(Resending as I forgot this old Bugzilla bug with non-BMP characters...)

-----------------

I'm opening this bug for reference, as I believe this will require WebKit/Gecko to expose the mathvariant property in the accessible tree.

Currently, Orca reads

data:text/html;charset=UTF-8,<math><mi>&#x1D504;</mi></math>

as "fraktur A".

However, MathML has an equivalent way to do that, using the mathvariant attribute (used by MathJax, Wikipedia and many LaTeX-to-MathML converters):

data:text/html;charset=UTF-8,<math><mi mathvariant="fraktur">A</mi></math>

or even

data:text/html;charset=UTF-8,<math><mstyle mathvariant="fraktur"><mrow><mrow><mi>A</mi></mrow></mrow></math>

In these cases, Orca just reads "A".

Some remarks:

* Internally, in Gecko the mathvariant MathML attribute is mapped to an internal CSS property "-moz-mathvariant" so I believe the obvious solution would be to map this CSS property to a text attribute, just like we do for font-family etc:

https://dxr.mozilla.org/mozilla-central/source/accessible/base/TextAttrs.h#289

* mathvariant is not supported by WebKit (https://bugs.webkit.org/show_bug.cgi?id=85735)

Marking as "blocked" because I cannot add support in Orca for things which Orca has no way of knowing.

Closing blocked bugs because we need the issue fixed in the blocker.