Re: gcc34 -- yikes!

On 8/15/06, Eric Albert <[email protected]> wrote:
>
> On Aug 15, 2006, at 3:24 AM, Tabitha McNerney wrote:
>
> > On 8/14/06, Ryan Schmidt <[email protected]> wrote:
> >>
> >>
> >> On Aug 15, 2006, at 02:41, Tabitha McNerney wrote:
> >>
> >> > Has anyone tried to install gcc34 version 3.4.6 on either PPC or
> >> > Intel arch?
> >> >
> >> > I tried a port install of gcc34 on a single processor G4 and it
> >> > took hours
> >> > and did not complete so I control-C'ed out of it and cleaned the
> >> > port rather
> >> > than using it. On an Intel Mac Mini Duo Core, gcc34 actually failed
> >> > (I will
> >> > reattempt the install with a -v to get more verbose output and file
> >> > a bug
> >> > report).
> >>
> >> There have been reports on this list before that gcc3 is in
> >> general not
> >> compatible with Intel Macs.
> >
> >
> > I wonder why. Maybe a big endian / little endian issue? Or
> > something more
> > complicated?
>
> Something much more complicated.  gcc 3.x is not ABI compatible with
> Mac OS X on Intel.  In other words, in certain cases  it will emit
> code which will not work with the OS.
>
> -Eric



Eric, apologies for the belated reply. Yes indeed, the gcc34 port built out
as a result of installing the octave port. So did the gcc40 port on PowerPC.
In fact, even though it took a long time, I was able to build out the octave
port and its dependencies on PowerPC no problem with Darwin Ports
1.3.1base. But I will refrain from trying this on Intel. Looks like,
as usual,
the hardware changes are always pacing ahead of the software changes (as
much as Apple deserves to be applauded for the rapid hardware architecture
switch).

Best regards,

T.