Types & literals

Chuchu checks types before generating LLVM code. Variables infer their types; function signatures use lowercase int, float, and bool annotations.

Type	Examples	Current backend representation
int	0, 42, -7	Signed 64-bit integer
float	1.0, 3.14159, 2e-3	64-bit floating point
bool	true, false	Boolean
Numeric array	[1, 2], [1.5, 2.5]	Fixed-size value containing 1–256 elements

Integers

Integer literals use decimal digits. Negation uses unary -. The backend accepts values from -9223372036854775808 through 9223372036854775807.

Arithmetic is checked: overflowing addition, subtraction, multiplication, negation, or abs terminates the executable with a runtime diagnostic. Out-of-range integer literals are rejected during code generation.

This width describes the current implementation; it is not a promise about every future language version.

Floating point

A decimal point or exponent produces a float:

show 1.25
show 2e3
show 2.5e-1

Float literals must parse as finite f64 values. Malformed or non-finite literals such as 1e and 1e9999 are rejected. Numeric separators such as 1_000 are unsupported.

Runtime floating-point calculations can produce infinity or NaN. For example, sqrt(-1.0) produces NaN. Division by positive or negative zero is checked and fails at runtime.

Numeric widening

Mixed scalar arithmetic and comparisons convert integer operands to floats:

show 1 + 2.5
show 1 == 1.0

Output: 3.5, then true. Integer-to-float conversion is also accepted for float function parameters, float return values, and reassignment to an existing float variable. Large integers can lose precision when converted to f64.

There is no implicit float-to-int conversion or explicit cast syntax.

Arrays

Array type identity includes element type and length. [1, 2] and [1, 2, 3] have different types, as do [1, 2] and [1.0, 2.0]. Array elements must all have the same numeric type; arrays do not widen mixed elements.

Empty arrays, boolean arrays, nested arrays, and arrays above 256 elements are unsupported. Array annotations in function signatures are also unsupported. See arrays.

Output formatting

show prints one scalar followed by a newline:

• Integers use signed decimal output.
• Floats use C printf format %.15g; 2.0 can print as 2.
• Booleans print true or false.

Strings, characters, null values, objects, and user-defined types are not currently implemented.