[R-SIG-Mac] Pb with gcc 4.0.3 (from R-2.4.1.dmg) and last Xcode

[Herve Pages]

Hi Simon,

Simon Urbanek wrote:
> Herve,
> 
> On Jan 3, 2007, at 11:31 PM, Herve Pages wrote:
> 
>> I recently updated Xcode to the last available version (2.4.1) on our
>> Intel Mac Mini (Mac OS 10.4.8).
>> Now when I compile a C program with gcc 4.0.3 provided with R-2.4.1.dmg,
>> the output I get contains several lines like this:
>>
>>   /var/tmp//ccUlRH08.s:2232:indirect jmp without `*'
>>
> 
> This is harmless, merely a warning and stems from the fact that Apple
> changed the behavior of the assembler on Intel Macs (for better - it is
> now more compatible with other assemblers - but consequently
> incompatible with its previous behavior).

Thanks for the details!

> Although the most recent Xcode fixed the fatal bugs that forced us to
> use our own compiler, we can't switch to Apple's gcc until R 2.5.0 due
> to binary incompatibility between the two (see R for Mac FAQ). If you
> build you own R, feel free to use Xcode's gcc.

That's good to know but no thanks :-) I like being able to update my R devel
in less than 1 min. by downloading the last universal build from
http://r.research.att.com/ (thanks a lot for providing this).
Also, I like being able to use the other universal stuff provided
there (like graphviz-2.8.dmg and glibgtk-bin-1.2.10-univ.tar.gz)
that would then become incompatible with my home made R (compiled
with Xcode's gcc). Is that correct?

Besides, Apple still does not provide a Fortran compiler, do they?

Cheers,
H.

> 
> Cheers,
> Simon
> 
> 
> 
>