Re: [R-pkg-devel] Making adjustments to the R CMD check/install commands
Yay! I got it to work! I used the ~/.R/Makevars, and all is well. Thanks again, Erin Erin Hodgess, PhD mailto: erinm.hodg...@gmail.com On Mon, Aug 27, 2018 at 2:29 PM Kevin Ushey wrote: > Hi Erin, > > The R extensions manual should be helpful here -- in particular, see the > sections on the 'src/Makevars' file and configuration of the Fortran > compiler: > > https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-Makevars > > In particular, in your 'src/Makevars' file, you'll likely want to set the > FC variable to choose the Fortran compiler, and FCFLAGS to choose the > appropriate compilation flags. As an example, you can see how one might set > these flags to compile code with the address sanitizer active: > > > https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-Address-Sanitizer > > (Note that this example uses the user-level '~/.R/Makevars' file rather > than the package-level 'src/Makevars' file, but the crux is the same) > > Hope this gets you on the right track, > Kevin > > On Mon, Aug 27, 2018 at 1:15 PM Erin Hodgess > wrote: > >> Hello! >> >> I'm building a package on Windows 10 which uses the PGI Fortran compiler. >> I have a Makefile set up in the src directory, and when I run "make all", >> the subroutines compile beautifully. >> >> Now I want to create my package. When I run R CMD check, naturally, >> gfortran is called. Is there a way that I can direct the check function >> to >> just use the make/Makefile, please? >> >> Or am I out of luck, please? (highly probable) >> >> Thanks, >> Erin >> >> Erin Hodgess, PhD >> mailto: erinm.hodg...@gmail.com >> >> [[alternative HTML version deleted]] >> >> __ >> R-package-devel@r-project.org mailing list >> https://stat.ethz.ch/mailman/listinfo/r-package-devel >> > [[alternative HTML version deleted]] __ R-package-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-package-devel
Re: [R-pkg-devel] Making adjustments to the R CMD check/install commands
Thank you, Kevin! I put together the Makevars file: PKG_LIBS = $(LAPACK_LIBS) $(BLAS_LIBS) FC=c:/progra~1/PGICE/win64/18.4/bin/pgf90 SAFE_FFLAGS="-Lc:/progra~1/PGICE/win64/18.4/lib/pgf90.dll -Lc:/progra~1/PGICE/win64/18.4/lib/pgc14.dll" rmpiFort.obj: rmpiFort.f90 $(FC) $(SAFE_FFLAGS) -Mmpi=msmpi -c rmpiFort.f90 -o rmpiFort.obj rmpiFort.dll: rmpiFort.obj $(FC) $(SAFE_FFLAGS) rmpiFort.obj -Mmpi=msmpi -Mmakedll -o rmpiFort.dll But when I run R CMD check, the 00install file shows that the gfortran compile is used, rather than the PGI compiler: * installing *source* package 'rmpiFort' ...^M ** libs^M ^M *** arch - i386^M C:/progra~1/R/R-35~1.1/share/make/winshlib.mk:13: warning: overriding commands for target `rmpiFort.dll' Makevars:10: warning: ignoring old commands for target `rmpiFort.dll' c:/Rtools/mingw_32/bin/gfortran -O2 -msse2 -mfpmath=sse -Mmpi=msmpi -c rmpiFort.f90 -o rmpiFort.obj gfortran.exe: error: unrecognized command line option '-Mmpi=msmpi'^M make: *** [rmpiFort.obj] Error 1 ERROR: compilation failed for package 'rmpiFort'^M * removing 'C:/Users/erinm/rmpiFort.Rcheck/rmpiFort'^M In R CMD INSTALL^M So I'm still puzzled. Any help much appreciated. Sincerely, Erin Erin Hodgess, PhD mailto: erinm.hodg...@gmail.com On Mon, Aug 27, 2018 at 2:29 PM Kevin Ushey wrote: > Hi Erin, > > The R extensions manual should be helpful here -- in particular, see the > sections on the 'src/Makevars' file and configuration of the Fortran > compiler: > > https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-Makevars > > In particular, in your 'src/Makevars' file, you'll likely want to set the > FC variable to choose the Fortran compiler, and FCFLAGS to choose the > appropriate compilation flags. As an example, you can see how one might set > these flags to compile code with the address sanitizer active: > > > https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-Address-Sanitizer > > (Note that this example uses the user-level '~/.R/Makevars' file rather > than the package-level 'src/Makevars' file, but the crux is the same) > > Hope this gets you on the right track, > Kevin > > On Mon, Aug 27, 2018 at 1:15 PM Erin Hodgess > wrote: > >> Hello! >> >> I'm building a package on Windows 10 which uses the PGI Fortran compiler. >> I have a Makefile set up in the src directory, and when I run "make all", >> the subroutines compile beautifully. >> >> Now I want to create my package. When I run R CMD check, naturally, >> gfortran is called. Is there a way that I can direct the check function >> to >> just use the make/Makefile, please? >> >> Or am I out of luck, please? (highly probable) >> >> Thanks, >> Erin >> >> Erin Hodgess, PhD >> mailto: erinm.hodg...@gmail.com >> >> [[alternative HTML version deleted]] >> >> __ >> R-package-devel@r-project.org mailing list >> https://stat.ethz.ch/mailman/listinfo/r-package-devel >> > [[alternative HTML version deleted]] __ R-package-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-package-devel
Re: [R-pkg-devel] Making adjustments to the R CMD check/install commands
Hi Erin, The R extensions manual should be helpful here -- in particular, see the sections on the 'src/Makevars' file and configuration of the Fortran compiler: https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-Makevars In particular, in your 'src/Makevars' file, you'll likely want to set the FC variable to choose the Fortran compiler, and FCFLAGS to choose the appropriate compilation flags. As an example, you can see how one might set these flags to compile code with the address sanitizer active: https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-Address-Sanitizer (Note that this example uses the user-level '~/.R/Makevars' file rather than the package-level 'src/Makevars' file, but the crux is the same) Hope this gets you on the right track, Kevin On Mon, Aug 27, 2018 at 1:15 PM Erin Hodgess wrote: > Hello! > > I'm building a package on Windows 10 which uses the PGI Fortran compiler. > I have a Makefile set up in the src directory, and when I run "make all", > the subroutines compile beautifully. > > Now I want to create my package. When I run R CMD check, naturally, > gfortran is called. Is there a way that I can direct the check function to > just use the make/Makefile, please? > > Or am I out of luck, please? (highly probable) > > Thanks, > Erin > > Erin Hodgess, PhD > mailto: erinm.hodg...@gmail.com > > [[alternative HTML version deleted]] > > __ > R-package-devel@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-package-devel > [[alternative HTML version deleted]] __ R-package-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-package-devel
