Avoiding the IE rollover bug

[tedster]

IE 5 and 6 have a bug that's been hard to pin down. With some javascript rollovers, even though the rollover image is preloaded and can be verified as in the cache, the browser still calls back to the server for a new download when there's a mouseover.

I just read that the problem only happens when an absolute URL is used for the path to the image. The obvious work-around then is only to use relative paths for BOTH the preload and the mouseover. This may be a bit tricky to notice if your preload function uses an array, but it should be doable by dissecting the code just a bit.

I haven't had a chance to check out this explanation and fix, however, and I'm not likely to get around to it for a while. I've got to fire up a nice, slow dial-up account in order to be sure of what I'm seeing.

Does this explanation line up with what other members are experiencing?

[amznVibe]

what you are saying is a known and documented bug... also what you are saying is probably the only fix... however there is a way to use CSS for rollovers that avoids this bug but its not as dynamic (a:hover background images)