In my current compiles of R-devel from source (using clang from Yosemite and gfortran from homebrew) I added "-Wl,-no_compact_unwind" to my LDFLAGS to get around the annoying warning "create compact unwind for foo: stack subq instruction is too different from dwarf stack" === Jan de Leeuw Distinguished Professor Emeritus and Founding Chair, UCLA Statistics Editor: Journal of Multivariate Analysis, Journal of Statistical Software email: deleeuw at stat.ucla.edu ? homepages: http://gifi.stat.ucla.edu ------------------------------------------------------------------------- No matter where you go, there you are. --- Buckaroo Banzai http://gifi.stat.ucla.edu/sounds/nomatter.wav -------------------------------------------------------------------------
linker warnings
2 messages · Jan de Leeuw, Brian Ripley
On 30/01/2015 18:28, Jan de Leeuw wrote:
In my current compiles of R-devel from source (using clang from Yosemite and gfortran from homebrew) I added "-Wl,-no_compact_unwind" to my LDFLAGS to get around the annoying warning "create compact unwind for foo: stack subq instruction is too different from dwarf stack"
Yes, but this is nothing to do with R and everything to do with the unstated version of gfortran you are using (I see it with 4.9.2 built from the sources). That gcc/gfortran do not play nicely with Apple's linker is something to take up with one of them.
=== Jan de Leeuw Distinguished Professor Emeritus and Founding Chair, UCLA Statistics Editor: Journal of Multivariate Analysis, Journal of Statistical Software email: deleeuw at stat.ucla.edu ? homepages: http://gifi.stat.ucla.edu
Brian D. Ripley, ripley at stats.ox.ac.uk Emeritus Professor of Applied Statistics, University of Oxford 1 South Parks Road, Oxford OX1 3TG, UK