STR09-C. Don't assume numeric values for expressions with type plain character

For portable applications, use only the assignment = operator, the equality operators == and !=, and the unary & operator on plain-character-typed or plain-wide-character-typed expressions.

This practice is recommended because the C Standard requires only the digit characters (0–9) to have consecutive numerical values. Consequently, operations that rely on expected values for plain-character- or plain-wide-character-typed expressions can lead to unexpected behavior.

However, because of the requirement for digit characters, other operators can be used for them according to the following restrictions:

• The binary + operator may be used to add integer values 0 through 9 to '0'.
• The binary - operator may be used to subtract character 0.
• Relational operators <, <=, >, and >= can be used to check whether a character or wide character is a digit.

Character types should be chosen and used in accordance with STR04-C. Use plain char for characters in the basic character set.

Noncompliant Code Example

This noncompliant code example attempts to determine if the value of a character variable is between 'a' and 'c' inclusive. However, because the C Standard does not require the letter characters to be in consecutive or alphabetic order, the check might not work as expected.

char ch = 'b';
if ((ch >= 'a') && (ch <= 'c')) {
  /* ... */
}

Compliant Solution

In this example, the specific check is enforced using compliant operations on character expressions:

char ch = 't';
if ((ch == 'a') || (ch == 'b') || (ch == 'c')) {
  /* ... */
}

Exceptions

STR09-C-EX1: Consecutive values for characters like a~z can be assumed on platforms where ASCII or Unicode is used. This recommendation is primarily concerned with platform portability, for example, if code is migrated from ASCII systems to non-ASCII systems.

Risk Assessment

Automated Detection

Related Guidelines

Bibliography