[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gmp-bugs
Subject:    gmp-4.3.2 and 5.0.1 on i386-apple-darwin10 link-fails cxx tests
From:       fang () csl ! cornell ! edu (David Fang)
Date:       2011-03-19 1:30:45
Message-ID: 20110318212541.J97589 () fiji ! csl ! cornell ! edu
[Download RAW message or body]

> On Thu, 10 Mar 2011, David Fang wrote:
>
>> 	Yes, setting host/build to i686 explicitly did the trick, i.e. works 
>> on my core2-apple-darwin10 machine.  I'll ask the original reporters of 
>> this problem to try that on their pre-core2 machines.  I suspect it will 
>> work because we're using the exact same compiler/tools.
>
> It would still be nice to know why gmp identifies those CPUs as i386...

Old follow-up:

It turned out those machines the reporters were using were in fact corei* 
processors.  The releases of gmp I packaged had outdated versions of 
config.guess (from GMP).  So I grabbed config.guess from gmp's trunk and 
patched it return 'corei' for gmp-5.0.1 and 'core2' for gmp-4.3.2, because 
those were the newest cpus that their respective configure{.ac}s were 
aware of.  All feedback on builds on their systems has been positive; 
'make check' passed.

Fang

David Fang
http://www.csl.cornell.edu/~fang/
http://www.achronix.com/


[prev in list] [next in list] [prev in thread] [next in thread]