IPA characters from subprocesses

[dk2993]

  I've been having trouble getting characters from the IPA (phonetics)
character set, encoded as UTF-8, to display correctly.  I can type them
just fine; but if I feed them to a subprocess and ask for them back, the
result is displayed as a missing-character box.  I've tried this in a'
couple of different ways; the easiest way to be sure that the subprocess is
sending the correct character is to put it in a file and cat it under
M-x shell.

  I'm using a stock Debian install of Emacs 21.3, with no relevant
customizations in ~/.emacs.

  To reproduce the problem:

First, create the shell buffer and set up the various options.
  M-x shell RET
  C-x RET C-\ ipa RET C-\
  C-x RET p utf-8 RET utf-8 RET

Create the test file.
  echo SPACE C-\ E C-\ SPACE > SPACE /tmp/z RET

Look at it.
  cat SPACE /tmp/z RET

  The "echo" line should have a symbol on it which looks a little bit like
a lowercase e, with all curvy lines instead of straight.  In a perfect world,
the line output by cat should have the same symbol, but instead it has a
missing-character box.

  Here's the output of C-u C-x = for the correct symbol:

  character: ɛ (014243, 6307, 0x18a3)
    charset: ipa (IPA (International Phonetic Association))
 code point: 35
     syntax: word
   category:
buffer code: 0x9A 0xA1 0xA3
  file code: not encodable by coding system nil
       font: -ETL-Fixed-Medium-R-Normal--16-160-72-72-C-80-MuleIPA-1

  Here it is for the missing-char box:

  character: ɛ (01210733, 332251, 0x511db)
    charset: mule-unicode-0100-24ff
             (Unicode characters of the range U+0100..U+24FF.)
 code point: 35 91
     syntax: word
   category:
buffer code: 0x9C 0xF4 0xA3 0xDB
  file code: 0xC9 0x9B (encoded by coding system utf-8)
       font: -Adobe-Courier-Medium-R-Normal--17-120-100-100-M-100-ISO10646-1

  I think I understand what's going on, and if I'm right, the solution is a
trivial change to Emacs, letting it know that characters in the IPA range of
ISO 10646 are part of the IPA charset?

  Please advise.

  Just incidentally - this is quite unrelated - this mail will be the first
I've sent in utf-8.  I'm curious whether other people will receive it
properly.  Grin.