This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: sorry not implemented: address of bound pointer-to-member function
- To: egcs at cygnus dot com, peter at mcs dot nl
- Subject: Re: sorry not implemented: address of bound pointer-to-member function
- From: mrs at wrs dot com (Mike Stump)
- Date: Tue, 29 Sep 1998 12:52:15 -0700
> Date: Mon, 28 Sep 1998 18:26:55 +0200
> From: "A.P. Zijlstra" <peter@mcs.nl>
> To: egcs mailing list <egcs@cygnus.com>
> I'm currently using egcs-19980921 (l8est snapshot), and ran into the
> error: "sorry not implemented: address of bound pointer-to-member
> function".
> What do I do?
Well, you can try and fix your code so that it conforms to the ANSI
standard for C++. What you want to do is not permitted, find another
way.
> Is there someone working on this implementation
No.
> or do I code around this problem?
Best approach for portable code.
> Also could someone explain the exact meaning of this message to
> me.
Sure, a bound pointer-to-member function is a virtual member function
that is associated with a concrete object. The address of such a
thing is the address of the actual function.
I suppose that with thunks such an extension is ill-advised, if
someone wants to do up a patch to change the working into something
like error ("ANSI C++ forbids ..."), that would clarify this thing.