Confusing font selection with UTF-8 coding

[Bulent Murtezaoglu]

I am having an odd problem with Emacs 21.3.  When I save some Turkish text 
with coding iso-8859-9-unix, things are fine but when I use UTF-8 some 
characters end up using a font I don't want.  Here's an example with 
'LATIN SMALL LETTER S WITH CEDILLA' (0xFE in iso-8859-9, 0x015F in Unicode
0xC59F in UTF-8).  I get the below with C-u C-x =

With coding: iso-8859-9-unix (where it is displayed correctly):

  character: ş (07376, 3838, 0xefe)
    charset: latin-iso8859-9
             (Right-Hand Part of Latin Alphabet 5 (ISO/IEC 8859-9): ISO-IR-148)
 code point: 126
     syntax: word
   category: l:Latin  
buffer code: 0x8D 0xFE
  file code: FE (encoded by coding system iso-latin-5-unix)
       font:
             -bitstream-bitstream vera sans 
mono-medium-r-normal--15-83-130-130-m-90-iso8859-9



With coding: UTF-8 (where it is displayed smaller an in a different font):

  character: ş (04472, 2362, 0x93a)
    charset: latin-iso8859-2
             (Right-Hand Part of Latin Alphabet 2 (ISO/IEC 8859-2): ISO-IR-101)
 code point: 58
     syntax: word
   category: l:Latin  
buffer code: 0x82 0xBA
  file code: 0xC5 0x9F (encoded by coding system utf-8-unix)
       font: -jmk-Neep Alt-Medium-R-Normal--15-140-75-75-C-80-ISO8859-2


What am I doing wrong?  How do I debug this?  Where is latin-iso8859-2 
coming from?  The default font _does_ have the right glyph (?) at the 
proper unicode code point (I double checked with pfaedit).

thanks,

BM