Changing the Unicode data in ICU?

[Shachar Shemesh]

Hi all,

I am trying to recompile ICU so that it has a different BiDi 
classification for some characters. ICU contains some of the Unicode 
Character Database files, so I tried simply changing the classification 
there and recompiling. Some files changed, but the end result did not do 
what I expected it to do.

I've tried to follow the code and see where it leads me, but so far, it 
seems like nothing actually specifies what BiDi class each character is 
(which is nonsense, because ICU has to know that).

Anyone ever tried anything similar?

Shachar

-- 
Shachar Shemesh
Lingnu Open Source Consulting Ltd.
http://www.lingnu.com