Hi, first, thank you very much for your hard work making our life easier

First, the bug. with pie 1.0 everything is fine but with 2.0 beta in ie9, when I hover over a rollover menu in pure css3, when the new box appear, the linear gradient background flashed. once the first rollover is open, the other one that open after are not flashing the background. This is not present in ie8 with pie 2.0beta, only ie9.

This bug alone prevent me from using pie 2.0 witch is a big problem for me as explained in my following implementation case.

I have a network of 364 websites. I reuse the same images on all of them. So for 1 website, Total image size 1M - Pie 1.0 size 80k
For 1 website pie is 12.5 time smaller then the images... Perfect

For all 364 websites, total images size is 1M and pie size is 29M so it is now a whopping 29 times bigger the the images.
I just cant use this solution.

But your new loading mechanism would completely resolve this issue and be 2k multiply by 364 = 732k so 40 times smaller. This is huge for a network.

But this IE9 bug I encounter is preventing me from using beta so pretty please, continue your hard work.

I believe you are on the right path with version 2.0. I will test all new beta and report here when it is fixed. As soon as this bug is fixed will implement it on my network.

Sylvano

Interesting, I haven't noticed this issue -- for example the rollover gradients on the menu on css3pie.com don't seem to have any flashing. Can you provide a standalone test case that shows the issue?

The gradient rendering in IE9 was switched from SVG to Canvas in 2.0 so it's possible there's some difference in timing.

For the rest of your post ... I'm very confused by your description. Are you implying that by hosting the images on a central domain you are able to take advantage of browser caching when users visit more than one of your sites? If so, then I suppose it's true you'd get some added network benefit by hosting PIE's secondary JS logic files on the same central domain, due to caching. But unless all your users visit all 364 of your sites, then your numbers of "29 times bigger" or "40 times smaller" seem extremely inflated.