On May 4, 9:00 pm, "Jerome (Guru)" <jero...@gmail.com> wrote:

Since the Google parent div has a specified height of auto the browser must compute its actual height based on the sum of the heights of its child nodes. In this instance, the sum comes out to 0 since the only child div has a position of absolute and, as a result, is removed from normal flow. The browser then applies overflow:hidden and, bam, you can't see anything inside of the parent div. I'm not sure how the CSS spec says you're supposed to climb out of this hole. But, in standards compliance mode the browsers do.

When you set the child div's position to relative it is returned to normal flow. The parent div's height is then calculated to be the height of the child div, or 130px. Another fix would be to set an explicit height on the parent div -- if it weren't being served automatically from the ig servers.

So directly this is related to the CSS in the div the ig servers automatically insert into gadgets. But indirectly, this is a consequence of running gadgets in quirks mode. If the gadget were rendered in standards compliance mode then IE would have correctly expanded the height of the parent div and this wouldn't have been an issue. It's the reverse of the bug detailed here: http://www.positioniseverything.net/explorer/expandingboxbug.html

I hope the Google dev team takes this as more evidence that being able to control the rendering mode of gadgets would be a valuable tool to developers.