So, in doing more testing, I have not seen this behavior in Internet Explorer. I have also not seen this behavior when in the standard (non sandbox) version of iGoogle. It seems to be isolated to FireFox (2.0.0.14 in this case) in the iGoogle sandbox.

I have captured the two versions of the Hot Trends Gadget that continue to cycle repeatedly.

http://google-gadgets-api.googlegroups.com/web/hottrendssmall.jpg?gsc=bEV1YgsAAAAgOXz-AY7SI75ez879OX-m

http://google-gadgets-api.googlegroups.com/web/hottrendsbig.jpg?gsc=bEV1YgsAAAAgOXz-AY7SI75ez879OX-m

(If those links don't work, I'll try...)

http://google-gadgets-api.googlegroups.com/web/hottrendssmall.jpg

http://google-gadgets-api.googlegroups.com/web/hottrendsbig.jpg

The hottrendssmall.jpg file shows when the Gadget is small, causing "paragon hard disk manager 8" line to wrap at the "8". This causes the "more" link to be cut off and creates a vertical scroll bar (in the iFrame or div?)

Then, the code seems to attempt to auto-adjust for this, and causes the Gadget to become bigger. At this point, it appears as the hottrendsbig.jpg file. However, since doing so removes the vertical scroll bar, it also makes it so that "paragon hard disk manager 8" no longer has to wrap. So, I believe the code then detects that the Gadget is too large, with too much white space beneath the "more" link... so it shrinks it again.

This cycle continues repeatedly until I leave the page, or repeatedly click at the STOP button in the browser to force the JavaScript code to end execution.

I have not looked into the code... so it is difficult to know if the auto-grow/auto-shrink code is a part of the Gadget itself, or a part of the general iGoogle code. However, when I leave the sandbox, this Gadget does not repeat this behavior.

I should warn, though, that it is very specific about sizing. If a particular item is wrapping one day, and growing the Gadget does not remove the wrapping for that line, the Gadget will not shrink again and everything will be fine. And since this Gadget is based on trends of the moment, it changes frequently and does not always exhibit this problem.

I hope this helps track down the problem if it does, in fact, exist in the code for the iGoogle sandbox.

On May 15, 10:08 am, "Dan (Google Employee)" <api....@google.com> wrote: