https://bugs.kde.org/show_bug.cgi?id=445119
Oliver Kellogg <okell...@users.sourceforge.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |okellogg@users.sourceforge.
| |net
--- Comment #5 from Oliver Kellogg <okell...@users.sourceforge.net> ---
(In reply to Robert Hairgrove from comment #0)
> Using Umbrello from Git/master:
>
> In file "component.cpp" at line 58 in the method
> "UMLComponent::saveToXMI1()":
>
> for (UMLObjectListIt objectsIt(m_objects); objectsIt.hasNext();) {
> UMLObject* obj = objectsIt.next();
> uIgnoreZeroPointer(obj);
> obj->saveToXMI1 (writer);
> }
>
> What happens if "obj" == nullptr? A warning is emitted by
> "uIgnoreZeroPointer", but "obj" is dereferenced after that anyway, which
> causes undefined behavior according to the C++ standard.
Looking at the definition of uIgnoreZeroPointer, in
umbrello/debug/debug_utils.h :
#define uIgnoreZeroPointer(a) if (!a) { uDebug() << "zero pointer detected" <<
__FILE__ << __LINE__; continue; }
The `continue` statement means that the null pointer will not be accessed.
Just a minor point... not to detract from the question of why lists contain
null pointers at all.
--
You are receiving this mail because:
You are watching all bug changes.
