If you have questions or comments, you may post them on this entry.

That is very odd that it displays the puzzle piece. That means that the FlashObject script found the correct plugin in the browser’s plugin array and wrote the object/embed tag out but then the browser couldn’t find a plugin to load the swf.

Are you sure you have the Flash player installed correctly on each browser you are using? Netscape 7 and 8 are very different browsers. Netscape 8 can switch between IE rendering mode and Gecko rendering mode, so be sure that if you installed a plugin, that plugin is installed for each browser.

I just downloaded Netscape 7 and 8 on my test PC, and both of them work fine on my example page. Does my example page work for you?

<script type="text/javascript">
// <![CDATA[
var fo = new FlashObject("/us/swf/75LogoAnim.swf?from=homepage", "fp_animation", "162", "188", 6, "#eee");
fo.addVariable("flashVarText", "this is passed in via FlashVars");
fo.write("fp_logo");
// ]]>
</script>

Apologies if this is something simple I’ve overlooked…I’m really quite incompetent with Flash.

David, you can set the wmode parameter without editing the js file…

Just use: fo.addParam("wmode", "transparent");

That way you can pick which movies are transparent and those which shouldn’t be.

Added to the end of the flashHTML += ‘<embed ..’ line inside the FOP.getHTML function:

wmode="transparent"

Added to PC IE architecture inside the FOP.getHTML function:

flashHTML += '<param name="wmode" value="transparent" />';

Now my ADxMenu drop-downs float on top of the Flash movie using CSS z-index.

Source Code Changes

   <object data="index.swf" type="application/x-shockwave-flash" width="633" height="395" hspace="0" vspace="0">
                        <param name="movie" value="index.swf" />
                        <img src="../../images/getflash.gif" alt="flash" width="633" height="395" hspace="0" vspace="0" />
                      </param>

I had originally used setTimeout() and it appeared to work OK for small numbers of flash elements (about eight or so). When the number of elements increased it didn’t actually fix the problem ( see http://www.dsol.com.au/test/index02.php ).
Also, it was difficult to control the timing of the non-flash components. The test page listed hides the flash divs and then makes them visible after the page loads. But the setTimer doesn’t activate until after the page has loaded and so the results can be ugly if the non-flash content is visually different from the flash content.
Do you have any examples of how you used setTimer() to solve your problem?

By the way, I should point out that this problem exists no matter how you place the flash content in the page – it isn’t a problem with Geoff’s script.

Anyway, that fixed the problem when it cropped up in a sIFR beta release a few months ago, so it may also fix this.

Since the script doesn’t write the object tag to the page if there is no plugin detected, the browser would never use the codebase attribute to locate the missing plugin.

If you have any examples of bugs or problems specifically with the FlashObject script, please post a link or e-mail me the files so I can test them out.

I didn’t consider the FOUC bug because of the script elements I have in the head. Typically, that works. But apparently not so for the FlashObject solution.

Thanks for your help.

BUT, I did look at the FOUC page, and it said that a single script element in the head will prevent it, and it looks like you are embedding the FlashObject script in the head? So I’m not sure. Maybe you found a completely new issue.

Anyway, if you want to try this fix: import all your stylesheets as you normally would, then add one more stylesheet and link it as a ‘print’ stylesheet – check the link to the FOUC page for an example.

When I link the CSS file or write the CSS directly on the page, the FlashObject embed works perfectly in IE. But, unfortunately, I don’t want to do this, because I don’t want older browsers choking on my CSS. I’d rather they get nothing.

Do you think this is just some strange IE bug? Is there any hope that the FlashObject method can be modified to correct this? Or is my only hope to detect the browser with JS and then link my CSS file if it’s a modern browser?

First of all I should outline what the problem appeared to be (please correct me if I have not explained the process):-

In IE6 and FlashPlayer 7, each Flash movie is loaded directly into memory as it loads in the page. If the movies are all different there is no problem because the FlashPlayer merely loads each according to its specification. If, however, the movie is duplicated then the FlashPlayer appears to start loading the movie while the rest of the page loads. When it receives the next command to load the same movie with different parameters, it uses the progress of the first movie as its starting point and, importantly, leaves the original movie in limbo. This continues through all the instances of the movie until the movie is finally loaded completely. Once it is loaded in one of the instances, all subsequent instances load properly but the original unloaded movies stay unloaded.

You most likely wont see this occur on your local test machine because it will load the movie file quickly. Online the download speed is much slower and this is where the problem shows itself. Make sure you delete all instances of the movie from your temporary files cache before each test as the page will simply use the cached version and not show the problem.

To correct this, the FlashPlayer needs a fully loaded instance of the movie before it will load the multiple instances correctly. To control this a dummy instance of the movie is placed in a hidden div block at the top of the page. Then the javascript window.onLoad call is used to activate all the movie instances after the page has finished loading.

The only problem with this solution is that the divs containing the instances of the movie during the loading process need to be hidden, then made visible in the window.onLoad call. If the viewer doesn’t have javascript turned on in their browser then the non-flash alternative will also remain hidden.