Software Internationalization Guide

Software Characteristics That Vary by Locale

Software Internationalization Guide—526225-002

2-11

Signed Versus Unsigned Character Data Types

supported character sets of 1, 2, and 4 bytes. To store 1-byte and 2-byte characters,

the implementation pads unused space in the 4-byte-wide character with nulls.

Figure 2-10 is an example of a 4-byte-wide character data stream. The unshaded

bytes in the figure indicate null padding.

Wide characters are not used for file codes, the external representation of data.

Because wide character size varies between systems, an application might not be able

to interpret wide character file codes defined on another system. Wide character types

are thus used only for internal processing. For more information about file codes and

internal processing codes, see File Codes and Process Codes on page 2-11.

Signed Versus Unsigned Character Data Types

ASCII-based programs do not need to differentiate between signed and unsigned

characters for most purposes, but programs using larger code sets often do. Large

code sets require eight bits or more to fully represent characters, so it must be possible

to distinguish between signed and unsigned characters without using the eighth bit for

the sign. Internationalized software must use an unsigned character data type to store

characters, so that the character can never be misinterpreted as a negative binary

value.

Integer Versus Character Data Types

When an application compares integer data types to character data types it might

produce unexpected results if multiple bytes are used for both character and integer

representation. For example, consider a conditional statement that compares the

variable b to the end-of-file (EOF) character:

int b;

if ((b = getchar()) == EOF)

Because the return value of getchar() is an integer, it is important that the return

value be stored as an integer also so that the comparison will work properly.

File Codes and Process Codes

A file code is the bit pattern representing a character in a character set. It is the

external representation of data and is the format of data stored on disk.

Figure 2-10. Four-Byte Wide Character Data Stream

byte

. . . . . .

4-byte character

VST011.vsd

2-byte character 1-byte character