bugnote – 2009-03-08

IE8rc1 collapses last container?

observed tested investigated problematic fixed
07.mar.2009 08.mar.2009 serious very 19.mar.2009
(IE8 final)

Update [19.mar.2009] IE8 final version shows no signs of the collapsing observed in IE8rc1, so for now I conclude that Microsoft have fixed this instability problem.

The footer randomly collapses completely in height on pages on this site in my copy of IE8rc1 on Vista. May have to do with the fact that footer is the last container in a page.

Update [09.mar.2009] further observation clarifies that IE8rc1 more or less frequent quits rendering the source-code too early, which means all markup/content after a given point gets ignored. The point in source-code at which IE8rc1 quits rendering seems to be stable for any given page, but is not necessarily the same point in different pages based on the same template.

Feedback from other web developers who have tested one or more of my pages on line in their IE8rc1 installations, indicates that failure-frequency varies a bit from installation to installation – no clear pattern emerges.

Sometimes a refresh fixes the problem – the entire source-code gets rendered. Other times a refresh makes IE8rc1 quit rendering too early. Completely unpredictable.

references:
Safari (web browser) dissected by Molly
…or any other regular (all scrolling) 2 or 3 column page on my site.
follow-up (note to self):

Collapsing-problem comes and goes both for local files and on live site. No apparent reason or known bug to track. No other browser has ever presented such a collapsing-problem. No solution found.

Problem does also occasionally appear in IE8rc1 in IETester on windows XP, but I have no way of knowing how accurate that version is compared to the real IE8rc1.

[09.mar.2009]: when served the reference-page (link above) IE8rc1 seems to quit rendering around the end-tags of #superwrapper1 and #superwrapper1.

Mainly looking for bug-triggers in my stylesheets, since modifications to source-code hasn't lead to anything so far. I'm actually not sure if a defined bug-trigger can be found, and that instead IE8rc1 simply cannot handle somewhat complex layouts reliable – maybe because IE8rc1' own error-handling interprets complex but correct and valid HTML/CSS wrong.