Name: js5519 Date: 02/09/99


When numerals are in an Arabic context, for example when they are
surrounded by Arabic letters, they should have Hindi shapes (Unicode
values from \u0660 to \u0669). Currently the BIDI algorithm always
sets numeral shapes to the Arabic (Roman) shapes (Unicode values
\u0030 to \u0039). Shaping the numbers is the responsibility of the
BIDI algorithm as specified by the Unicode standard.
Note that shaping the numbers should only happen in Arabic blocks
and not in Hebrew blocks, since Hebrew always uses the Roman numerals.

This is very important because the Hindi numerals are the only numerals
known in most of the Arab countries, especially in the Gulf region.

I suggest that there should be an attribute in the TextAtribute class
as follows:
TextAttribute.NUMERALS_SHAPE
and it could be set to:
TextAttribute.NUMERALS_SHAPE_ROMAN //numerals are always Roman
TextAttribute.NUMERALS_SHAPE_HINDI //numerals are always Hindi
TextAttribute.NUMERALS_SHAPE_CONTEXT //the BIDI algorithm will shape the numerals depending on the context they're in
(Review ID: 53978)
======================================================================