Alan Gresley wrote: > Gunlaug Sørtun wrote: >> If a browser can't stack various layers of one element together in >> the right order on top of all layers of another element, without >> explicitly being "told" to group and stack element layers by using >> a nonsensical property/value for the case, then it is a serious >> stacking bug. > > > I know the fix/bug. As I said previously this is deeper.
Not deeper than how a rendering engine (all rendering engines) splits up and treats elements as a set of "layers" which are stacked in a certain order depending on rules - of which CSS takes part. The fact that we can "fix" this case doesn't help much either, since the missing "hot-spot" on an anchor is more like a symptom for something potentially much worse. We should be able to interact with - :hover on (amongst other things) - all elements, so a simple test should reveal if this is a general flaw or isolated to one element in IE8. Add something like this to any ordinary web page... html *:hover {color: red;} html * *:hover {color: green;} html * * *:hover {color: yellow;} ...and compare what part(s) of (the whole of) each element in the page reveals hot-spots. Compare with the results in latest versions of other major browsers, to see if IE8 deviates from "the norm" in any way. It shouldn't if they've gotten the element layering and stacking correct. >>> I don't understand what you mean by "How does it handle that >>> filter if rolled back to IE7?" I see that my band pass filter >>> works. >> Toggle IE8 to render like IE7, and see how it handles such filters >> that are known to work in a certain way in IE7. > > > Sorry Georg, remember that Ingo had to help me find that hidden blue > arrow after I had replied to you. I didn't know what you meant. :-) That's ok Alan, it takes time to figure out an entirely new browser. I can relax and wait for answers, while piecing IE8 together in my head without having to do the work :-) >>> Sorry Georg, but you are using a xml prolog like myself, and this >>> is throwing IE8 into quirks mode. >> [...] > I not completely correct here in a true sense since this can only > happen when a comment appears between the xml prolog and doctype. Not quite unexpected... <http://www.satzansatz.de/cssd/quirksmode.html> ...and not really a problem. May change if they stop counting comments as source-code nodes in later betas and final though, so we should keep an eye on it. >> If IE8 doesn't emulate previous versions/modes *identical to* those >> previous versions/modes, the whole version target mess that has >> been "the talk of the town" lately, becomes "utter nonsense" in any >> order. > [...] The perfect emulation of previous versions in all modes is essential, since any deviation means we'll have yet another version-set to go through and fix up while designing - and then another and yet another as new main IE-versions are released. That would be a disaster. regards Georg -- http://www.gunlaug.no ______________________________________________________________________ css-discuss [EMAIL PROTECTED] http://www.css-discuss.org/mailman/listinfo/css-d List wiki/FAQ -- http://css-discuss.incutio.com/ List policies -- http://css-discuss.org/policies.html Supported by evolt.org -- http://www.evolt.org/help_support_evolt/