Don't use PyNumber_Float() since it's is nearly useless. Instead, rely on
PyFloat_AsDouble to generate a good error message.
This commit is contained in:
parent
293fdeb31e
commit
47ec6c0790
|
@ -214,11 +214,9 @@ select_select(PyObject *self, PyObject *args)
|
|||
return NULL;
|
||||
}
|
||||
else {
|
||||
tout = PyNumber_Float(tout);
|
||||
if (!tout)
|
||||
timeout = PyFloat_AsDouble(tout);
|
||||
if (timeout == -1 && PyErr_Occurred())
|
||||
return NULL;
|
||||
timeout = PyFloat_AS_DOUBLE(tout);
|
||||
Py_DECREF(tout);
|
||||
if (timeout > (double)LONG_MAX) {
|
||||
PyErr_SetString(PyExc_OverflowError,
|
||||
"timeout period too long");
|
||||
|
|
Loading…
Reference in New Issue