funny Opera behaviour

include not showing straight away

[teylyn]

Redesigning our web site, I use two html includes, one for the navigation and one for some footer content with images.

It works fine in all the browsers (FF, IE5.01, IE5.5, IE 6), but Opera throws a wobbly. The page loads but sans the footer. BUT as soon as I hover over the menu, (which is also an include, remember) and the first menu drops down, Voila! - the footer pops up too.

Duuhh?

The footer is static html containing the XHTML validator images. The menu is some nested ul li construct fashioned after the sons of suckerfish dropdown menu. It formats purely with CSS and uses a few lines of JavaScript to force the a:hover pseudoclass down IE's throat.

In the html code, the div that calls the footer include is positioned before the div that calls the menu include.

Now, for this particular instance I don't much care if the few Opera browsers see the footer but the behaviour is definitely strange and I worry what will happen to other includes.

Anyone come across that before?

[teylyn]

(off topic: does anyone know why I can't edit my own post anymore?)

Narrowing down: Seems that the footer pop-up problem only occurs when the page is longer than the Opera window. Pages that fit into the available space will show their footer straight away. Pages where I have to scroll to see the footer only display it after I have triggered the drop down menu.

strrrrange

teylyn

[tedster]

An include happens on the server -- so the browser itself does not execute any code but only receives the full page source as sent from the server. But, there is a sometimes weirdeness in Opera's interaction with some servers - especially IIS - where the last packet is never fully acknowledged. And so the browser always looks like the page is still loading (the location bar icon stays partial.) So -- does your browser show the page is fully loaded?

And second, have you validated the final page source code?

W3C Validator - HTML [validator.w3.org]
W3C Validator - CSS [jigsaw.w3.org]

<off topic - the "owner edit" button has a relatively
short timer on it. We vary it from time to time, but no
matter what, it always time out relatively quickly.>

[teylyn]

Yep, validates with no errors. The pages are served on my IIS localhost, so you've touched down on the potential source of the problem.

thanks

[tedster]

If you can zero in any further, please let us know what you find. I've struggled with this little nit for many many months. I'm currently suspecting some default configuration may need changing - perhaps in the area of content negotiation.

It's not enough of an issue to be a worry, and it doesn't happen with all IIS servers. Tthat's why I suspect some default configuration that can be changed. But it happens quite often, even on major sites that are hosted on IIS.

[encyclo]

Back a few years ago, IIS functioned in a way which gave a speed advantage to IE over Netscape - the way that the HTTP connections were made was non-standard, missing certain parts of the RFC-defined sequence. I don't know if the same is happening still today with more modern versions of IIS (for the above we're talking about 3 or 4 years ago with IIS3.0 and 4.0 on WinNT) however you can be certain that IIS functions best when combined with IE, whereas Apache follows the HTTP specifications to the letter.

Note: the above is from memory, I'm struggling to find the article/documentation that will back this up...

However, is this really the cause of your problem? If once the page is loaded (but the footer is invisible), can you view the source code and see whether the markup for the footer is present? If so, then the problem is more likely to do with hide/show visibility in the Javascript than anything else.

[tedster]

If it is the bug that I've been wrestling with, it will go away if you click on "stop". If that doesn't change anything, them we'll need to look elsewhere for your situation -- and javascript incompatibilites are often at work with Opera.

[teylyn]

encyclo, thanks for your suggestions. I've checked the source code while the footer is still invisible, and Yes, it is in the code.

I will experiment a bit more and see if the phenomenon persists if I remove all javascript to do with the includes. But these tests will have to wait a bit because I'm now off work for 10 days

HAPPY HOLIDAYS to you all.

cheers

teylyn

[tedster]

Scratch my IIS focus -- at least let's say that the problem I mentioned is not exclusive to IIS. It's reproducible on the Wired site every time, and Wired is served on Apache (Unix).