[aspect-devel] Compilation errors with old compilers since rev.2414

Rene Gassmoeller rengas at gfz-potsdam.de
Fri May 2 06:44:26 PDT 2014


Problem solved.

Sorry, line numbers were from r.2414.

Problem is solved for me. GCC 4.4.4 and 4.1.2 compile flawlessly from
Revision 2563 on.
Thanks a lot and looking forward to see you soon.

Rene

On 05/02/2014 03:28 PM, Wolfgang Bangerth wrote:
> 
> Hi Rene & Cedric,
> I can't quite match the line numbers, but I think I know what's going on
> anyway. Can you try again?
> Best
>  Wolfgang
> 
> 
> On 05/02/2014 06:32 AM, Rene Gassmoeller wrote:
>> Hi all,
>> I today updated Aspect on two systems with older compilers and noticed
>> that since rev.2414 gcc versions 4.4.4 and 4.1.2 (antique I know) throw
>> the following errors during compilation:
>>
>> /home/mod/rengas/Software/aspect-dev/source/simulator/nullspace.cc: In
>> member function ‘void
>> aspect::Simulator<dim>::remove_net_angular_momentum(aspect::LinearAlgebra::BlockVector&,
>>
>> aspect::LinearAlgebra::BlockVector&) [with int dim = 2]’:
>> /home/mod/rengas/Software/aspect-dev/source/simulator/nullspace.cc:155:
>> error: using ‘typename’ outside of template
>> /home/mod/rengas/Software/aspect-dev/source/simulator/nullspace.cc:172:
>> error: using ‘typename’ outside of template
>> /home/mod/rengas/Software/aspect-dev/source/simulator/nullspace.cc:173:
>> error: using ‘typename’ outside of template
>> At global scope:
>> cc1plus: warning: unrecognized command line option
>> "-Wno-unused-local-typedefs"
>> make[2]: *** [CMakeFiles/aspect.dir/source/simulator/nullspace.cc.o]
>> Error 1
>> make[2]: *** Waiting for unfinished jobs....
>> make[1]: *** [CMakeFiles/aspect.dir/all] Error 2
>> make: *** [all] Error 2
>>
>> A quick Google Search told me that this seems to be a new feature in
>> C++11 that is not understood by older compilers. Is there a way to
>> reformulate this code for older compilers (as they are used on some
>> 'stable' cluster systems)? If yes, then maybe someone who is more
>> familiar with the details of C++03 and C++11 could change this for the
>> convenience of some users ;-)? If no, then we should mention the
>> dependency on newer compiler version or compiler compatibility
>> somewhere, to let users know about it. Unfortunately it is know in the
>> release version (my fault, only checked aspect-1.0 on my laptop and
>> another cluster, gcc 4.6.3 and 4.8.1 are working well, by the way).
>>
>> Any thoughts or suggestions?
>>
>> Best,
>> Rene
>> _______________________________________________
>> Aspect-devel mailing list
>> Aspect-devel at geodynamics.org
>> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>>
> 
> 


More information about the Aspect-devel mailing list