Runtime errors

When a program fails while running, it prints error: and a message to stderr and exits with status 1. Output printed before the error stays printed. The interpreter prefixes the message with the file name, and a compiled executable prints it on its own:

$ stone run average.st
7
average.st: error: division by zero
$ stone build average.st -o average && ./average
7
error: division by zero

Both backends stop at the same point with the same message, since operands are evaluated left to right in both.

message	cause
division by zero	/ or % with a divisor of 0, or 0.0 (as in Python, rather than IEEE 754’s infinity or nan)
integer overflow in division	the smallest int divided by -1, which has no int result. % by -1 is always 0
negative exponent	** with an int base and a negative exponent, which has no int result
list index out of range	an index at or past the length, or a negative index before the start. The interpreter adds the index and length
cannot convert float to int (nan or out of range)	int() of nan, an infinity, or a float outside the int range
invalid literal for int() with base 10: '...'	int() of a string that is not a decimal int, quoted as given
int() argument out of range: '...'	int() of a string holding an int too large or small to fit
could not convert string to float: '...'	float() of a string that is not a float, quoted as given
empty separator	split("")
recursion is too deep (more than 1000 nested calls)	more than 1,000 calls active at once
'name' is used before it is assigned	a function read a global that had not been assigned yet when it ran

Integer +, -, *, and ** wrap around on overflow instead of failing. Float arithmetic follows IEEE 754 apart from division by zero, so it produces inf, -inf, and nan instead of errors. A float ** multiplies by repeated squaring and takes the reciprocal for a negative exponent, so 0.0 ** -1 is inf.

A ret at the top level is not an error: it ends the program with status 0.