You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, we've recently triggered null pointer deref in that place, right on dynamic_cast. From the first glance code does not look correct, dynamic_cast occurs after explicitly calling destructor on object. In our combination of clang16, gcc13 and python3.11 (I don't know what exactly triggered this) this destructor compiles in such way that vtable pointer is set to 0 in the end, which does not work so well with dynamic_cast.
Using object after destroying it sound like UB anyway, can it be fixed somehow? Using dynamic_cast before calling object destructor?
The text was updated successfully, but these errors were encountered:
Hi, we've recently triggered null pointer deref in that place, right on dynamic_cast. From the first glance code does not look correct, dynamic_cast occurs after explicitly calling destructor on object. In our combination of clang16, gcc13 and python3.11 (I don't know what exactly triggered this) this destructor compiles in such way that vtable pointer is set to 0 in the end, which does not work so well with dynamic_cast.
Using object after destroying it sound like UB anyway, can it be fixed somehow? Using dynamic_cast before calling object destructor?
The text was updated successfully, but these errors were encountered: