Re: Backporting to 4_0 the latest friend bits

[Joe Buck <Joe dot Buck at synopsys dot COM>]

On Fri, Apr 29, 2005 at 10:12:46PM +0200, Paolo Carlini wrote:
> >> I know that, technically, we are not talking about regressions wrt
> >> 3.x, still, important packages that used to compile and, well,
> >> apparently at least, *work* well, now don't even compile (see
> >> c++/19403, c++/21235, many others linked from there). Would be a big
> >> deal having more of Kriang's recent work in 4_0 too, or is there
> >> something else that you can suggest?
> >
> > Do we know that (a) the affected programs are valid, and (b) the
> > patches will make them compile again?
> >
> > I understand the compiler did the wrong thing before, but it might be
> > that the programs were valid, the compiler did a wrong-but-harmless
> > thing, and now 4.0.0 does a wrong-but-harmful thing.  If so, I think
> > it's reasonable to consider patches (if simple) that change the
> > compiler to do the right thing; going back to the wrong-but-harmless
> > thing doesn't seem so attractive.
> 
> Thanks for your assessment of the problem: indeed, I can tell you for
> sure that (b) it's true and, as reported by Kriang, the patches are
> rather simple (but the details of this judgement are up to you, of
> course). I'm not 100% sure about (a) but Michael can tell you better:
> AFAIK, KDE Kopete used to compile and work well pre-4.0.

I don't quite understand your answer.  It seems that (a) is the important
issue; if the programs are valid, they compiled before, and they worked
before, then it seems there really is a regression, even if we can argue
that we were "right by accident" in the past.

It seems we need an answer to (a) to settle whether this is a high
priority bug.