hi,
Since 1.7 mmbase has a nice fiter that prevents jpg's with xml data in the file from crashing internet explorers image thread. That's nice.
I've noticed some differences between the handling of images that are filtered and those that aren't. It looks like the images are considered to be 'new' at every request for internet explorer so they will be downloaded again instead of (I cannot reproduce it in firefox).
Probably one of the headers is set to something different by applying the filter, but I don't see a real difference when I look in HandleServlet.java -- does anyone have an idea what might be the difference? Maybe the check if an image has been modified checks against the original image or something?
Simon
