Re: Neutral directional characters in Mozilla/ Firefox

Thank you for your reply.
I understand my mistake (well, I think!): I didn't consider that fact
that, when the neutral directional characters are resolved, the weak
directional characters have already been resolved (i.e. have received a
direction).
So in the example...
<div dir="rtl">The date is:<br>2006 YLUJ 17</div>
...
1) "1" is given a left-to-right direction according to the W7 rule you
pointed out to me;
2) the colon is seen between two strong directional characters ("s" and
"1") and receives the same direction as these.

Patrick

Simon Montagu schreef:

> [email protected] wrote:
> > Mozilla or Firefox for Linux display the following HTML extract (with
> > the capital letters indicating Arabic text):
> >
> > <div dir="rtl">
> > The date is:<br>
> > 2006 YLUJ 17
> > </div>
> >
> > ... like this:
> >
> > The date is:
> > 2006 YLUJ 17
> >
> > Yet, I understand from the litterature on bidirectional Unicode text
> > that neutral directional characters (like the colon) surrounded by
> > strong directional characters of different directions (in this case:
> > the "s" and the "A") receive the same direction as the paragraph (in
> > this case: right-to-left). So I would have expected the colon to appear
> > at the other end of the line, like so:
> >
> > :The date is
> > 2006 YLUJ 17
> >
> > Moz/FF behave as expected if the second line starts with some
> > right-to-left text. For example:
> >
> > <div dir="rtl">
> > The date is:<br>
> > 2006 YLUJ 17 YADNOM
> > </div>
> >
> > is displayed:
> > :The date is
> > 2006 YLUJ 17 YADNOM
> >
> >
> > Do Moz/FF wrongly consider the digit at the beginning of the second
> > line as a strong character, giving the colon the same (left-to-right)
> > direction? Or am I missing something else?
>
> See http://www.unicode.org/reports/tr9/#W7
>
> In the first example the first strong character before the digit is
> left-to-right, so the digit is considered left-to-right.
> 
> > 
> > Thanks
> > 
> > Patrick
> >