diff options
Diffstat (limited to 'Objects')
| -rw-r--r-- | Objects/floatobject.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/Objects/floatobject.c b/Objects/floatobject.c index a42be71..b571ca8 100644 --- a/Objects/floatobject.c +++ b/Objects/floatobject.c @@ -289,7 +289,7 @@ float_repr(PyFloatObject *v) * may lose info from fractional bits. Converting the integer to a double * also has two failure modes: (1) a long int may trigger overflow (too * large to fit in the dynamic range of a C double); (2) even a C long may have - * more bits than fit in a C double (e.g., on a a 64-bit box long may have + * more bits than fit in a C double (e.g., on a 64-bit box long may have * 63 bits of precision, but a C double probably has only 53), and then * we can falsely claim equality when low-order integer bits are lost by * coercion to double. So this part is painful too. |