On 09/24/2015 10:02 AM, Pavel Hrdina wrote:
Coverity cannot introspect the cpython code and it thinks, that if the PyTuple_New or similar will always give the same result. In functions where the python function is called more than once with the same arguments, it will assume that if the first call don't return NULL, then also the every other call cannot return NULL. This is a wrong assumption and it's a false-positive.
My turn to make the obvious question... Have you asked the Coverity folks? This just seems very strange. I agree it seems to be a false positive, but something isn't adding up. John
Signed-off-by: Pavel Hrdina <phrdina@redhat.com> --- libvirt-override.c | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/libvirt-override.c b/libvirt-override.c index 3192e74..5330d0f 100644 --- a/libvirt-override.c +++ b/libvirt-override.c @@ -5409,6 +5409,7 @@ libvirt_virEventAddTimeoutFunc(int timeout, Py_INCREF(python_cb); VIR_PY_TUPLE_SET_GOTO(pyobj_args, 1, python_cb, cleanup);
+ /* coverity[notnull] */ if ((cb_args = PyTuple_New(3)) == NULL) goto cleanup;
@@ -6125,6 +6126,7 @@ libvirt_virConnectDomainEventGraphicsCallback(virConnectPtr conn ATTRIBUTE_UNUSE libvirt_constcharPtrWrap(local->service), cleanup);
+ /* coverity[notnull] */ if ((pyobj_remote = PyDict_New()) == NULL) goto cleanup;