>";
};
>
Firebug complains that REF.div1.innerH TML is null . Why?

Does myDiv already exist, when REF is being defined? Your first example
needs the reference only upon being called (perhaps long after the
document has completed loading), the latter one when REF gets defined -
presumably somewhere in a script element.

Gregor

**disappearedng** · Nov 12 '08, 09:05 PM

Re: weird var problem

thx that helped it

**disappearedng** · Nov 12 '08, 09:05 PM

Re: weird var problem

Wait
If I have an html
<html>
<head>
<script type="text/javascript">
.... DEFINITION of the javascript here ...
</script>
<body>
<div id="myDiv"/>
</body>
</html>

My question is:
1) Does Javascript gets loaded before the HTML?
2) If what you said is indeed true, myDiv not defined, then how come
when my calling of print inside the Javascipt definition area works
then? If myDiv is not defined upon compile time, then why will

var print = function(string ){
var reference = document.getEle mentById
("myDiv");
reference.inner HTML += string + "<br/>";
};

this work since myDiv has not been created (or I misintreprid what you
said wrong)

**disappearedng** · Nov 12 '08, 09:55 PM

Re: weird var problem

Wait
If I have an html
<html>
<head>
<script type="text/javascript">
.... DEFINITION of the javascript here ...
</script>
<body>
<div id="myDiv"/>
</body>
</html>

My question is:
1) Does Javascript gets loaded before the HTML?
2) If what you said is indeed true, myDiv not defined, then how come
when my calling of print inside the Javascipt definition area works
then? If myDiv is not defined upon compile time, then why will

var print = function(string ){
var reference = document.getEle mentById
("myDiv");
reference.inner HTML += string + "<br/>";
};

this work since myDiv has not been created (or I misintreprid what you
said wrong)

**Gregor Kofler** · Nov 12 '08, 11:25 PM

Re: weird var problem

disappearedng meinte:

Wait
If I have an html
<html>
<head>
<script type="text/javascript">
... DEFINITION of the javascript here ...
</script>
<body>
<div id="myDiv"/>
</body>
</html>
>
My question is:
1) Does Javascript gets loaded before the HTML?

Loaded? No. Or did you mean parsed? Since the scripts are parsed along
with the rest of the document - it depends.

2) If what you said is indeed true, myDiv not defined, then how come
when my calling of print inside the Javascipt definition area works
then?

When you call the function, myDiv has come to existence due to the
complete parsing of your HTML document.

Gregor

**disappearedng** · Nov 13 '08, 01:55 AM

Re: weird var problem

Dear Greg,
I don't think you are right
Please look at what I meant

<html>
<head>
<script type="text/javascript">
var print = function(string ){
var reference = document.getEle mentById("myDiv ");
reference.inner HTML += string + "<br/>";
};

/*
REF = {
div1 : document.getEle mentById("myDiv ")

};

var print = function(string ){
REF.div1.innerH TML += string + "<br/>";

};
*/

function init() {
print("Hello");
}

</script>
<body onload="init(); ">
<div id="myDiv"/>
</body>
</html>

Try uncommenting the commented part.
What you said was not possible because of the following scenario:

If the javascript is parsed before div was defined, then the top print
function will not work anyway because what revoked it was init() and
init() was called by body.

Both functions are executed at the same time in both scenarios. It
would be not be possible that in one scenario Div was defined and the
other not.

**Gregor Kofler** · Nov 13 '08, 08:55 AM

Re: weird var problem

disappearedng meinte:

Dear Greg,
I don't think you are right

Sigh...

<html>
<head>
<script type="text/javascript">
REF = {
div1 : document.getEle mentById("myDiv ")
>
};

REF.div1 is now "undefined" . Since the browser hasn't come across the
div yet.
....

.... now "myDiv" becomes available. Alas, too late.

....

Much later: The load event fires.

What you said was not possible because of the following scenario:

Said to me by someone, who is completely clueless.

If the javascript is parsed before div was defined, then the top print
function will not work anyway because what revoked it was init() and
init() was called by body.

Revoked? init() would be called once the body is completely loaded. But
REF.div1 was assigned long before that.

Gregor

Both functions are executed at the same time in both scenarios. It
would be not be possible that in one scenario Div was defined and the
other not.

**disappearedng** · Nov 13 '08, 06:05 PM

Re: weird var problem

Alright that clear things up.

Let me confirm with you
1) the "onload="init() ;"" part of my body is executed AFTER the entire
HTML has been parsed?
2) Hence the second print works because init() is called AFTER myDiv
has been created, and when init calls print THEN the javascript parser
looks up the already created myDiv element and provides a non null
reference?

I tried implementing the first print function in init and everything
worked. Thank you.

It's just that I had no idea that body's onload will only be called
when the ENTIRE html has been parsed.

Thank you

**Thomas 'PointedEars' Lahn** · Nov 13 '08, 06:35 PM

Re: weird var problem

disappearedng wrote:

1) the "onload="init() ;"" part of my body is executed AFTER the entire
HTML has been parsed?

No, only after it has been *rendered* also (with some exceptions).

2) Hence the second print works because init() is called AFTER myDiv has
been created, and when init calls print THEN the javascript parser looks
up the already created myDiv element and provides a non null reference?

Roughly speaking, yes.

What really happens in the majority of cases is: The client-side script
source code is parsed into language tokens according to the (modified)
ECMAScript grammar, and JIT-compiled into byte-code before execution.

The Virtual Machine that interprets (executes) that byte-code program then
calls a method of a DOM object that is an implementation of an interface
(method), as the result of the compilation. That method cannot return a
non-null value (with the meaning of `null' being implementation-dependent)
if the DOM object that represents the (HTML) element (an element object) has
not been registered in the DOM tree yet.

Sometimes element objects are created and registered in the DOM tree when
the corresponding element has been parsed. But since the parse tree of the
markup parser is incomplete until the last element (usually the `html'
element) has been parsed or inferred (because of optional tags), you cannot
be sure that the DOM tree is complete either until the `load' event of the
document (standardized in the W3C DOM Level 2 Events Specification) occurs.
That event can be handled with the intrinsic `onload' event-handler
attribute of the `body' element (among other, less reliable methods), as
standardized in the HTML 4.01 Specification.

PointedEars

P.S.: I selected the wrong menu item again, sorry for the e-mail.

weird var problem

weird var problem

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment