Stan Brown wrote:
I'm sorry -- in stripping some of the excess from the original
(validated) page that displayed the problem, I accidentally created
invalid HTML.
That explains that :)
I've fixed that, and validated both the HTML and CSS with no
warnings or errors. Now if you look at
http://www.acad.sunytccc.edu/instruc...tat15/demo.htm
http://www.acad.sunytccc.edu/instruc...n/npscreen.css
you see the same behavior I reported earlier -- text doesn't wrap
around the floated tables, and even an explicit clear doesn't seem
to work.
Indeed - this time with Firefox 0.9.1.
I'm using Moz 1.4. You mentioned that IE and Opera seemed okay but
Firefox didn't. Maybe this is a Mozilla bug -- if so, does anyone
know a workaround?
<fiddles>
Ah, interesting. I don't know what the CSS spec implies is the correct
interpretation, but I think I do see why Gecko is getting upset.
I hope someone will be able to explain whether Gecko or Opera is taking
the correct approach.
You insert your floated table as the last object within a list item,
then in the following list item, want the paragraphs contained therein
to flow around the preceeding table. With Gecko, whilst they are
rendered to the side of the table, they refuse to flow below it.
Although nothing jumps out as being blatantly wrong with such a
structure, it does feel somewhat unnatural to me. Without knowing what
the spec implies, I'd be inclined to think that logically, as the text
below is in a separate list item, the browser could reasonably render it
in a way which makes it quite separate from the preceeding list item.
It would seem more natural to me to make the table the first item within
the subsequent list item, particularly as structurally, it contains
information related to that list item (hence the reason you want the
text near it).
Doing this seems to avoid your problem & the code renders, in respect of
it, fine in Gecko, IE 6 & Opera 7.23 (though the latter two do seem a
bit unsure as to where to place the list item number - Opera puts it
inside the table; I imagine this can be worked around).
I'd be very interested to hear what others think a browser ought to do
with your construct.
--
Michael
m r o z a t u k g a t e w a y d o t n e t