Re: [css3-text] line break opportunities are based on *syllable* boundaries?

On 30/01/2011 2:39 PM, Ambrose LI wrote:
> 2011/1/29 Boris Zbarsky<>:

>> If it's marked up (<name>,<span class="name">, whatever), then it seems
>> that all that's needed on the CSS end is saying that "white-space:nowrap"
>> should prevent linebreaks even at non-whitespace breaking opportunities.
>>   Which is what UAs do in practice, and what the spec seems to say already.
>> Am I just missing something?
> I asked the same question a few years ago. Unfortunately, the problem
> is that all the attribute that SHOULD be able to do such things either
> do not act on CJK characters or do not act on anything that is not a
> block. As to why we have these restrictions, this is completely beyond
> my understanding.

Hello Ambrose and Boriz,

 From my work with rtl and block formatting, inline formatting, etc., 
etc., I now theorized that the problems we see with bidi reordering, 
run-ins, overflow, white-space, line breaks,etc. and how these are 
different among UA rendering engines is due to construction (nesting and 
arrays ?) of the code base of these engines.

Take this patch,

for this bug.

Each rendering engines has a different tree of C++. Initially, 
especially with Trident, there was a bias toward the writing-mode with 
left to right inline formatting and top to bottom block formatting. IE8 
corrected this and implemented multiple writing modes. Opera 11 (Presto) 
still chokes with RTL and WebKit a little less.

Considering all of this, we have a long way to go to deal with the full 
complexities of CJK or other scripts.

Somehow expecting a reply from Boriz saying I'm completely wrong. Such 
is life. :-)


Armies Cannot Stop An Idea Whose Time Has Come. - Victor Hugo

Received on Sunday, 30 January 2011 07:57:36 UTC