Re: How to recognize keyboard insertion?

[David De La Harpe Golden]

Eli Zaretskii wrote:


> Maybe so, but UAX#9 treats European digits and Arabic digits
> differently.  They have different bidirectional properties.

The eastern arabic digits U+06F0 to U+06F9 and usual western digits 
U+0030 to U+0039 do have separate code point ranges.  Apparently
the former code points are less used - the arabic keyboard layout 
returns the western codes for the 0-9 keypresses.

gtk+/pango seems to be choosing to merely _display_ the latter with the 
glyphs of the former depending on surrounding language.
Based on post dug up here http://markmail.org/message/72on34u7nupadioh
they might be availing of a "higher level protocol" freedom granted
to them. Or something.