Re: [css-fonts] Conflicting tags in font-feature-settings

I also agree. This proposal is both undesirable and impossible.

—Myles

> On Mar 16, 2016, at 10:10 AM, Tab Atkins Jr. <jackalmage@gmail.com> wrote:
> 
> On Wed, Mar 16, 2016 at 5:02 AM, Jonathan Kew <jfkthame@gmail.com <mailto:jfkthame@gmail.com>> wrote:
>> On 16/3/16 00:12, Xidorn Quan wrote:
>>> 
>>> On Wed, Mar 16, 2016 at 7:59 AM, Myles C. Maxfield <mmaxfield@apple.com
>>> <mailto:mmaxfield@apple.com>> wrote:
>>> 
>>> 
>>>> On Mar 14, 2016, at 7:03 PM, Xidorn Quan <quanxunzhen@gmail.com
>>> <mailto:quanxunzhen@gmail.com>> wrote:
>>>> 
>>>> Hi,
>>>> 
>>>> It seems to me that the spec is not clear about what should happen
>>> if conflicting tags are specified for font-feature-settings. For example, if
>>> you set "font-feature-settings: 'hwid', 'twid', 'qwid'", only one of them
>>> can be chosen, but it is not clear which would win in that case.
>>> 
>>>    Is there a repository somewhere listing every combination of
>>>    conflicting font features?
>>> 
>>> 
>>> I believe at least most of variants conflict with each other. Not sure
>>> about other features.
>>> 
>> 
>> IMO, this is out of scope for CSS-Fonts. The font-feature-settings property
>> simply specifies some OpenType feature tags to be set.
>> 
>> What happens from there on -- and in particular, how "competing" features
>> interact -- is a matter for the OpenType spec,[1] and in most cases (except
>> where script-specific shaping specifications mandate a particular order of
>> feature application) it's the font designer who determines the interaction
>> or relative priority of the features, based on the ordering of lookups
>> within the font.
>> 
>> So for example, if the style calls for
>> 
>>  font-feature-settings: 'hwid', 'twid', 'qwid';
>> 
>> all of these feature tags will be set; which glyphs will actually result is
>> dependent on how the font designer orders the relevant lookups within the
>> font.
> 
> Strongly agree.  Authors shouldn't probably *shouldn't* set
> conflicting things, but there's nothing inconsistent or undefined
> about doing so.  Nothing needs to be done on our end.
> 
> ~TJ