McLachlan in ET/master needs an updated KrancNumericalTools/GenericFD

[Frank Löffler]

McLachlan in ET/master needs (among other things) kisgn(), which is defined in Kranc/Tools/CodeGen/CodeGenCactus.m, but not mentioned in KrancNumericalTools/GenericFD/src/GenericFD.h.

Could it be that something needs to be updated from the Mathematica sources?

COMPILING /home/knarf/Cactus/arrangements/McLachlan/ML_BSSN/src/ML_BSSN_RHS1.cc
/home/knarf/Cactus/configs/sim/build/ML_BSSN/ML_BSSN_RHS1.cc: In function ‘void ML_BSSN_RHS1_Body(const cGH*, int, int, const double*, const double*, const double*, const int*, const int*, int, double* const __restrict__*)’:
/home/knarf/Cactus/configs/sim/build/ML_BSSN/ML_BSSN_RHS1.cc:648: error: ‘kisgn’ was not declared in this scope

Keyword:

[Ian Hinder]

As far as I can tell, kisgn is defined in LSUThorns/Vectors/src/vectors.h (http://git.barrywardell.net/arrangements/LSUThorns/Vectors.git/blob/HEAD:/src/vectors.h):

    220 #define kisgn(a) (-42424242)

This was added in August:

    Author: eschnett <eschnett@105869f7-3296-0410-a4ea-f4349344b45a>
    Date:   Thu Aug 16 18:52:18 2012 +0000

    Add dummy kisgn function

    git-svn-id: https://svn.cct.lsu.edu/repos/numrel/LSUThorns/Vectors/trunk@64 105869f7-3296-0410-a4ea-f4349344b45a

and seems to be related to this ticket: https://github.com/ianhinder/Kranc/issues/64

Do you have a clean checkout and a fresh build? I just tried building on my laptop with GCC, using origin/master branches of Kranc and McLachlan, and I don't see this error. I also tried in the ubuntu VM with vectorisation enabled, and it was also fine. Could this be an intel compiler issue?

[Peter Diener]

I don't see this problem on my laptop using the Intel 12 compilers, so I don't think it's an intel compiler issue.

[Frank Löffler]

Updating Vectors helped, thanks. Another confirmation that ticket:241 would be nice to have.