This is all very interesting, maybe there is a bug in jQuery with the
ready event for IE6?
Because after the ready event fires, keep in mind, the ready event is
NOT part of the JS standard and it is entirely up to jQuery to
determine what "ready" is, execution control should be sent back to
whatever was running beforehand.
I was just reading about how IE's memory manager for DOM is separate
from their JScript engine, could it be that DOM processing occurs
while dcs_main() is running and then triggers the "ready" event when
it reaches the bottom of the document?
Could an expert chime in on this?
- jake
On 6/15/07, Theo Welch <[EMAIL PROTECTED]> wrote:
> Thanks for the help, the clear explanations and the time you spent, guys.
>
> I should point out that none of my jQuery code inside the
> $(document).ready() is manipulating the <img /> tag created by the WebTrends
> script. My jQuery scripts are totally separate and unrelated to the
> WebTrends scripts (at the moment).
>
> I just want to try to make this all as clear as possible for myself and for
> anyone coming upon this thread in the future. As you deftly pointed out
> Jake, the $(document).ready() is *delaying the execution* of my jQuery code
> inside it. So while:
>
> $(document).ready( function() {
> /do stuff
> });
>
> ...is *evaluated* (loaded) first in the browser, none of the "//do stuff"
> code inside $(document).ready() is *actually executed* before the dcs_main()
> function is invoked (since the dcs_main() function is invoked directly in a
> script tag, embedded in the page).
>
> So the code inside the dcs_main() function actually *does execute before*
> the code inside $(document).ready() starts to execute, even though the
> $(document).ready() code comes first in the document. So, I don't think that
> moving the jQuery code to be below the call to dcs_main() would have any
> effect on this IE6 problem.
>
> And (I may be wrong here) but I suspect that IE 6 is firing the
> $(document).ready() event exactly when it thinks the document is ready, but
> it doesn;t consider any currently executing JavaScript in whether or not the
> document is "ready." Therefore my jQuery code gets started even though the
> dcs_main() function hasn't finished its execution. I think this "bad
> behavior" on IE's part is what causes the so-called "DOM collision" in IE6
> that effectively prevents the WebTrends <img /> tag from getting appended to
> the DOM.
>
> It's as if IE6 "hijacks" the single JavaScript thread that dcs_main() is
> currently executing within when the $(document).ready() event is triggered,
> so the dcs_main() function never (or very rarely) gets a chance to finish.
> It's like the JavaScript thread is simply "reset" in IE 6. I hope that makes
> sense.
>
> And there also lies "the rub" with Jon's good idea of attaching a .load()
> event to the WebTrends <img /> tag: the <img /> tag never gets inserted into
> the DOM by the WebTrends JS (since the dcs_main() function is
> short-circuited when the JS thread is hijacked), so the image will actually
> never load. :)
>
> I think I will go ahead and use the "backup measure" methodology I
> envisioned. I will add some code to my $(document).ready() that will check
> (in IE 6 only) for the existence of the WebTrends <img /> tag in the DOM and
> if it does not exist, I will invoke the dcs_main() function from within that
> conditional (which will be inside the $(document).ready() ). I'll post the
> final, (hopefully) working code after I have tested it.
>
> If anyone else can think of any other ideas, I'm all ears. :)
>
> Thanks again, guys.
> -THEO-
>
>
>
>
>
> On 6/15/07, Jake McGraw < [EMAIL PROTECTED]> wrote:
> >
> > JavaScript is a single thread/process and executed in the order (top
> > to bottom) of the page. To provide additional functionality, JS has a
> > series of events, which interrupts the standard execution order
> > (onclick, onload, onblur, etc...), the $(document).ready() binds a
> > callback function to the document object to fire during the "ready"
> > event. What is most likely occurring in your case:
> >
> > <script>
> > $(document).ready(function(){
> > // Do some stuff
> > });
> > </script>
> > ...
> > <body>
> > ...
> > <script>
> > dcs_main();
> > </script>
> > </body>
> >
> > The "ready" event won't fire until the HTML for the entire document is
> > loaded, while the dcs_main() function will fire as soon as it's
> > reached by the parser. Normally, I would suggest using:
> >
> > $(document).load(/* callback goes here */);
> >
> > To halt the execution of your code until all of the document images
> > are loaded, but since you're using JS to modify your DOM during
> > parsing, the onload event may fire before the image is actually
> > loaded, because it isn't detected as part of the original document.
> > So, my suggestion is to do one of the following:
> >
> > A. Try changing $(document).ready() to $(document).load()
> >
> > B. Place all your JS at the bottom of your document, like so:
> > <script>
> > dcs_main();
> > /* jQuery code goes here */
> > </script>
> > </body>
> >
> > C. Continue doing what you're doing, there's nothing wrong, as far as
> > I know, with using the $(document).ready() callback to call other
> > functions.
> >
> > Disclaimer: This one really made me think, so everything I said above
> > may not be right.
> >
> > - jake
> >
> > Mike answered the main question here, but just to clarify:
> >
> > .load() or .bind('load') is the jQuery way to respond to the "onload"
> > event. This event only gets fired when the element it is attached to
> > is *completely* loaded. That means images, etc. must be completely
> > transferred before the event fires. The .ready() method of jQuery
> > registers code that will run when the DOM is ready, which could well
> > be quite a bit before the "onload" event occurs for the page.
> >
> > It is possible to attach an onload handler to an individual element
> > (e.g. image) though, which can be useful.
> >
> > --
> > Jonathan Chaffer
> > Technology Officer, Structure Interactive
> >
> > On 6/15/07, skyeflye <[EMAIL PROTECTED]> wrote:
> > >
> > > Hi all,
> > >
> > > I have a client that is running WebTrends for a large, enterprise web
> > > site. The way WebTrends works is very simple, and similar to many
> > > other user-tracking software. There is a chunk of WebTrends JavaScript
> > > code (they call it the "SmartSource Data Collector Script"), embedded
> > > in the bottom of every page's source code. This script basically just
> > > gathers several environment variables and then appends a new <img />
> > > tag to the DOM. The source (src) for the image tag is loaded from a
> > > separate web server that is running the WebTrends software. So, when
> > > the WebTrends server receives the request for this tiny image, it logs
> > > the visit to the page in its WebTrends database.
> > >
> > > The "SmartSource" script (embedded at the bottom of every web page)
> > > just defines a function and then immediately invokes it, like this:
> > >
> > > function dcs_main() {
> > > // collect client info;
> > > // create a new <img /> object;
> > > }
> > > dcs_main(); //immediately invoke the above function
> > >
> > >
> > > The problem I am having is that 99% of the time, IE6 is not finishing
> > > the execution of the WebTrends code, so the special <img /> tag is not
> > > getting appended to the DOM. This is obviously bad because then the
> > > visit to the page is not recorded by WebTrends.
> > >
> > > I also have a few simple jQuery DOM manipulations firing on $
> > > (document).ready(). Based on some testing, my theory is that in IE6,
> > > the WebTrends dcs_main() function is beginning its execution but it
> > > never finishes because while it is running, my simple jQuery DOM
> > > manipulations also begin, and the two scripts "collide" when they both
> > > try to change the DOM at the same time. There are never any errors or
> > > anything. I guess I am actually lucky that only IE6 is having this
> > > problem since any browser might have a problem with DOM manipulation
> > > collisions.
> > >
> > > One potential solution is to re-invoke the WebTrends dcs_main()
> > > function from inside the $(document).ready() event IF the special
> > > image tag has not already been added to the DOM. That way, if the
> > > dcs_main() function was "short-circuited" by the jQuery scripts (or if
> > > it was somehow not invoked at all), it would definitely be re-invoked
> > > from inside the $(document).ready() event. This "backup measure" could
> > > help to ensure that the <img /> was appended to the DOM no matter what
> > > and that the visit to the page was properly recorded by WebTrends.
> > >
> > > I am just wondering if anyone else could recommend a solution with a
> > > different methodology that might not require this kind of "backup
> > > measure" approach. For example, if there was any way to perhaps delay
> > > my jQuery scripts in $(document).ready() from running until after the
> > > WebTrends <img /> tag had been appended to the DOM (which will happen
> > > nearly instantaneously anyway).
> > >
> > > Thanks!
> > >
> > >
> >
>
>
