FreeTDS Development Group

["James K. Lowden" <jklowden AT schemamania.org>]

On Thu, 15 May 2003 11:57:58 +0800, "Varley, David (CBorn at Alcoa)"
<David.Varley AT alcoa.com.au> wrote:
> Jim,
> 
> From what I can see, the conversions are:
> 
> ISO8859-1-GL_ISO8859-1
> ISO8859-1-GR_ISO8859-1
> ISO8859-1_ISO8859-1-GL
> ISO8859-1_ct
> 
> with aliases by symbolic links as:
> 
> ct_ISO8859-1ASCII-GR_ISO8859-1 -> ISO8859-1_ISO8859-1-GL
> ISO8859-1_ASCII-GR -> ISO8859-1-GL_ISO8859-1
> ISO8859-1_ISO8859-1-GR -> ISO8859-1-GR_ISO8859-1
> ISO8859-1_fold7 -> ISO8859-1_ct
> ISO8859-1_fold8 -> ISO8859-1_ct
> fold7_ISO8859-1 -> ct_ISO8859-1
> fold8_ISO8859-1 -> ct_ISO8859-1
> 
> if this is of any help. My version of iconv doesn't support the
> --list option of the GNU one, I've just taken these from the
> /usr/lib/nls/loc/iconv directory (algorithmic converters).

It's ~some~ help.  What I need to know is, what encodings those are.  For
lack of a better term, I need their GNU names.  I found this [1]:

  [Tru64 UNIX]  Current industry standards do not specify the name format,
  implementation, and location of the codeset converters used by this com-
  mand.  On this operating system, the command searches directories of
algo-
  rithmic and table converters to find one that performs the requested
  conversion. The names of codeset converters adhere to the following
format:

  fromcode_tocode

I have to think your support is broader than those names, too, btw.  You
mentioned "/usr/lib/nls/loc/iconv/".  What about table converters in
/usr/lib/nls/loc/iconvTable ?

The work of establishing the meaning of all those conversions may well be
more than you care to take on.  That's fine.  What we need at least is
Frediano's favorite four, plus any you require.  (From what I understand,
you require ASCII and nothing else, yea?)

FFF are:

        ISO-8859-1
        UTF-8
        UCS-2LE
        UCS-2BE

Freddy, Tru64 introduces a couple of other things to noodle around.  It
uses environment variables to control iconv behavior.  These might just be
documentation notes; I don't know.  They are:

  ICONV_BYTEORDER
      Sets byte ordering for UCS-2 or UCS-4 converters

  ICONV_DEFSTR
      If these variables are not defined or are set to the null string,
the
      characters that cannot be converted are skipped and have no
representa-
      tion in converted output.  [Not GNU's behavior!]

  ICONV_NOBOM
      Disables generation of the byte-order mark at the beginning of UCS-2
or
      UCS-4 output. 

--jkl
[1] http://btrcx1.cip.uni-bayreuth.de/cgi-bin/manpages/iconv/1