This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: [Patch] Nathan's improvements to basic_string
On Thu, Jun 12, 2003 at 11:15:27PM +0200, Paolo Carlini wrote:
> Loren James Rittle wrote:
>
> >My only concern (in terms of committing it ASAP to mainline) might be
> >that this is at least a reverse incompatibility. It may be (i.e. I
> >think it is but haven't personally tested it) forward compatible.
> >
> I believe that for 3.4 we are completely free to break the ABI, don't
> you?... see, f.i., what we have already committed for iostreams (not to
> mention what's around the corner ;)
Yes, I think 3.4 is already binary-incompatible with 3.3. The day
is fast approaching, though, when we will have to stop breaking the
ABI so routinely. Probably there will be a lot more resistance to
an incompatible 3.5. I would expect reduced uptake of 3.4 if the
performance improvement wasn't so great. (Yay, you guys.)
What's changing is that people are finally beginning to move off
2.95/2.96, partly because of x86-64, where they have had at least the
illusion of stability, for years at a stretch. They're landing on 3.2
or 3.3, and they expect to stay there as long. Pray that more hold
out a little longer and land on 3.4 than on an earlier release.
Numerics people I talk to, by the way, say that each C compiler
release since 2.7 has generated slower code than the last, and
they're starting to get very testy about it.
Nathan Myers
ncm-nospam@cantrip.org