dummy-line-for-utf8-textfile-start-character.

-- This is a script documenting a strange problem I encountered with Arabic:
-- I am working on PC / Windows XP with MySQL 4.1.10a - just updated from 4.1.7:
-- ONE letter (Feh, Unicode 0641) seems to get corrupted when entered in a table, 
-- while all the other Arabic letters I used were OK.

use test;
tee outfile.txt;

drop table if exists ALetters;

create table ALetters (
  letter VARCHAR(2) CHARACTER SET UTF8, 
  name VARCHAR(5) CHARACTER SET UTF8);

-- entering 3 different letters for comparison:

INSERT INTO ALetters VALUES ('ب','ba'),('ف','fa'),('ك','kaf');
SELECT * FROM ALetters;

-- seems OK so far, BUT the letter 'fa' is incorrectly rendered in the table (column letter),
-- It comes out as a ? in the outfile.txt (if the textfile is encoded as utf8), and
-- on my html page using perl/CGI/DBI it is displayed as another letter (Tehe: ٿ, Unicode 067F)? 
-- But the letter comes out correctly when using the following statements:

select hex('ف');
select unhex('d981');
select unhex(hex('ف'));

-- select unhex(hex('ب'));
-- select unhex(hex('ف'));
-- select unhex(hex('ك'));

-- BUT when using the same statement with values from the table,
-- the letter 'fa' is of course corrupted as before:

INSERT INTO ALetters select unhex(hex(letter)), name from Aletters;
SELECT * FROM ALetters;

-- WHY is the letter 'fa' corrupted? How do I put it right?