On Thu, Aug 08, 2002 at 09:58:57PM +0200, Martin v. Loewis wrote: > Chris Halls <chris.halls@gmx.de> writes: > > > I have posted the test script and preprocessed source here: > > http://apt-proxy.sourceforge.net/ooo/g++test > > http://apt-proxy.sourceforge.net/ooo/test.cxx.bz2 > > Maybe I'm missing something, but... Is it the case that > ucb::ContentProviderImplHelper::~ContentProviderImplHelper is not > implemented in this file? Is it further the case that this is the > first virtual function of ContentProviderImplHelper? Hmm, yes it looks like that is the case. > If so, the virtual table for that class should not be emitted in > test.o, and it appears that g++ 3.1 behaves correctly in this respect. I take it that the 'typeinfo' should also not be emitted, as well as the virtual table? g++ 3.1 is actually emitting this typeinfo, along with 3.0 and 2.95 - it is the preliminary g++ 3.2 that is no longer emitting this information. This is going to break OpenOffice and any other programs that rely on this behaviour. So, should this change be considered to be a bug fix in g++ 3.2, and anything which breaks because of it should be fixed, or is it a problem with g++ 3.2? I'm pretty agnostic either way - I'm just trying to determine which program gets the blame and which needs fixing :) Thanks, Chris
Attachment:
pgp7kEJakp4cl.pgp
Description: PGP signature