Re: [stan-users] Re: I cannot compile source code of version 2.0 in windows
23 views
Skip to first unread message
Bob Carpenter
Oct 18, 2013, 10:13:44 PM10/18/13
to stan...@googlegroups.com
[moved to stan-dev]
You should be able to release RStan 2.0.1 at any point now.
That's one of the reasons we wanted to decouple Stan and RStan.
Whatever gets worked out on the Stan side for adaptation won't
affect the RStan interface.
Our hope is that any RStan with the same major and minor version (2.0
here) will work with any Stan version with matching major and minor
versions.
- Bob
On 10/18/13 9:44 PM, Jiqiang Guo wrote:
> It seems that I was not understood. Essentially, there is a makefile for building some of the c++ source for rstan
> package. The makefile uses 'find' to locate the folder of boost and eigen inside the rstan package. But for some
> reason, this part failed for you.
>
> Anyway, I am gonna fix the path for boost and eigen as what is done in 1.3.0, so for the next uploaded version, you
> would not have this problem. I could release a 2.0.1 for rstan after the issue in stan 2.0 is sorted out.
>
> --
> Jiqiang
>
>
> On Fri, Oct 18, 2013 at 5:44 PM, Juan Miguel Marin <juanmigu...@gmail.com <mailto:juanmigu...@gmail.com>> wrote:
>
> Hi Daniel,
>
> yes it is puzzling :-(
> I reinstalled R and Rtools without success at last.
>
> There a possible solution. In version 1.3.0 after compiling it, I had the rstan folder (with the corresponding
> *.dll) in the main library folder of R.
> Well, if I zipped directly this rstan folder and copy and unzipped in other computer (with Rcpp and Inline) rstan
> worked fine without needing to compile it.
>
> So If you (or someone) have an installed version of rstan in a windows machine it may possible to zip the folder and
> try to uncompress it in my machine. Perhaps it could worl like in the case of version 1.3.0
> As the zipped folder will be large, it might be transfered via the "Public" folder of dropbox .
>
> I got the same error of installation in three different computers :-/
>
> Anyway, 'my' environment variables are
>
> > names(s <- Sys.getenv())
> [1] "ALLUSERSPROFILE" "APPDATA"
> [3] "APR_ICONV_PATH" "CommonProgramFiles"
> [5] "CommonProgramFiles(x86)" "CommonProgramW6432"
> [7] "COMPUTERNAME" "ComSpec"
> [9] "FP_NO_HOST_CHECK" "GFORTRAN_STDERR_UNIT"
> [11] "GFORTRAN_STDOUT_UNIT" "HOME"
> [13] "HOMEDRIVE" "HOMEPATH"
> [15] "LOCALAPPDATA" "LOGONSERVER"
> [17] "MOZ_PLUGIN_PATH" "MuPAD_SWP_310"
> [19] "NUMBER_OF_PROCESSORS" "OS"
> [21] "PATH" "PATHEXT"
> [23] "PROCESSOR_ARCHITECTURE" "PROCESSOR_IDENTIFIER"
> [25] "PROCESSOR_LEVEL" "PROCESSOR_REVISION"
> [27] "ProgramData" "ProgramFiles"
> [29] "ProgramFiles(x86)" "ProgramW6432"
> [31] "PSModulePath" "PUBLIC"
> [33] "R_ARCH" "R_HOME"
> [35] "R_LIBS_USER" "R_USER"
> [37] "SESSIONNAME" "SVN_EDITOR"
> [39] "SystemDrive" "SystemRoot"
> [41] "TEMP" "TMP"
> [43] "USERDOMAIN" "USERNAME"
> [45] "USERPROFILE" "windir"
>
> El viernes, 18 de octubre de 2013 20:37:43 UTC+2, Daniel Lee escribi�ｿｽ:
>
> Hi Juan,
>
> I think the problem is with your configuration of R 3.0.2. The steps you followed to package RStan just
> repackages RStan and should be identical to downloading the rstan_2.0.0.tar.gz from
> https://github.com/stan-dev/__rstan/releases/download/v2.0.__0/rstan_2.0.0.tar.gz
> <https://github.com/stan-dev/rstan/releases/download/v2.0.0/rstan_2.0.0.tar.gz>.
