Lexical Structure¶

Quxlang makes keywords and user-defined names visually distinct. Keywords use uppercase spelling; source identifiers use lowercase spelling.

Identifiers and keywords¶

A user identifier:

begins with a through z;
continues with lowercase letters, decimal digits, or _; and
does not end with _.

Valid examples include value, value2, and source_module. Value, SOURCE_MODULE, _value, and value_ are not user identifiers.

Language keywords and reserved compiler names are uppercase, such as FUNCTION, STATIC_IF, and STRING_CONSTANT. This division lets Quxlang add new keyword spellings without colliding with ordinary source names.

Whitespace and comments¶

Spaces, tabs, and line breaks separate tokens. A line comment begins with // and continues to the end of the line:

VAR count I32 := 4; // Number of items to process.

Block-comment syntax is not part of the current source grammar.

Numeric literals¶

Numeric literals use decimal digits with at most one decimal point:

VAR integer I32 := 42;
VAR fraction F64 := 12.5;

The literal itself remains an exact compile-time literal until context selects a concrete numeric type. A leading minus can be part of a numeric literal, including the minimum value of a signed type:

VAR negative I32 := -7;
VAR minimum I8 := -128;

Current source literals do not use C-style hexadecimal, octal, binary, digit separator, or exponent notation.

String and character literals¶

Strings use double quotes and characters use single quotes:

VAR message STRING_CONSTANT := "line one\nline two";
VAR newline BYTE := '\n';

String escapes are \n, \t, \r, \0, \\, and \". Character literals support the same byte escapes, with \' in place of the escaped double quote. A character literal represents one BYTE, not a multi-byte Unicode scalar.

See Primitive types and literals and Source files and imports.