On 08/30/2012 05:53 PM, Douglas Bates wrote:
You're right. I didn't see that -I/opt/local/include was in the compiler call. If you look at $RHOME/etc/Makeconf you will probably see that CPPFLAGS=-I/opt/local/include
Correct!
(in my case /usr/local/include) for reasons that are somewhat mysterious to me. It seems like the sort of thing that a prominent member of R-core would yell and scream about if he had not put that there himself.
I can't blame anyone else. That's part of our standard setup which was working beautifully until 2 days ago. But, I will re-think it.
It happens that the Ubuntu package for eigen3 installs the Eigen header files in /usr/include/eigen3 so I don't get bitten by that. It seems that you options are 1) Change R's default CPPFLAGS, which seems unlikely to happen or 2) change the location of the Eigen header files.
When the next release of R appears, #1 is probably what I should look into since it is probably self-inflicted and, otherwise, may lead to other unusual bug reports in the future. For a quick and dirty fix, I tweaked my original hack (after playing around with the cpp flag -I- and giving up in frustration). Replace line 9 of PREFIX/include/Eigen/CholmodSupport with #ifdef RcppEigen__RcppEigen__h #include <RcppEigenCholmod.h> #else #include <suitesparse/cholmod.h> #endif Thanks for all of your help! -- Rodney Sparapani, PhD Center for Patient Care and Outcomes Research Sr. Biostatistician http://www.mcw.edu/pcor 4 wheels good, 2 wheels better! Medical College of Wisconsin (MCW) WWLD?: What Would Lombardi Do? Milwaukee, WI, USA _______________________________________________ Rcpp-devel mailing list Rcpp-devel@lists.r-forge.r-project.org https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel