'this' going out of scope

[Emanuele D'Arrigo]

Hi everybody,

in the small snippet of code below, you can see two identical
debugMsg(this.m yname) statements, one correctly returns "manu"
but the other returns "undefined" . This is due to 'this' getting out
of scope. Is there a way I can retain the 'this' object inside the
function() block? So that I can affect the object calling this method
from there? Any workarounds instead?

Here's the code:

----------------------------------------------
function AreasManager()
{
this.myname = "manu";
}

AreasManager.pr ototype.addArea sFromXmlFile = function(xmlFil e_)
{
debugMsg(this.m yname);
var request = GXmlHttp.create ();
request.open("G ET", xmlFile_, true);
request.onready statechange = function()
{
if (request.readyS tate == 4)
{
var xmlContent = request.respons eXML;
debugMsg(this.m yname);
}
}
request.send(nu ll);
}
----------------------------------------------

Thanks for your help.

Manu

[RobG]

Emanuele D'Arrigo wrote:

Hi everybody,

in the small snippet of code below, you can see two identical
debugMsg(this.m yname) statements, one correctly returns "manu"
but the other returns "undefined" . This is due to 'this' getting out
of scope. Is there a way I can retain the 'this' object inside the
function() block? So that I can affect the object calling this method
from there? Any workarounds instead?

The value of the this keyword is determined entirely by how you call a
function, not by how you declare it. Try this post:

<URL:
http://groups.google.com.au/group/co...78eba47a1a7dc9

>

Or search for the recent thread with subject:

Assigning methods to objects, and assigning onreadystatecha nge to an
XMLHttpRequest -- an inconsistency?

>
Here's the code:

----------------------------------------------
function AreasManager()
{
this.myname = "manu";
}

AreasManager.pr ototype.addArea sFromXmlFile = function(xmlFil e_)
{
debugMsg(this.m yname);

The this value here refers to an instance of an AreasManager object
*provided* it is called as a method of an AreasManager object.

var request = GXmlHttp.create ();
request.open("G ET", xmlFile_, true);
request.onready statechange = function()
{
if (request.readyS tate == 4)
{
var xmlContent = request.respons eXML;
debugMsg(this.m yname);

It a common assumption that in the above case, the this value of the
anonymous function refers to the request object, but as the song says,
"it ain't necessarily so". The link above gives the reason why.
--
Rob

[Emanuele D'Arrigo]

Rob, thanks for your help.

This closure thing is giving me an headache. I've gone through quite
a few posts and webpages on the matter and I feel fairly confused now.
I understand why it doesn't work, but I still can't quite understand
how
to make it work or how to change my approach to work around it.

I.e, now I tried this:

function AreasManager()
{
this.myname = "manu";
}

AreasManager.pr ototype.addArea sFromXmlFile = function(xmlFil e_)
{
debugMsg(this.m yname + "1");

var request = GXmlHttp.create ();
request.open("G ET", xmlFile_, true);

var self = this;
var req = request;
function handler()
{
debugMsg(self.m yname + "2");
if (req.readyState == 4)
{
debugMsg(self.m yname + "3");
}
}

request.onready statechange = handler();
request.send(nu ll);
}

The first two debugMsg() statements work fine and will print out
the expected manu1 and manu2. But the flow never gets to
the third debugMsg() statement, because it's now the req object
to be out of scope. Why is it?

Shouldn't var req = request; act as var self = this; ??

What am I missing?

Thanks again for your help.

Manu

[Michael Winter]

Emanuele D'Arrigo wrote:

This closure thing is giving me an headache. I've gone through quite
a few posts and webpages on the matter and I feel fairly confused now.
I understand why it doesn't work, but I still can't quite understand
how to make it work or how to change my approach to work around it.

You're alterations are almost correct, but you introduced a new error
not present in your original post.

AreasManager.pr ototype.addArea sFromXmlFile = function(xmlFil e_)

[snip]

var req = request;

This is unnecessary, by the way. Simply use request - both are local
variables, and both will be available within the scope chain of the
function that follows. All this does is add a second reference.

function handler()

There's no need for this to be included as a function declaration. A
function expression, as in the original, will do.

[snip]

request.onready statechange = handler();

Your problem is here. Instead of assigning a reference to the function
to the property, you call the function and assign the return value.

[snip]

The first two debugMsg() statements work fine and will print out
the expected manu1 and manu2. But the flow never gets to
the third debugMsg() statement, because it's now the req object
to be out of scope. Why is it?

Because when the function is called, the readyState property isn't 4.
You should have checked that. :-)

[snip]

Mike

[Emanuele D'Arrigo]

Aaaah!!! Thank you Michael!
At last I managed to come up with a working method:

AreasManager.pr ototype.addArea sFromXmlFile = function(xmlFil e_)
{
var self = this;

var request = GXmlHttp.create ();
request.open("G ET", xmlFile_, true);
request.onready statechange = function()
{
if (request.readyS tate == 4)
{
var xmlContent = request.respons eXML;
self.addAreasFr omXmlElement(xm lContent);
}
}

request.send(nu ll);
}

Thank you very much, this seems to work exactly as
I intended! In fact it's extremely similar to what I had
in the first place, except for the use of 'self'. Thank you
again.

Manu

[Laurent Bugnion]

Hi,

Emanuele D'Arrigo wrote:

Aaaah!!! Thank you Michael!
At last I managed to come up with a working method:

AreasManager.pr ototype.addArea sFromXmlFile = function(xmlFil e_)
{
var self = this;

var request = GXmlHttp.create ();
request.open("G ET", xmlFile_, true);
request.onready statechange = function()
{
if (request.readyS tate == 4)
{
var xmlContent = request.respons eXML;
self.addAreasFr omXmlElement(xm lContent);
}
}

request.send(nu ll);
}

Thank you very much, this seems to work exactly as
I intended! In fact it's extremely similar to what I had
in the first place, except for the use of 'self'. Thank you
again.

Manu

I would recommend using against "self", because in web-browser based
JavaScript, "self" means the same as "window", it's a predefined name.
It's hidden by the local "var" declaration, but it will confuse your
audience.
http://developer.mozilla.org/en/docs/DOM:window.self

Greetings,
Laurent
--
Laurent Bugnion, GalaSoft
Software engineering: http://www.galasoft-LB.ch
PhotoAlbum: http://www.galasoft-LB.ch/pictures
Support children in Calcutta: http://www.calcutta-espoir.ch

[Emanuele D'Arrigo]

Laurent Bugnion wrote:

I would recommend using against "self", because in web-browser based
JavaScript, "self" means the same as "window", it's a predefined name.

I was wondering why my editor did highlight it... thanks Laurent, will
do.

Manu