[css-text-decor-3] Processing model for text-underline-position: under | left | right

http://dev.w3.org/csswg/css-text-decor-3/#text-underline-position-property
says:

  # ‘under’
  #   In horizontal writing modes, the underline is positioned
  #   relative to the under edge of the element's content box. In
  #   this case the underline usually does not cross the descenders.
  #   (This is sometimes called "accounting" underline.) If the
  #   underline affects descendants with a lower content edge, the
  #   user agent should shift the underline down further to the
  #   lowest underlined content box edge. The user agent may ignore
  #   elements with ‘vertical-align’ values given as lengths,
  #   percentages, ‘top’, or ‘bottom’ when making this adjustment.
  #   (Note that images that are not affected by the underline per
  #   ‘text-decoration-skip’ will not affect the position of the
  #   underline.) 

and similar for 'left' and 'right'.


It's not clear to me how these rules integrate with the processing
model for text decorations defined in
http://dev.w3.org/csswg/css-text-decor-3/#line-decoration

I expect the specification to define a processing model that clearly
answers questions such as:

 * Does relative positioning of descendants affect the resulting
   position?  (Answer:  of course not!)

 * Are descendants within the entire element considered, or is the
   rule applied line-by-line?  (If the latter, are descendant blocks
   to which the underline is propagated considered?)

 * Are elements to which the underline does not apply because they
   have 'visibility: hidden' still considered when positioning the
   underline?  (Answer:  of course!)

 * Is overflow from atomic inline descendants considered?  (Answer:
   no, I hope.)

 * Are content box edges inside atomic inline descendants
   considered?  (Answer:  No!)

(All but the second of these questions relate to the vagueness of
"affects descendants".   However, I think it's important to actually
integrate this behavior into the processing model rather than try to
define it one question at a time.)

-David

-- 
𝄞   L. David Baron                         http://dbaron.org/   𝄂
𝄢   Mozilla                           http://www.mozilla.org/   𝄂

Received on Monday, 25 March 2013 04:09:28 UTC