Re: [css-text] vertical-align and shaping (again)

fantasai wrote:

>>> - RESOLVED: Text shaping MUST be broken across inline element
>>>              boundaries when:
>>>    A. one of margin/border/padding are non-zero
>>>    B. vertical-align is not 'baseline'
>>>    C. it is a bidi isolation boundary
>>>     Text shaping MUST NOT be broken across inline element
>>>     boundaries when there is no change in formatting.
>>>     Text shaping SHOULD NOT be broken across inline
>>>     element boundaries otherwise, if it is reasonable and
>>>     possible for that case given the limitations of the
>>>     font technology.
>>>     http://lists.w3.org/Archives/Public/www-style/2014Aug/0295.html
>>
>> I should note again that I think (B) is misguided and it should not
>> be viewed as consensus opinion. I doubt any implementor will actually
>> implement this so we should remove it.
>>
>> There was a small thread pointing this out [2], including Simon's comment [3]:
>>
>>    There are all sorts of cases where you don't want Arabic
>>    letters to join, e.g. acronyms like IBM, and I don't think
>>    it's a good idea to multiply cases where it gets triggered
>>    by other properties.
> 
> Right now we have implementations that break joining in this case
> (because they break joining across any inline boundary) and
> implementations that don't break shaping even when the baseline is
> vastly different between the two letters to be joined.
> 
> Which behavior are we going to require? Since they're both clearly
> implementable, this isn't a case that should be non-interoperable.
> 
> I'm not opposed to making it required to join, but while I know of
> use cases where the joining should break, nobody's presented one
> where the joining should not break.

Breaking when the value of vertical-align is different is fine but
simply breaking when it's not 'baseline' doesn't make sense. Shaping
should be the same when two adjacent inline elements have the same
value of vertical-align.

As for whether to require coalescing or not, I'm not sure we need to
require things one way or another.

Cheers,

John Daggett
Mozilla Japan

Received on Monday, 27 October 2014 05:41:58 UTC