Re: [css-ruby] What does it mean for "ruby-position: inter-character" to force writing-mode to be vertical?

On Mon, Dec 15, 2014 at 11:00 AM, Daniel Holbert <dholbert@mozilla.com>
wrote:
>
> On 12/15/2014 03:22 AM, Xidorn Quan wrote:
> > On Thu, Dec 11, 2014 at 9:36 AM, Daniel Holbert <dholbert@mozilla.com
> > <mailto:dholbert@mozilla.com>> wrote:
> >      (1) Does this spec-text influence the *computed value* of the
> >     'writing-mode' property? (I hope not; there's added complexity when
> >     properties influence other properties' computed values on the same
> >     element.)
> >
> > Actually, I hope the answer is yes. I'm not quite sure which way is
> > better, but we have had some style fixup like this, and I don't think
> > it's a big problem to add one more rule there. The only problem might be
> > that it would add memory footprint.
> >
> > I feel that, if we don't do this, we need to maintain another path for
> > this value, and handle inheritance ourselves.
>
> That's fair.  Behind my "I hope not", I really just meant that I want to
> avoid adding channels of inter-property influence, if they aren't
> actually intended/needed, because they add complexity (and foil memory
> optimizations, at least in Gecko, as you indicated). But, it sounds like
> the influence may be both intended & needed here -- which is fine, as
> long as it's clearly explained.
>
> In this case, the desired inter-property influence seems to be:
>    For any element with a computed "display" of
>    "ruby-text-container" *and* a computed "ruby-position"
>    of "inter-character", the UA must force the computed
>    "writing-mode" to be $SOME_VERTICAL_WRITING_MODE.
>

This complexity could easily be avoided if the spec says "if ruby-position
is inter-character, and writing-mode is not a vertical mode, then treat
ruby-position as if the value 'initial' were specified".

There is no reason ruby position should be forcing writing mode.


>
> >      (2) If the answer to (1) is "yes" (I hope not): is this
> "writing-mode"
> >     computed-value influence restricted to elements with "display:
> >     ruby-text",
> [...]
> >
> > If that is style fixup, then it certainly only influences elements with
> > "display: ruby-text-container", since ruby-position is inherited by
> > default, and author may specify it in an outer container.
>
> (Right, sorry - I meant "ruby-text-container", not "ruby-text". Thanks
> for the correction on that.)
>
> Like you, I think the answer to question (2) should really be "yes".  My
> point in asking it was to highlight that this needs clarification in the
> spec.
>
> Thanks,
> ~Daniel
>
>