Open System Services Shell and Utilities Reference Manual (G06.29+, H06.08+, J06.03+)

genxlt(1) OSS Shell and Utilities Reference Manual
as non-identical conversions.
There is no requirement that the target_sub character for a conversion from a source code set to
a target code set be the source_sub character in a table that specifies the inverse conversion.
Comments can contain any characters, but it is recommended that only characters in the ASCII
code set be used. Using non-ASCII characters might cause genxlt to report an error if it finds
characters that are invalid in the current locale.
EXAMPLES
The following is an excerpt of a code set conversion table:
#source_sub 0x1a # SUB (default source_sub)
target_sub 0xffff Replacement Character
0x0 0x00 NULL
0x1 0x01 SOH
0x001a 0x1A SUB
0x80 0xc7 C cedilla
0x81 0xfc u diaeresis
0x82 0xe9 e acute
0x83 0xe2 a circumflex
0x84 0xe4 a diaeresis
0x85 invalid not in use
0x86 0x40 a grave
0xa1a1 0xa2a3 special
0xa1a2 0xa2a4 special
0xa1a3 0xa2a5 special
0xa1a4 0xa2a6 special
In the preceding example, a comment serves only as a reminder that the default source_sub
value is 0x1a. The target_sub is explicitly specified. The input values 0x1a and 0x001a are the
same input value; input values can be padded with leading NULL bytes. The table indicates that
the input character corresponding to the source_sub character is translated to the character SUB
(0x1A), instead of to the target_sub character. It is not required that the source_sub character
be translated to the target_sub character.
Note that the source value 0x85 in the preceding example is specifically declared invalid in the
target column. Source values between 0x86 and 0xa1a1 are implicitly declared valid by their
omission from the table. The omitted characters are converted to the target_sub character.
The name of the file generated by the genxlt utility must adhere to the following naming con-
vention for the iconv subsystem to recognize it as a conversion-table file:
fromcode: "ISO8859-1-GL"
tocode: "ISO8859-1"
conversion table file: "ISO8859-1-GL_ISO8859-1"
The code-set conversion table file name is formed by concatenating the to_code file name onto
the from_code file name, with an underscore character between the two.
Input code sets representable by the tables are restricted to single-byte and double-byte code sets.
It is not possible to represent a multibyte input code set, such as SJIS. Conversion of multibyte
input code sets to other code sets can be done by the iconv command only, using iconv convert-
ers.
4−8 Hewlett-Packard Company 527188-021