First: there is no onload attribute for div's. EDIT: please read comments below, very interesting!
Secondly, you should place the url between quotes (escaping them if needed): url('http://lamininbeauty.co.za/images/products/facial.jpg')
Third, there was no image called preloader.gif, yet there was a image called loader.gif, so I used that one to 'fix' your css part for my solution in the jsfiddle demo link at the bottom.
During SO's server-move, I wrote a simple custom function for you that does exactly what you want.
Tested in IE6 and FF12.
To test this: please clear your browsers buffer, otherwise you can't SEE it in action (it would go too fast), since the images will probably be buffered on second view (again, perfect for your goal)!
JavaScript:
var repBg=function(a, t){ t=t||'*'; // by GitaarLAB
var c=document.getElementsByTagName(t), i=c.length, r=[];
while(i--){if (c[i].getAttribute(a)){r.push(c[i]);}} c=r; i=c.length;
repBg.exec=function(){
c[this['data-i']].style.background="#ebebeb url('"+this.src+"') no-repeat top right";
};
while(i--){ if (c[i].getAttribute(a)) {
r=new Image();
r.onload=repBg.exec;
r['data-i']=i;
r.src=c[i].getAttribute(a);
}}
};
// one could run repBg onload, but better to run it when the image has actually loaded, see html!
// window.onload=function(){ repBg('data-bg_img','div'); };
In your BODY: Add the attribute 'data-bg_img' (as per html5 convention, start with data-) to the elements you want to use this technique on and have it contain your background url, like this:
<div id="glykopeels" data-bg_img="http://lamininbeauty.co.za/images/products/glykopeel.jpg">Glykopeels Content</div>
The 'optional' initialization in your BODY:
<!--
trigger the replace background function when the loader image has actually loaded!
rewriting the onload with nothing to prevent infinite loop in IE6 (and greater?) !!
-->
<img src="http://lamininbeauty.co.za/images/products/loader.gif" style="display:none;" onload="this.onload=null; repBg('data-bg_img','div');">
Manual/explanation:
Images DO have a onload-event, so we place a loading-image in the html (at the bottom), that will trigger it's own onload-event, calling repBg() as soon as the browser has actually downloaded this loading-image!!!
The function repBg() takes up to 2 arguments:
- the first mandatory string that contains the attribute that should be selected,
- the second optional argument can define tagname (to limit the first argument).
When invoked, function repBg() will then search the body for elementTagNames that adhere to the second argument or * and then filter them with the first argument.
For each htmlObject that remains in the filtered htmlObjectCollection, a new image is created (not appended to the body) with the htmlObject's attribute-value (url) corresponding to the function's first argument as image-source, together with the htmlObjectCollection's referring id (attribute data-id) for reference.
As soon as these images load, they fire their onload event: calling repBg's exec method that replaces the background of the referenced htmlObject with the new freshly loaded (big) background-image (and the rest of your css). For further modularity you could expand on that function.
Lastly, note: the background images load in order they appear in source, aka the way you expect things to work!!
You can see it in action in this fiddle: http://jsfiddle.net/epdDa/
UPDATE VERSION 2: GRACEFUL FALLBACK!! AND COPY-PASTE NOBRAIN SOLUTION
It annoyed the living daylights out of me that my first solution did not provide graceful fallback. So I made a different solution that provides graceful fallback.
Also fully tested in IE6 and FF12
It works like this:
In your BODY: SIMPLY set your div's class to 'preload' and set it's style-attribute to the backgroundimage it should normally load. Like this:
<div id="facials" class="preload" style="background: #ebebeb url('http://lamininbeauty.co.za/images/products/facial.jpg') no-repeat top right;">Facials Content</div>
That was easy right?
Then place the following script in the HEAD (this is important) of the HTML:
// getElementsByClass original by dustin diaz, modified by GitaarLAB
document.getElementsByClassName=document.getElementsByClassName||function(searchClass,node,tag) {
var classElements = [], i=0, j=0;
if (!node){node = document;}
if (!tag){tag = '*';}
var els = node.getElementsByTagName(tag);
var elsLen = els.length;
var pattern = new RegExp('(^|\\\\s)'+searchClass+'(\\\\s|$)');
for (; i < elsLen; i++) {
if ( pattern.test(els[i].className) ) {
classElements[j] = els[i]; j++;}
} return classElements;
};
var repBg=(function(n,u,p,a,i,r){ // by GitaarLAB
window.onload=function(){repBg(1);};
i=new Image(); i.onload=function(){this.onload=null; repBg(2);};
document.write("<style>."+n+"{background:"+p+" url("+u+") "+a+
" !important; background-size: contain !important;}"+
"</style>");
i.src=u; r=0;
return function(t){
r=r+t; if(r>2){
var c=document.getElementsByClassName(n), i=0, l=c.length, s;
repBg.exec=function(){
document.getElementById(this['data-id']).className='';
};
for(;i<l;i++){
r=new Image();
r.onload=repBg.exec;
r['data-id']=c[i].getAttribute('id');
s=c[i].getAttribute('style');
try { // sane browsers
r.src=s.match(/url\('?([^'"]*)'?\)/i)[1];
} catch(e) { // <IE8
r.src=s.cssText.match(/url\('?([^'"]*)'?\)/i)[1];
}
}
}
};
})('preload','http://lamininbeauty.co.za/images/products/loader.gif','#ebebeb','no-repeat top right');
Explanation:
It took me all night.. but I found a way.
