Dan Kogai <dankogai(_at_)dan(_dot_)co(_dot_)jp> writes:
On Wednesday, May 21, 2003, at 03:21 PM, Jarkko Hietaniemi wrote:
I agree that there should be no reason do convert the ISO 8859-6
decimal digits to the Unicode Arabic-Indic digits. I don't know where
this curious data might actually be coming from, since most of our
data comes from the Unicode Consortium's web site's legacy mapping
data, and there it's simple a business of 0x31 -> 0x00031. Thanks for
the catch.
I wonder where they came from, too. I think it slipped in when Encode
switched from Tcl-based .enc to .ucm. I will check ICU and regen
ISO-8859-X and release Encode 1.95 as soon as I can because I consider
this rather severe.
Hmm - Tcl/Tk might have done that deliberately as its primary use of
encodings is for font glyph lookup. So it may have wanted to use the
Arabic-Indic codepoints to find those glyphs.
(Though as Tk does not do BIDI or deal with Arabic initial/medial/final
presentation forms the value of having arabic digit glyphs in a sea of
mis-presented characters seems marginal.)
--
Nick Ing-Simmons
http://www.ni-s.u-net.com/