>
> When you install RStan from within R, your compiler does this:
>
> g++ -m64 -I"C:/R-30~1.2/include" -DNDEBUG -I"../inst/include/stansrc" -I"" -I"/unsupported" -I
> -I"../inst/include" -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DECLTYPE -DBOOST_DISABLE_ASSERTS
> -I"C:/R-3.0.2/library/Rcpp/__include" -I"d:/RCompile/CRANpkg/__extralibs64/local/include" -O3 -c
> stanc.cpp -o stanc.o
>
>
> When it's installed properly, it should look like:
>
> g++ -m32 -I"C:/R/R-30~1.2/include" -I"../inst/include/stansrc"
> -I"../inst/include/stanlib/__eigen_3.2.0" -I"../inst/include/stanlib/__eigen_3.2.0/unsupported"
> -I../inst/include/stanlib/__boost_1.54.0 -I"../inst/include" -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DECLTYPE
> -DBOOST_DISABLE_ASSERTS -I"C:/R/R-3.0.2/library/Rcpp/__include"
> -I"d:/RCompile/CRANpkg/__extralibs64/local/include" -O3 -Wall -mtune=core2 -c stanc.cpp -o stanc.o
>
>
>
> These are the differences, correct on the left, your configuration on the right:
>
> 1. -I"../inst/include/stanlib/__eigen_3.2.0" vs -I""
> 2. -I"../inst/include/stanlib/__eigen_3.2.0/unsupported" vs -I"/unsupported"
> 3. -I../inst/include/stanlib/__boost_1.54.0 vs -I
>
>
> So... I'm at a loss now. I'm not sure how the Boost and Eigen library locations in your makefiles have gotten lost.
>
> I think you mentioned that you reinstalled R. Is that true? Perhaps there's some interaction with environment
> variables stomping on the library locations? Can you check which environment variables have been set?
>
> Do you know of a way we can replicate your setup?
>
>
>
>
> Daniel
>
>
>
>
>
> On Fri, Oct 18, 2013 at 12:52 PM, Juan Miguel Marin <juanmigu...@gmail.com> wrote:
>
>
> I followed the instructions in
>
> https://github.com/stan-dev/__rstan/wiki/How-to-build-rstan-__package%3F
> <https://github.com/stan-dev/rstan/wiki/How-to-build-rstan-package%3F>
>
> I think without any relationship with R installing
>
> But everything remains the same...
>
> C:\kk\rstan-develop>cd rstan
> C:\kk\rstan-develop\rstan>ls
> INSTALL.txt R_Makevars example makefile rstan tests
>
> C:\kk\rstan-develop\rstan>make build
> uname: not found
> uname: not found
> FIND: formato de par�ｿｽmetros incorrecto
> FIND: formato de par�ｿｽmetros incorrecto
> INFORMACI�ｿｽN: no se pudo encontrar ning�ｿｽn archivo para los patrones dados.
> INFORMACI�ｿｽN: no se pudo encontrar ning�ｿｽn archivo para los patrones dados.
> v�ｿｽnculo simb�ｿｽlico creado para .\rstan\inst\include\stanlib <<===>> ..\..\..\..\s
> tan\lib
> v�ｿｽnculo simb�ｿｽlico creado para .\rstan\inst\include\stansrc <<===>> ..\..\..\..\s
> tan\src
> R --vanilla CMD build rstan --md5 # --no-build-vignettes --no-manual
> * checking for file 'rstan/DESCRIPTION' ... OK
> * preparing 'rstan':
> * checking DESCRIPTION meta-information ... OK
> * cleaning src
> * installing the package to process help pages
> ------------------------------__-----
>
> * installing *source* package 'rstan' ...
> ** libs
> running 'src/Makefile.win' ...
>
> FIND: formato de par metros incorrecto
> FIND: formato de par metros incorrecto
> FIND: formato de par metros incorrecto
> g++ -m64 -I"C:/R-30~1.2/include" -DNDEBUG -I"../inst/include/stansrc" -I"" -I
>
> "/unsupported" -I -I"../inst/include" -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DEC
> LTYPE -DBOOST_DISABLE_ASSERTS -I"C:/R-3.0.2/library/Rcpp/__include" -I"d:/RCompi
> le/CRANpkg/extralibs64/local/__include" -O3 -c stanc.cpp -o stanc.o
> stanc.cpp:4:28: fatal error: stan/version.hpp: No such file or directory
> compilation terminated.
> make[1]: *** [stanc.o] Error 1
>
> ERROR: compilation failed for package 'rstan'
> * removing 'C:/Users/jmmarin/AppData/__Local/Temp/Rtmpi0sOgy/__Rinst1d085236390e/rst
> an'
> ------------------------------__-----
> ERROR: package installation failed
> make: *** [build] Error 1
>
>
>
>
> El viernes, 18 de octubre de 2013 10:25:53 UTC+2, Juan Miguel Marin escribi�ｿｽ:
>
> Hello Ben,
>
> yes it has non sense that line. Moreover, I have not any folder like "RCompile" or "CRANpkg/extralibs64"
> anywhere, so it is impossible that the system could find it.