If javascript is enabled, function repBg will start by writing an extra style-block to the documents head (where it is located, note to place it after your last css script), that sets the loader-background-image for all elements with the class 'preload' (thus displaying the load-image at pageload).
If a load-test image for the loading-image is loaded AND the window is loaded (to get to all the elements in the body), then it does basically the same as version 1. Only this time we fetch and match the url from the element's style-atribute and onload subsequently empty the element's style-attribute.
Since this function auto-executes and overwrites itself with a version similar to version 1 (as above), you can simply adjust parameters at the last line of function 'repBg'.
Note that: in it's initial sate repBg accepts a maximum of 4 arguments: className, Url, cssPrepend and cssAppend.
To see it in action (don't forget to clean your browsers buffer as explained),
click this fiddle: http://jsfiddle.net/epdDa/1/
Whoever uses this function, I would greatly appreciate it if you credit me!
UPDATE:
Extra explanations and answers to comments.
Main differences between the 2 versions
Technically both versions use almost the same techniques so there is no real difference there.
- With version 1 the javascript is the glue that IS NEEDED to make the page work, but works in valid true xhtml and plain html.
However, people with javascript turned off will get a nonfunctional site (with only loading-gifs displayed). Note that all other current answers, including the direction you where going, suffer from this problem!
- With version 2 the javascript is only the spice that enhances the page-interaction (the way websites should be coded), but only works in html (or invalid xhtml).
However this should make sure that people with javascript turned off still see a normal functioning page. IE: javascript is NOT NEEDED to display the site correctly. This concept is called 'graceful fallback' or 'degrading gracefully'. My vote no 1 for version 2.
Extra bonus: this path gives you plain vanilla validating and SEMANTIC html since you use ancient trusty in-line style, id and class. My vote no 2 for version 2
Why did I choose to use in-line css? Why 'must' you use in-line css for this to work?
First of all, I spent hours to avoid in-line css. (I did not loose them, I learned way's that did not work, just as useful). Next, I want to point out that again all current answers including the direction you were going, had the actual background image url separated from the css, while in the css you were setting the loader image on each div separately, something where a class would have made more sense. Version 2 simply uses a configurable classname.
Both reading and writing css blocks in the document's HEAD is kind of a mess..
And did I mention linked external css files..??
In my opinion, all this would need so much extra code and cpu-cycles AND blocking/halting the browser on every pageload, in order for the core-priciple to work: the last valid css-rule applies. So the loading image is displayed as soon as possible since it is the specified background image at pageload, exactly what one would want from such a feature. And if the element is no longer part of the 'preload' class? Right: it's inline css takes effect, updated as fast as the browsr can render (if the image is already loaded). Nice.
So if you sacrifice (true) xhtml-support by simply using document.write, it currently still turns out this way is the 'best' way to go (as you can read in the previous 2 links). AND it would still work with an external linked css. My third (KISS-)vote for version 2.
My fourth vote for version 2 is because: the repBg function is prepared to have it's exec method(=function) 'upgraded' so you can only take out the 'preload' value from the class' valuelist. A simple replace() would suffice (currently left out for speed).
My fifth and final vote for version 2 is that because of it's graceful fallback setup, it is also relatively easy to fix for or block some browsers to use the extra 'spice'.
Finally, as for speed: I think version 2 will always feel snappier: onload-image is displayed almost as fast as the browser can fetch it (as if this extra css was always there to begin with), the loading-animations load snappy since: their load is already initiated in the head, and the browser will not download the overruled images until called for by the function. Plus they look interactive without distraction. But.. when the actual background images are loaded and the css updates: bam! the image is there, without the top-to-bottom-scanning'-effect'. That effect feels damn snappy to. Actually I'm convinced and will be doing an adaptation for images in the galary, for the snap-feel and increased perceived initial pageload.. Note, this is my opinion. Your mileage may vary haha.
Good luck!!
(and please vote if you like/use this idea/function, thank you!!)
