Created attachment 172357 [details] svn patch for math/libR portlint: OK testport: OK (poudriere: 9.3-RELEASE-p36, i386, default options) testport: OK (poudriere: 9.3-RELEASE-p30, amd64, default options) testport: OK (poudriere: 10.3-RELEASE, i386, default options) testport: OK (poudriere: 10.3-RELEASE, amd64, default options) I also tested a variety of option combinations and all built without error. However, math/atlas cannot be chosen as the BLAS implementation in poudriere, because it must be built on the target hardware for optimization. This is handled somewhat gracefully with ATLAS_VARS= MANUAL_PACKAGE_BUILD=... This applies to math/R as well. See http://math-atlas.sourceforge.net/atlas_install/node27.html.
Created attachment 172359 [details] svn diff to fix math/libR Same as the last patch, but with the line below. ATLAS_USES= blaslapack:atlas
The math/libR port has been deleted from the tree. https://svnweb.freebsd.org/ports?view=revision&revision=424082 As I don't fully understand this and the related math/libRmath request to separate it from math/R, and the port deletion was done by the reporter, I'm going to ask if this is okay to close first? Ok to close?
After further discussion, it seemed to appropriate action was to remove the port. So, yes, this can be closed. Thanks.