Calculate smallest properly (it was off by one) and use proper ssize_t types for Win64
diff --git a/Modules/_struct.c b/Modules/_struct.c
index 1491b2e..6a3c11f 100644
--- a/Modules/_struct.c
+++ b/Modules/_struct.c
@@ -309,24 +309,24 @@
_range_error(const formatdef *f, int is_unsigned)
{
if (is_unsigned == 0) {
- long smallest = 0, largest = 0;
+ Py_ssize_t smallest, largest = 0;
Py_ssize_t i = f->size * 8;
while (--i > 0) {
- smallest = (smallest * 2) - 1;
largest = (largest * 2) + 1;
}
+ smallest = -largest - 1;
PyErr_Format(StructError,
- "'%c' format requires %ld <= number <= %ld",
+ "'%c' format requires %zd <= number <= %zd",
f->format,
smallest,
largest);
} else {
- unsigned long largest = 0;
+ size_t largest = 0;
Py_ssize_t i = f->size * 8;
while (--i >= 0)
largest = (largest * 2) + 1;
PyErr_Format(StructError,
- "'%c' format requires 0 <= number <= %lu",
+ "'%c' format requires 0 <= number <= %zu",
f->format,
largest);
}