> I re-installed R, I uninstalled Rcpp and re-installed it afterwards. I run R witn the --vanila option
> and everything remains the same, with the same error :-(
>
> So I cannot do anything... I get back to version 1-3.0 that works OK :P
>
> Thank you anyway for your time :D
>
> El viernes, 18 de octubre de 2013 03:19:50 UTC+2, Ben Goodrich escribi�ｿｽ:
>
> On Thursday, October 17, 2013 11:14:43 AM UTC-4, Juan Miguel Marin wrote:
>
> g++ -m32 -I"C:/R-30~1.2/include" -I"../inst/include/stansrc" -I"" -I"/unsupported" -I
> -I"../inst/include" -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DECLTYPE -DBOOST_DISABLE_ASSERTS
> -I"C:/R-3.0.2/library/Rcpp/__inc__lude"
> -I"d:/RCompile/CRANpkg/__extrali__bs64/local/include" -O3 -Wall -mtune=core2 -c stanc.cpp
> -o stanc.o
> In file included from ../inst/include/stansrc/stan/__g__m/compiler.hpp:4:0,
> from stanc.cpp:5:
> ../inst/include/stansrc/stan/__g__m/ast.hpp:10:47: fatal error:
> boost/variant/recursive___varian__t.hpp: No such file or directory
> compilation terminated.
> make: *** [stanc.o] Error 1
>
>
> I'm going to guess the culprit is
>
> |
> -I"d:/RCompile/CRANpkg/__extrali__bs64/local/include"
> |
>
> Do you maybe have another (part of) boost in that directory? I think since it comes last, it is
> searched first so if there is some boost there, then it may be looking in the wrong place for its
> dependencies.
>
> In general, can you try installing rstan after starting R with the --vanilla option?
>
> Ben
>
> --
> You received this message because you are subscribed to the Google Groups "stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
> stan-users+...@__googlegroups.com.
>
> For more options, visit https://groups.google.com/__groups/opt_out <https://groups.google.com/groups/opt_out>.
>
>
> --
> You received this message because you are subscribed to the Google Groups "stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
> stan-users+...@googlegroups.com <mailto:stan-users%2Bunsu...@googlegroups.com>.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>
> --
> You received this message because you are subscribed to the Google Groups "stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
You should be able to release RStan 2.0.1 at any point now.
That's one of the reasons we wanted to decouple Stan and RStan.
Whatever gets worked out on the Stan side for adaptation won't
affect the RStan interface.
Our hope is that any RStan with the same major and minor version (2.0
here) will work with any Stan version with matching major and minor
versions.
- Bob
On 10/18/13 9:44 PM, Jiqiang Guo wrote:
> It seems that I was not understood. Essentially, there is a makefile for building some of the c++ source for rstan
> package. The makefile uses 'find' to locate the folder of boost and eigen inside the rstan package. But for some
> reason, this part failed for you.
>
> Anyway, I am gonna fix the path for boost and eigen as what is done in 1.3.0, so for the next uploaded version, you
> would not have this problem. I could release a 2.0.1 for rstan after the issue in stan 2.0 is sorted out.
>
> --
> Jiqiang
>
>
> On Fri, Oct 18, 2013 at 5:44 PM, Juan Miguel Marin <juanmigu...@gmail.com <mailto:juanmigu...@gmail.com>> wrote:
>
> Hi Daniel,
>
> yes it is puzzling :-(
> I reinstalled R and Rtools without success at last.
>
> There a possible solution. In version 1.3.0 after compiling it, I had the rstan folder (with the corresponding
> *.dll) in the main library folder of R.
> Well, if I zipped directly this rstan folder and copy and unzipped in other computer (with Rcpp and Inline) rstan
> worked fine without needing to compile it.
>
> So If you (or someone) have an installed version of rstan in a windows machine it may possible to zip the folder and
> try to uncompress it in my machine. Perhaps it could worl like in the case of version 1.3.0
> As the zipped folder will be large, it might be transfered via the "Public" folder of dropbox .
>
> I got the same error of installation in three different computers :-/
>
> Anyway, 'my' environment variables are
>
> > names(s <- Sys.getenv())
> [1] "ALLUSERSPROFILE" "APPDATA"
> [3] "APR_ICONV_PATH" "CommonProgramFiles"
> [5] "CommonProgramFiles(x86)" "CommonProgramW6432"
> [7] "COMPUTERNAME" "ComSpec"
> [9] "FP_NO_HOST_CHECK" "GFORTRAN_STDERR_UNIT"
> [11] "GFORTRAN_STDOUT_UNIT" "HOME"
> [13] "HOMEDRIVE" "HOMEPATH"
> [15] "LOCALAPPDATA" "LOGONSERVER"
> [17] "MOZ_PLUGIN_PATH" "MuPAD_SWP_310"
> [19] "NUMBER_OF_PROCESSORS" "OS"
> [21] "PATH" "PATHEXT"
> [23] "PROCESSOR_ARCHITECTURE" "PROCESSOR_IDENTIFIER"
> [25] "PROCESSOR_LEVEL" "PROCESSOR_REVISION"
> [27] "ProgramData" "ProgramFiles"
> [29] "ProgramFiles(x86)" "ProgramW6432"
> [31] "PSModulePath" "PUBLIC"
> [33] "R_ARCH" "R_HOME"
> [35] "R_LIBS_USER" "R_USER"
> [37] "SESSIONNAME" "SVN_EDITOR"
> [39] "SystemDrive" "SystemRoot"
> [41] "TEMP" "TMP"
> [43] "USERDOMAIN" "USERNAME"
> [45] "USERPROFILE" "windir"
>
> El viernes, 18 de octubre de 2013 20:37:43 UTC+2, Daniel Lee escribi�ｿｽ:
>
> Hi Juan,
>
> I think the problem is with your configuration of R 3.0.2. The steps you followed to package RStan just
> repackages RStan and should be identical to downloading the rstan_2.0.0.tar.gz from
> https://github.com/stan-dev/__rstan/releases/download/v2.0.__0/rstan_2.0.0.tar.gz
> <https://github.com/stan-dev/rstan/releases/download/v2.0.0/rstan_2.0.0.tar.gz>.
>
> When you install RStan from within R, your compiler does this:
>
> g++ -m64 -I"C:/R-30~1.2/include" -DNDEBUG -I"../inst/include/stansrc" -I"" -I"/unsupported" -I
> -I"../inst/include" -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DECLTYPE -DBOOST_DISABLE_ASSERTS
> -I"C:/R-3.0.2/library/Rcpp/__include" -I"d:/RCompile/CRANpkg/__extralibs64/local/include" -O3 -c
> stanc.cpp -o stanc.o
>
>
> When it's installed properly, it should look like:
>
> g++ -m32 -I"C:/R/R-30~1.2/include" -I"../inst/include/stansrc"
> -I"../inst/include/stanlib/__eigen_3.2.0" -I"../inst/include/stanlib/__eigen_3.2.0/unsupported"
> -I../inst/include/stanlib/__boost_1.54.0 -I"../inst/include" -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DECLTYPE
> -DBOOST_DISABLE_ASSERTS -I"C:/R/R-3.0.2/library/Rcpp/__include"
> -I"d:/RCompile/CRANpkg/__extralibs64/local/include" -O3 -Wall -mtune=core2 -c stanc.cpp -o stanc.o
>
>
>
> These are the differences, correct on the left, your configuration on the right:
>
> 1. -I"../inst/include/stanlib/__eigen_3.2.0" vs -I""
> 2. -I"../inst/include/stanlib/__eigen_3.2.0/unsupported" vs -I"/unsupported"
> 3. -I../inst/include/stanlib/__boost_1.54.0 vs -I
>
>
> So... I'm at a loss now. I'm not sure how the Boost and Eigen library locations in your makefiles have gotten lost.
>
> I think you mentioned that you reinstalled R. Is that true? Perhaps there's some interaction with environment
> variables stomping on the library locations? Can you check which environment variables have been set?
>
> Do you know of a way we can replicate your setup?
>
>
>
>
> Daniel
>
>
>
>
>
> On Fri, Oct 18, 2013 at 12:52 PM, Juan Miguel Marin <juanmigu...@gmail.com> wrote:
>
>
> I followed the instructions in
>
> https://github.com/stan-dev/__rstan/wiki/How-to-build-rstan-__package%3F
> <https://github.com/stan-dev/rstan/wiki/How-to-build-rstan-package%3F>
>
> I think without any relationship with R installing
>
> But everything remains the same...
>
> C:\kk\rstan-develop>cd rstan
> C:\kk\rstan-develop\rstan>ls
> INSTALL.txt R_Makevars example makefile rstan tests
>
> C:\kk\rstan-develop\rstan>make build
> uname: not found
> uname: not found
> FIND: formato de par�ｿｽmetros incorrecto
> FIND: formato de par�ｿｽmetros incorrecto
> INFORMACI�ｿｽN: no se pudo encontrar ning�ｿｽn archivo para los patrones dados.
