For a simple character set named
MYSET, sorting rules are specified in
the
configuration file using MYSET.xml<map> array
elements within <collation> elements.
If the sorting rules for your language are too complex to be
handled with simple arrays, you must define string collating
functions in the
ctype-
source file in the MYSET.cstrings directory.
The existing character sets provide the best documentation and
examples to show how these functions are implemented. Look at
the ctype-*.c files in the
strings directory, such as the files for
the big5, czech,
gbk, sjis, and
tis160 character sets. Take a look at the
MY_COLLATION_HANDLER structures to see how
they are used. See also the
CHARSET_INFO.txt file in the
strings directory for additional
information.