The settings in the configuration file NATCONV.INI apply to the A format. For the U format, the ICU library is used.
This document describes how Natural supports different character sets. It covers the following topics:
The support of multiple languages with different character sets represents Natural's approach towards internationalization. It can help you when using:
terminals and printers with different character sets, all communicating with the same Natural environment;
several Natural environments sharing one database and located on different platforms;
upper-/lower-case translation of language-specific characters;
language-specific characters in Natural identifiers, object names and library names;
language-specific characters in an operand compared with a mask definition (see MASK Option in the Programming Guide).
Natural supports any single-byte character set that conforms to the ASCII character set in the lowest seven bits.
Natural distinguishes between an internal character and several external character sets; the internal character set is used by Natural itself.
As illustrated below, conversion between the internal and an external character set is performed after the input from a terminal and before the output to a terminal or printer. There is no conversion to an external character set available for work file I/Os, database I/Os and reading/writing of Natural objects.
By default, Natural uses the internal character set ISO8859_1. If the default character set does not meet your requirements, you can choose either one of the predefined character sets provided by Natural or any other standard character set.
Note:
Problems may occur if you run computers with different
internal character sets sharing the same database, or if you try to exchange
data or Natural objects between such computers.
You can define an external character set for any terminal and printer.
For a terminal, the name of its character set is defined by the TCS entry in the terminal database, for example:
:TCS = usascii:
You can also use the UNIX environment variable
$NATTCHARSET which overrides all TCS settings.
If neither a TCS entry nor the logical
NATTCHARSET (which is set with the environment variable
$NATTCHARSET) is defined, no conversion is
performed during terminal I/O.
For a printer, the name of an external character set name can be defined in the printer profile. This is part of the global configuration file. See Printer Profiles in the Overview of Configuration File Parameters of the Configuration Utility documentation.
All check, translation and classification tables used by Natural to support language-specific characters reside in the configuration file NATCONV.INI. By default, this file is located in Natural's etc directory.
You can modify NATCONV.INI to support local or application-specific character sets.
In a standard application, NATCONV.INI need not and should not be modified, because this could lead to serious inconsistencies, in particular if Natural objects and database data are already present.
Modifications are necessary if you want to do any of the following:
use an internal character set other than the default one,
use a terminal or printer whose character set is not supported by NATCONV.INI,
allow or disallow the use of certain characters in identifiers,
support local characters when evaluating the
MASK
option.
Any modifications of NATCONV.INI should be well considered and carefully performed, otherwise problems might occur that are difficult to locate.
NATCONV.INI is subdivided in sections and subsections. The following sections are defined:
| Section | Description |
|---|---|
CHARACTERSET-DEFINITION |
This section defines the name of the internal
character set. The default is ISO8859_1.
If you choose a different character set, subsections for this character set must be contained in the sections described below. |
CHARACTERSET-TRANSLATION |
This section contains the tables
required for the conversion between the internal character set and external
character sets.
If you use, for example, a terminal with an entry in
|
CASE-TRANSLATION |
This section contains the tables required for the
conversion from upper-case to lower-case which is performed when one of the
following is specified:
|
IDENTIFIER-VALIDATION
|
This section contains the tables required for the
validation of identifiers (that is, user-defined variables in source programs),
object names and library names. It contains a subsection for each defined
internal character set.
The special characters "#" (for non-database variables), "+" (for application-independent variables), "@" (for SQL and Adabas null or length indicators) and "&" (for dynamic source generation) can be redefined in this section. In addition, the set of valid first and subsequent characters for identifiers, object names and library names can be modified. Note: |
CHARACTER-CLASSIFICATION
|
This section contains the tables required for the
classification of characters, which, for example, are used when evaluating the
MASK
option. It contains a subsection for each defined internal character
set.
|
The section CHARACTERSET-DEFINITION and each
subsection contain lines which describe how characters are to be converted and
which characters are related with which attributes. These lines are represented
as follows:
line ::= key = value
key ::= name_key | range_key
name_key ::= keyword{ CHARS }
keyword ::= INTERNAL-CHARACTERSET | NON-DB-VARI | DYNAMIC-SOURCE |
GLOBAL-VARI | FIRST-CHAR | SUBSEQUENT-CHAR |
LIB-FIRST-CHAR | LIB-SUBSEQUENT-CHAR | ALTERNATE-CARET
ISASCII | ISALPHA | ISALNUM | ISDIGIT | ISXDIGIT |
ISLOWER | ISUPPER | ISCNTRL | ISPRINT | ISPUNCT |
ISGRAPH | ISSPACE
range_key ::= hexnum | hexnum-hexnum
value ::= val {, val }
val ::= hexnum | hexnum-hexnum
hexnum ::= xhexdigithexdigit | xhexdigithexdigit
Notes:
range_key variable is
specified on the left-hand side, the number of values specified on the
right-hand side must correspond to the number of values specified in the key
range, unless only one value is specified on the right-hand side, which is then
assigned to each element of the key range.
name_key variable is
specified on the left-hand side and the corresponding list of character codes
does not fit in one line, it can be continued on the next line by specifying
name_key = again. You must not start the lines with leading blanks
or tabulators.
| x00-x1f = x00 | All characters between x00 and
x1f are converted to x00.
|
| x00-x7f = x00-x7f | All characters between x00 and
x7f are not converted.
|
| x00-x08 = x00,x01-x07,x00 | The characters x00 and x08
are converted to x00 and characters between x01 and
x07 are not converted.
|
|
ISALPHA = x41-x5a,x61-x7a,xc0-xd6,xd8 |
The attribute ISALPHA
is assigned to all characters specified in these two lines.
|
| x41 = 'A' | All characters must be specified in hexadecimal format. |
| 0x00-0x1f = 0x00 | Hexadecimal values have to be specified in either of
the following ways:
xdigitdigit |
| x00-x0f = x00,x01 | The number of specified values does not correspond to the number of elements in the key range. |