Matthew Emmett

Is there a particular case/example that tripped a bug? If so, I will double check it...

I will get some regression tests running... do they need to be MPI enabled at this stage?

I have added two regression tests. The easiest (fewest lines of code) way to do this was by monkey patching the `build_variant_arg_dicts` routine in `pyclaw.utils`. A better long term solution...

Oops, didn't notice that hardcoded path -- I will make it relative. Unfortunately I don't have access to an OSX machine, so wasn't able to test the build step on...

Wow, I've never seen that error message before. There are two versions of BoxLib: a pure F90 version, and a C++/F90 version. PyBoxLib wraps the C++ version using SWIG. Looks...

@mandli @ketch What versions of OS X are you guys running? Can you give some details on how you installed your compilers? I will try to recreate your environments and...

@ketch I just pushed a change to BoxLib in an attempt to get it to link properly. Can you try it out?

@ketch It looks like the version of g++ you are (currently) using is 4.2.1, but your gfortran is 4.8.1. Is that intentional? Did something in your environment change recently? I...

Could you try with g++ 4.8? The offending code in `FArrayBox.cpp` was changed recently to support the latest Cray compilers. Perhaps the calls to `isnan` and `isinf` were not being...

Just realised that the Makefile is setting flags appropriate for GCC, but your mpic++ is picking up the LLVM compilers. Are you using MPICH? Can you set MPICH_CXX=/path/to/g++?