Hello all, I'm trying hard to make R CMD SHLIB work on Windows XP (32-bit) - please note that I don't have the admin rights on the computer. In terms of setup, I have followed the procedure : => installed Rtools 2.14 => changed my path : PATH=C:\Trading\R\RTools\bin;C:\Trading\R\RTools\MinGW\bin;C:\Trading\R\R-2.14.1\bin\i386;C:\WINDOWS\system32;%OTHER_PATHS% But when I try R CMD SHLIB *.* I have litterally nothing, no output at all. (and R CMD INSTALL debug tells me, of course that the operation failed while calling R CMD SHLIB) I tried the command R CMD SHLIB -n, which gives me : make -f "C:/Trading/R/R-2.14.1/etc/i386/Makeconf" -f "C:/Trading/R/R-2.14.1/share/make/winshlib.mk" SHLIB= OBJECTS= So I had a closer look at these two files, and found some strange things in Makeconf, for example a variable LOCAL_SOFT with a d: path which has nothing to do with my configuration : LOCAL_SOFT = d:/RCompile/CRANpkg/extralibs/local Ok, the thing is, my Makeconf is strange, but when I actually use the """ make -f "C:/Trading/R/R-2.14.1/etc/i386/Makeconf" -f "C:/Trading/R/R-2.14.1/share/make/winshlib.mk" SHLIB= OBJECTS= """ directly in the command prompt, it works more or less. I managed to build a simple dll based on C code (convolve2.dll, from the example), but couldn't build any dll based on c++ code. Is my SHLIB properly set up ? Why do I have a strange Makeconf ? Should I try to find a proper version of the file ? or reinstall R ? Does anyone have any idea how to fix that ? Been trying to fix it for a long time now ! Thanks in advance ! Lorenzo -- View this message in context: http://r.789695.n4.nabble.com/R-CMD-SHLIB-in-Windows-XP-No-output-at-all-tp4385797p4385797.html Sent from the R devel mailing list archive at Nabble.com.
On 13 February 2012 at 16:09, Lorenzo wrote: | Hello all, | | I'm trying hard to make R CMD SHLIB work on Windows XP (32-bit) - please Have a look at the countless examples and hundreds of unit tests in the Rcpp package. Almost all of them work by using the inline package to compile, link and load shared libraries on the fly. And yes, it all works fine on Windows XP. [ We do here that people on Windows 7 have to fight with security and admin settings though. ] You "merely" need to install Rtools right and get your PATH settings sorted out. This is documented in numerous places. