From: Troy Louden (firstname.lastname@example.org)
Date: Fri May 15 2009 - 13:26:16 CDT
I'm implementing the Unicode Bidi algorithm (http://www.unicode.org/reports/tr9/) for a project at work and I don't understand why, according to the spec, Arabic Numbers are displayed Left-to-Right instead of Right-to-Left.
I've gone through the spec and, as far as I can see, the only explicit instructions for assigning direction to Arabic numbers (class AN) is in 3.3.5 table 5. Here AN is moved to the next even level making it LTR.
European numbers do the same thing even though, by this time, they should also resolve RTL sine any LTR EN will already be resolved with the other weak types.
Example if you have the text:
"abc \u+0x668\u+0x638\u+0x62d ABC" (which resolves to classes: [L][L][L][WS][AN][AL][AL][WS][L] [L] [L])
Should it be displayed:
"abc \u+0x62d\u+0x638\u+0x668 ABC"
"abc \u+0x668\u+0x62d\u+0x638 ABC"
Common sense and even your demo app says the first result is correct, but your spec says the second is correct
Am I missing something here?
This archive was generated by hypermail 2.1.5 : Fri May 15 2009 - 14:00:13 CDT