Lexical grammar

Whitespace and comments

Unicode whitespace separates tokens. # starts a comment that continues to the next newline. Jazz has no block-comment form.

Identifiers

An identifier starts with a Unicode alphabetic character or _. Continuation characters may also contain Unicode alphanumerics, _, ', and !. The ! suffix has purity meaning, but is lexically part of the identifier.

Reserved words are module, import, as, data, value, if, then, else, and case. True and False are literal names. Declaration words such as class, impl, and operator are contextual identifiers and are recognized only in their declaration shapes.

Literals

Integers are unsigned decimal digit sequences; negative values are expressed with subtraction. Fractional literals require adjacent whole, dot, and fractional digits, such as 1.25. Optional adjacent suffixes f16, f32, and f64 select a concrete floating width. Other suffix spellings are not numeric syntax.

Character literals use single quotes and contain exactly one Unicode scalar. Text literals use double quotes. Both accept \\, \', \", \n, \r, \t, \0, and \u{HEX} with one to six hexadecimal digits. Surrogates, values above 0x10FFFF, raw newlines, invalid escapes, and unterminated literals are E0001 errors.

Symbols

Structural symbols are ::, :, @, =, ->, ., braces, parentheses, brackets, comma, and backslash. Operator runs use characters from !%&*+-/<>?^|~; $ is also built in. Arrow and several comment-like symbols are reserved from user-defined operators.

All tokens retain one-based line and column positions for diagnostics.