Source files

Cuarzo source uses the .cz extension and must be valid UTF-8 without NUL bytes. Statements are normally separated by newlines. Blocks use explicit keywords rather than braces.

// A line comment
/* A block
   comment */

The lexical rules are byte-oriented and ASCII outside of string and comment contents:

• identifiers match [A-Za-z_][A-Za-z0-9_]*;
• numeric literals use ASCII digits; a float needs digits on both sides of the dot, such as 3.14;
• whitespace is space or tab, and a newline is LF, CRLF, or a bare CR;
• strings and comments may contain any UTF-8 text, which is kept byte for byte;
• diagnostics have the form [ERROR E0042] file:line:col: message, with a code that Diagnostics explains, a one-based line and a one-based UTF-8 byte column. Errors are reported up to ten at a time: after a syntax error, the compiler reads on from the next statement, and reports every syntax error of every file before it stops; without one, each statement with an error is left out while the rest are checked.

Integer literals keep their decimal meaning, so 007 is the integer 7.

A line break inside parentheses or brackets does not end a statement, so long calls, conditions, and array literals can span several lines:

str[] words = [
    "one",
    "two",
]
print(join(
    words,
    ", "
))

Code is laid out with four spaces for each block, one space around operators and after commas, and none inside brackets or before a call's parenthesis. cuarzo fmt lays files out that way in place, keeping their line breaks and comments; cuarzo fmt --check lists the files it would change, for CI.