Re: Bidi, HTML5 and CSS3, test bidi-html5-019

>  What does that mean in terms of implementation?

I am not sure what you mean.

The Writing Modes spec says that unicode-bidi:embed on an element
"corresponds to adding a LRE (U+202A), for ‘direction: ltr’, or RLE
(U+202B), for ‘direction: rtl’, at the start of the element and a PDF
(U+202C) at the end of the element." Similarly, unicode-bidi:bidi-override
"corresponds to adding a LRO (U+202D), for ‘direction: ltr’, or RLO
(U+202E), for ‘direction: rtl’, at the start of the element and a PDF
(U+202C) at the end of the element." This is a plausible implementation of
these unicode-bidi values.

Now, when elements with these unicode-bidi values include a paragraph break
(like <br>), the Writing Modes spec says that "the bidi control codes
corresponding to the end of the element are added before the interruption
and the codes corresponding to the start of the element are added after
it." This is referring to the LRE|RLE|LRO|RLO|PDF specified above.

> <span style='direction:d; unicode-bidi:u'>...<br>...</span>
> is equivalent to:
> <span style='direction:d; unicode-bidi:u'>...<span><br><span
style='direction:d; unicode-bidi:u'>...</span>

That is correct (in terms of bidi ordering) for unicode-bidi:embed,
unicode-bidi:bidi-override, and (in a null sort of way) unicode-bidi:normal.

It is not correct for unicode-bidi:isolate and unicode-bidi:plaintext,
which specify that "for the purpose of bidi resolution in its containing
bidi paragraph (if any), the element itself is treated as if it were an
Object Replacement Character (U+FFFC)." U+FFFC is in ON bidi class. Thus, a
<br> inside a unicode-bidi:isolate element acts as a paragraph separator
within that isolated element (with the same "reopening" as above), but the
paragraph around the isolated element remains unbroken, since the isolated
element has the same effect on that paragraph as an ON-class character (not
B).

(This is implemented correctly in Firefox, but not in WebKit)

On Fri, Mar 9, 2012 at 6:28 PM, Eric Muller <emuller@adobe.com> wrote:

>  Thanks for the answer.
>
> I now understand that <br> is really intended to be a paragraph separator,
> as far as bidi is concerned.
>
> I also understand better the relationship between HTML and CSS.
>
> I am still a bit confused but before I can express that, a couple of
> questions:
>
>
> 'If an inline element is
> broken around a bidi paragraph boundary (e.g. if split by a block or
> forced paragraph break), then the bidi control codes corresponding to
> the end of the element are added before the interruption and the codes
> corresponding to the start of the element are added after it. (In
> other words, any embedding levels or overrides started by the element
> are closed at the paragraph break and reopened on the other side of it.)
>
>
> What does that mean in terms of implementation?
>
> - that in the bidi computation, the state after X1-X6 processing of the
> paragraph before the paragraph boundary is used as the initial state for
> the X1-X6 processing of the paragraph after that boundary (really,
> overriding the second X1)? That state being: the current embedding level,
> the directional status override, the stack of underlying X1-X6.
>
> - that only some of that state is carried over? if so, how much?
>
> - the meaning is in terms of HTML/CSS only, something like:
>
> <span style='direction:d; unicode-bidi:u'>...<br>...</span>
>
>
> is equivalent to:
>
> <span style='direction:d; unicode-bidi:u'>...<span><br><span
> style='direction:d; unicode-bidi:u'>...</span>
>
>
> (with the property present on the second only if it is present on the
> first; and of course, with the similar treatment of relevant HTML elements
> attributes). If that is the case, what are the exact rules?
>
> - something else?
>
>
> Also, what is the eor of the last run of the paragraph before, and the sor
> of the first run of the paragraph after?
>
>
> Thanks,
> Eric.
>
>