Re: Debugging missing vtables

[Florian Weimer <fw at deneb dot enyo dot de>]

* Ian Lance Taylor:

> Florian Weimer <fw@deneb.enyo.de> writes:
>
>> On some large application (not written by me), GCC 4.0 does not emit
>> some vtables, and link errors are the result.
>> 
>> If I understand things correctly, a certain member is deemed the
>> controlling member, and the vtable is emitted in the translation unit
>> which contains this member.
>> 
>> Is there a way to cause GCC to print this member?  I think this would
>> help me to debug this issue.  (It's probably a bug in the application,
>> but I'm not completely sure.)
>
>>From the documentation:
>
>     C++ virtual functions are implemented in most compilers using a lookup
>     table, known as a vtable.  The vtable contains pointers to the virtual
>     functions provided by a class, and each object of the class contains a
>     pointer to its vtable (or vtables, in some multiple-inheritance
>     situations).  If the class declares any non-inline, non-pure virtual
>     functions, the first one is chosen as the ``key method'' for the class,
>     and the vtable is only emitted in the translation unit where the key
>     method is defined.
>
> I don't know of any way to print the member.

It wouldn't have helped in this case anyway.

The code looks like this:

template <int I>
struct Foo
{
  virtual ~Foo();
};

template struct Foo<5>;

template <int I>
Foo<I>::~Foo()
{
}

int
main()
{
  return 0;
}

This used to work with GCC 3.4.  It would be nice if GCC 4.x could
issue some warning, instead of running into a link error later.  (It's
probably not a GCC bug because the program is invalid and the error is
detected at link time.)

Does this issue look familiar?  Is there a PR about it?