> INFORMACI�ｿｽN: no se pudo encontrar ning�ｿｽn archivo para los patrones dados.
> v�ｿｽnculo simb�ｿｽlico creado para .\rstan\inst\include\stanlib <<===>> ..\..\..\..\s
> tan\lib
> v�ｿｽnculo simb�ｿｽlico creado para .\rstan\inst\include\stansrc <<===>> ..\..\..\..\s
> tan\src
> R --vanilla CMD build rstan --md5 # --no-build-vignettes --no-manual
> * checking for file 'rstan/DESCRIPTION' ... OK
> * preparing 'rstan':
> * checking DESCRIPTION meta-information ... OK
> * cleaning src
> * installing the package to process help pages
> ------------------------------__-----
>
> * installing *source* package 'rstan' ...
> ** libs
> running 'src/Makefile.win' ...
>
> FIND: formato de par metros incorrecto
> FIND: formato de par metros incorrecto
> FIND: formato de par metros incorrecto
> g++ -m64 -I"C:/R-30~1.2/include" -DNDEBUG -I"../inst/include/stansrc" -I"" -I
>
> "/unsupported" -I -I"../inst/include" -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DEC
> LTYPE -DBOOST_DISABLE_ASSERTS -I"C:/R-3.0.2/library/Rcpp/__include" -I"d:/RCompi
> le/CRANpkg/extralibs64/local/__include" -O3 -c stanc.cpp -o stanc.o
> stanc.cpp:4:28: fatal error: stan/version.hpp: No such file or directory
> compilation terminated.
> make[1]: *** [stanc.o] Error 1
>
> ERROR: compilation failed for package 'rstan'
> * removing 'C:/Users/jmmarin/AppData/__Local/Temp/Rtmpi0sOgy/__Rinst1d085236390e/rst
> an'
> ------------------------------__-----
> ERROR: package installation failed
> make: *** [build] Error 1
>
>
>
>
> El viernes, 18 de octubre de 2013 10:25:53 UTC+2, Juan Miguel Marin escribi�ｿｽ:
>
> Hello Ben,
>
> yes it has non sense that line. Moreover, I have not any folder like "RCompile" or "CRANpkg/extralibs64"
> anywhere, so it is impossible that the system could find it.
> I re-installed R, I uninstalled Rcpp and re-installed it afterwards. I run R witn the --vanila option
> and everything remains the same, with the same error :-(
>
> So I cannot do anything... I get back to version 1-3.0 that works OK :P
>
> Thank you anyway for your time :D
>
> El viernes, 18 de octubre de 2013 03:19:50 UTC+2, Ben Goodrich escribi�ｿｽ:
>
> On Thursday, October 17, 2013 11:14:43 AM UTC-4, Juan Miguel Marin wrote:
>
> g++ -m32 -I"C:/R-30~1.2/include" -I"../inst/include/stansrc" -I"" -I"/unsupported" -I
> -I"../inst/include" -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DECLTYPE -DBOOST_DISABLE_ASSERTS
> -I"C:/R-3.0.2/library/Rcpp/__inc__lude"
> -I"d:/RCompile/CRANpkg/__extrali__bs64/local/include" -O3 -Wall -mtune=core2 -c stanc.cpp
> -o stanc.o
> In file included from ../inst/include/stansrc/stan/__g__m/compiler.hpp:4:0,
> from stanc.cpp:5:
> ../inst/include/stansrc/stan/__g__m/ast.hpp:10:47: fatal error:
> boost/variant/recursive___varian__t.hpp: No such file or directory
> compilation terminated.
> make: *** [stanc.o] Error 1
>
>
> I'm going to guess the culprit is
>
> |
> -I"d:/RCompile/CRANpkg/__extrali__bs64/local/include"
> |
>
> Do you maybe have another (part of) boost in that directory? I think since it comes last, it is
> searched first so if there is some boost there, then it may be looking in the wrong place for its
> dependencies.
>
> In general, can you try installing rstan after starting R with the --vanilla option?
>
> Ben
>
> --
> You received this message because you are subscribed to the Google Groups "stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
> stan-users+...@__googlegroups.com.
>
> For more options, visit https://groups.google.com/__groups/opt_out <https://groups.google.com/groups/opt_out>.
>
>
> --
> You received this message because you are subscribed to the Google Groups "stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
> stan-users+...@googlegroups.com <mailto:stan-users%2Bunsu...@googlegroups.com>.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>
> --
> You received this message because you are subscribed to the Google Groups "stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
Reply all
Reply to author
Forward
0 new messages