svn commit: r336343 - in head: . Mk/Uses

[Tijl Coosemans]

On Thu, 27 Feb 2014 15:13:44 +0100 John Marino wrote:
> On 2/27/2014 15:09, Tijl Coosemans wrote:
>> On Thu, 27 Feb 2014 15:02:51 +0100 John Marino wrote:
>>> On 2/27/2014 14:56, Tijl Coosemans wrote:
>>>> On Wed, 26 Feb 2014 23:03:08 +0100 (CET) Gerald Pfeifer wrote:
>>>>
>>>>> And it makes USES=fortran picking up the compiler specified with
>>>>> USE_GCC=X.Y, as a number of committers/users have asked the last
>>>>> months, harder.
>>>>
>>>> There are no ports with that combination at the moment, but I'm also not
>>>> sure it would be a good thing to let the choice of Fortran compiler
>>>> depend on USE_GCC=x.y.  It's much safer to have all Fortran code on a
>>>> system compiled with one compiler.
>>>
>>> Er, no, Gerald is correct.
>>> The fortran compiler used needs to be the same as USE_GCC.  I know this
>>> from experience with use of my ALTCOMPILER patch (pending approval).
>>> The fact that fortran was hardcoded to gcc4.6 was the source of
>>> breakage.  Gerald fixed something that I've had a pending patch for a
>>> few months now.
>> 
>> Can you explain what that problem is exactly?
> 
> Unreferenced symbols during linking IIRC.  I don't have a representative
> error log handy, nor do I remember a specific port that would fail if
> built with gcc4.8 and fortran 4.6, but there are some.

Hmm, one thing I can think of is the combination of the -rpath flags
from both versions.  If they end up in the wrong order then it's
possible that gcc48 tries to use the runtime libraries from gcc46
during linking.