colldef(1) BSD General Commands Manual colldef(1)
NAME
colldef -- convert collation sequence source definition
SYNOPSIS
colldef [-I map_dir] [-o out_file] [filename]
DESCRIPTION
The colldef utility converts a collation sequence source definition into
a format usable by the strxfrm() and strcoll() functions. It is used to
define the many ways in which strings can be ordered and collated. The
strxfrm() function transforms its first argument and places the result in
its second argument. The transformed string is such that it can be cor-
rectly ordered with other transformed strings by using strcmp(),
strncmp(), or memcmp(). The strcoll() function transforms its arguments
and does a comparison.
The colldef utility reads the collation sequence source definition from
the standard input and stores the converted definition in filename. The
output file produced contains the database with collating sequence infor-
mation in a form usable by system commands and routines.
The following options are available:
-I map_dir
Set directory name where charmap files can be found, current
directory by default.
-o out_file
Set output file name, LC_COLLATE by default.
The collation sequence definition specifies a set of collating elements
and the rules defining how strings containing these should be ordered.
This is most useful for different language definitions.
The specification file can consist of three statements: charmap,
substitute and order.
Of these, only the order statement is required. When charmap or
substitute is supplied, these statements must be ordered as above. Any
statements after the order statement are ignored.
Lines in the specification file beginning with a `#' are treated as com-
ments and are ignored. Blank lines are also ignored.
charmap charmapfile
Charmap defines where a mapping of the character and collating element
symbols to the actual character encoding can be found.
The format of charmapfile is shown below. Symbol names are separated
from their values by TAB or SPACE characters. Symbol-value can be speci-
fied in a hexadecimal (\x??) or octal (\???) representation, and can be
only one character in length.
symbol-name1 symbol-value1
symbol-name2 symbol-value2
...
Symbol names cannot be specified in substitute fields.
The charmap statement is optional.
substitute "symbol" with "repl_string"
The substitute statement substitutes the character symbol with the string
repl_string. Symbol names cannot be specified in repl_string field. The
substitute statement is optional.
order order_list
Order_list is a list of symbols, separated by semi colons, that defines
the collating sequence. The special symbol ... specifies, in a short-
hand form, symbols that are sequential in machine code order.
An order list element can be represented in any one of the following
ways:
o The symbol itself (for example, a for the lower-case letter a).
o The symbol in octal representation (for example, \141 for the letter
a).
o The symbol in hexadecimal representation (for example, \x61 for the
letter a).
o The symbol name as defined in the charmap file (for example,
<letterA> for letterA \023 record in charmapfile). If character map
name have > character, it must be escaped as />, single / must be
escaped as //.
o Symbols \a, \b, \f, \n, \r, \v are permitted in its usual C-language
meaning.
o The symbol chain (for example: abc, <letterA><letterB>c, \xf1b\xf2)
o The symbol range (for example, a;...;z).
o Comma-separated symbols, ranges and chains enclosed in parenthesis
(for example ( sym1, sym2, ... )) are assigned the same primary
ordering but different secondary ordering.
o Comma-separated symbols, ranges and chains enclosed in curly brackets
(for example { sym1, sym2, ... }) are assigned the same primary
ordering only.
The backslash character \ is used for continuation. In this case, no
characters are permitted after the backslash character.
DIAGNOSTICS
The colldef utility exits with the following values:
0 No errors were found and the output was successfully created.
!=0 Errors were found.
FILES
/usr/share/locale/<language>/LC_COLLATE
The standard shared location for collation orders under the
locale <language>.
SEE ALSO
mklocale(1), setlocale(3), strcoll(3), strxfrm(3)
BSD January 27, 1995 BSD
Mac OS X 10.7 - Generated Thu Aug 11 05:56:54 CDT 2011