NAME

chrtbl, wchrtbl − generate character classification and conversion tables

SYNOPSIS

chrtbl [file]
wchrtbl [file]

DESCRIPTION

chrtbl creates character type and numeric layout files for single byte locales.  wchrtbl does the same for multibyte locales.  The two commands are links to each other. 

Character classification tables contain information on character attributes, upper- to lowercase conversion, and codeset character width.  The LC_CTYPE file is an array of bytes encoded so simple table lookups can determine character type or perform case mapping, using ctype(3C) or wctype(3I) library routines.  Other routines can find the byte count and screen width of characters in supplementary code sets.  The LC_NUMERIC file contains format information for numbers.  The first byte specifies the decimal delimiter, and the second byte specifies the thousands separator. 

Both commands read character classification and conversion information from file and create three output files in the current directory.  If no input file is given, these commands read from standard input.  The example section below contains the source file for ISO 8859-1.  For multibyte locales this example needs to be extended. 

The first output file, [w]ctype.c is a C language source file, which application programs can use as needed.  It contains a (257∗2)+7 byte array generated from processing file. Review the contents of the C source to verify that the array is set up as planned. The first 257 bytes of the array are used for character classification. Symbols used for initializing these bytes represent character classifications defined in <ctype.h>; for example, _L means a character is lower case and _S|_B means the character is both a spacing character and a blank.  The second 257 bytes of the array are used for character conversion.  These bytes are initialized so that characters without conversion information are converted to themselves.  If you provide conversion information, the first value of the pair is stored where the second one would normally be stored, and vice versa.  For example, if you provide <0x41 0x61>, then 0x61 is stored where 0x41 would normally be stored, and 0x61 is stored where 0x41 would normally be stored.  The last 7 bytes are used for character width information for up to three supplementary code sets. 

The second output file is binary data containing the same information, but structured for efficient use by the ctype(3C) and wctype(3I) routines.  The name of this output file is the value you assign to the keyword LC_CTYPE in file. The superuser should install this file as /usr/lib/locale/locale/LC_CTYPE/ctype.  It must be readable by user, group, and other; execute permission is not necessary.  Application programs consult this file when the LC_CTYPE environment is set appropriately, upon calling setlocale(3C). 

The third output file is binary data created only if numeric formatting information is specified.  The name of this output file is the value you assign to the keyword LC_NUMERIC in file. The superuser should install this file as /usr/lib/locale/locale/LC_NUMERIC.  It must be readable by user, group, and other; execute permission is not necessary.  Application programs consult this file when the LC_NUMERIC environment is set appropriately, upon calling setlocale(3C). 

For supplementary codesets, there are three sets of tables.  The first set contains three pointer arrays that point to supplementary codeset information tables.  If supplementary codeset information is not specified, the contents of the pointers are zeros.  The arrays are full of null pointers.  The second set contains three supplementary codeset information tables, each specifying minimum and maximum code values to be classified and converted, and also pointers to character classification and conversion tables.  If there is no corresponding table, the contents of the pointers are zeros.  The third set contains character classification and conversion tables that contain the same information as the single byte table, except codes are represented as process codes and table size is variable.  The characters used for initializing values of the character classification table represent character classifications defined in <wctype.h>; _E1 through _E8 are for international use and _E9 through _E24 are for language-dependent use. 

The syntax of file provides for data file naming, assignment of characters to character classifications, upper- to lower-case mapping, byte and screen widths for up to three supplementary code sets, plus numeric formatting information.  The keywords recognized by [w]chrtbl are:

LC_CTYPE name of the first data file created by [w]chrtbl

isupper character codes classified as upper-case letters

islower character codes classified as lower-case letters

isdigit character codes classified as numeric

isspace character codes classified as spacing (delimiter) characters

ispunct character codes classified as punctuation characters

iscntrl character codes classified as control characters

isblank character code for the blank (space) character

isxdigit character codes classified as hexadecimal digits

ul relationship between upper- and lower-case characters

cswidth byte count and screen width information

LC_NUMERIC name of the second data file created by [w]chrtbl

decimal_point decimal delimiter, may be \NNN octal or \xNN hexadecimal

thousands_sep thousands separator, may be \NNN octal or \xNN hexadecimal

LC_CTYPE1 begin definition of supplementary codeset 1

LC_CTYPE2 begin definition of supplementary codeset 2

LC_CTYPE3 begin definition of supplementary codeset 3

isphonogram(iswchar1) character codes classified as phonograms in supplementary code sets

isideogram(iswchar2) character codes classified as ideograms in supplementary code sets

isenglish(iswchar3) character codes classified as English letters in supplementary code sets

isnumber(iswchar4) character codes classified as numeric in supplementary code sets

isspecial(iswchar5) character codes classified as special letters in supplementary code sets

iswchar6 character codes classified as other printable letters in supplementary code sets

iswchar7 - iswchar8 reserved for international use

iswchar9 - iswchar24 character codes classified as language-dependent letters/characters

Any lines with a sharp (#) in the first column are treated as comments and are ignored, as are blank lines. 

To indicate character codes, use either hexadecimal or octal constants.  For example, the letter a can be represented as 0x61 in hexadecimal or 0141 in octal.  Constants may be separated by one or more spaces and/or tabs.  Use a dash (−) to indicate a range of consecutive numbers.  Zero or more spaces may separate the dash from its numbers.  Use a backslash (\) for line continuation; only the newline is permitted after a backslash. 

The relationship between upper- and lower-case letters (ul) is expressed as ordered pairs of octal or hexadecimal constants: <upper-case_character lower-case_character>.  One or more space characters may separate these two constants.  Zero or more space may separate angle brackets (<>) from numbers.

The following is the format of an input specification for cswidth: n1:s1,n2:s2,n3:s3

where

n1byte width for supplementary code set 1, required
s1screen width for supplementary code set 1
n2byte width for supplementary code set 2
s2screen width for supplementary code set 2
n3byte width for supplementary code set 3
s3screen width for supplementary code set 3

decimal_point and thousands_sep are specified by a single character, the delimiter. 

EXAMPLE

Here is the input file used to create the iso_8859_1 codeset definition table. 

FILES

/usr/lib/locale/locale/LC_CTYPE/ctype
data file containing character classification, conversion, and codeset width information

/usr/lib/locale/locale/LC_NUMERIC
data file containing numeric layout information

/usr/include/ctype.h declarations used by character classification and conversion routines

/usr/include/wctype.h declarations used by wide character classification and conversion routines

SEE ALSO

ctype(3C), setlocale(3C), environ(5)

NOTES

Do not change files under the C locale, as this could cause undefined or nonstandard behavior. 

SunOS 5.1/SPARC  —  Last change: 24 Feb 1992