This is the mail archive of the
mailing list for the GCC project.
Re: Bug in MI?
- To: oliva at dcc dot unicamp dot br, jason at cygnus dot com, mmitchell at usa dot net, mrs at wrs dot com, egcs-bugs at cygnus dot com, Manfred dot Hollstein at ks dot sel dot alcatel dot de
- Subject: Re: Bug in MI?
- From: Manfred Hollstein <manfred at s-direktnet dot de>
- Date: Fri, 6 Mar 98 10:27:21 +0100
- References: <9803060830.AA22277@lts.sel.alcatel.de>
- Reply-To: Manfred Hollstein <manfred at s-direktnet dot de>
On Fri, 6 March 1998, 09:30:32, email@example.com wrote:
> I'm just looking at a problem some of my collegues are having. It's
> actually quite a big source file (around 70000 lines after
> preprocessing) and uses a very complex inheritance lattice (8 parents
> and around 30 ancestors - dammed I'd really appreciat having something
> like Source-Navigator around...), but I've been able to reduce it to
> 30 lines.
> Compiling it with all gcc versions I have around here (2.6.3, 2.7.2,
> egcs in various versions), results in messages as follows:
> t153.cc: In function `int main()':
> t153.cc:29: cannot allocate an object of type `ProtectedTTPBidirectionalData'
> t153.cc:29: since the following virtual functions are abstract:
> t153.cc:29: int ProtectedTTPSourceData::convertAddedPtrToCrossConnectionObjectPointer()
> BUT, there is absolutely _no_ pure virtual function flying around!?!
Forget about this problem. I just realized, that gcc is doing the
right thing ;-)
According to C++WP 10.3 [class.virtual], the `final overrider' is
missing, hence the program is ill-formed!
But, gcc's error message is IMHO absolutely misleading.