the big character set change

[samba-technical at thewrittenword.com]

On Wed, Jul 04, 2001 at 05:23:27AM -0700, Andrew Tridgell wrote:
> If your OS does not have the iconv() calls in its C library then the
> only character sets you will have available are "ASCII", "UCS2" and
> the test character set "WEIRD". If you do have the iconv library then
> all of your system character sets defined in your iconv system will be
> available. On Linux you can list these character sets using "iconv --list"
> On my system I have 892 character sets available although many of
> those are aliases and even more of them are useless as they don't obey
> the character set rules I outlined in my last email.
> 
> I may end up building UTF8 into Samba so those of you with broken or
> old OSes can still use a good character set, but meanwhile you will
> have to set "dos charset = ASCII" and "unix charset = ASCII" if you
> don't have iconv(). Samba will detect whether you have iconv() at
> compile time using autoconf. Look for HAVE_NATIVE_ICONV in
> include/config.h for the result of this test.

How about allowing use of GNU libiconv on systems where iconv is not
good enough or it doesn't exist:
  http://clisp.cons.org/~haible/packages-libiconv.html

Gettext 0.10.38 has an autoconf test, AM_ICONV, that adds
--with-libiconv-prefix=DIR to search for libiconv in DIR/include and
DIR/lib. This might be a useful addition.

-- 
albert chin (china at thewrittenword.com)