Re: [gtkmm] Re: gcc 3.4-related gtkmm patches needed ASAP

[Jonathan Brandmeyer <jbrandmeyer earthlink net>]

On Fri, 2004-03-12 at 19:21, Murray Cumming wrote:
> On Thu, 2004-02-12 at 20:44, Jonathan Brandmeyer wrote:
> > Murray Cumming wrote:
> > > I'd also like to know how best to get these "g++ 3.4 pre-releases"
> > > onto
> > > my system. Is there some easy way to apt-get them onto my debian
> > > installation? Obviously I don't want to risk messing up my regular
> > > (Red
> > > Hat 9) development machine.
> > > 
> > 
> > I'm using Debian Sid at home,
> [snip]
> 
> Can your g++ 3.4 build the current libsigc++ 2 from source? Mine can
> not.
> 
> If yes, please tell me when you checked gcc 3.4 out of cvs, so that I
> can get the same version.

No, it doesn't work for me either.  The first class of errors can be
fixed as per http://gcc.gnu.org/gcc-3.4/changes.html.  See the section
that says 'in a template unqualified names will no longer find members
of a dependent base'  Therefore, changing line 1781 of mem_fun.h to "   
{ return (obj_ptr_->*(this->func_ptr_))(); }" works.

I think g++ is telling you what to fix for deduce_result_type.h.

signal.h line 340 should be this: 
	if (!i_->empty() && !i_->blocked() && !invoked_)

HTH,
Jonathan Brandmeyer