L. David Baron wrote:
> Seems reasonable, although I'm not sure if the interaction with
> 'min-width' and 'max-width' is what we want.
It looked right in my testing... where "right" is "what we do before
> I was thinking about moving more of this into the frame classes, though.
Hmm... we could add a "width override" method. In fact, if we're going
to have nsLayoutUtils::IntrinsicForContainer produce results consistent
with the nsHTMLReflowState code, we either need that or to add checks
for these two frame types there and do similar magic.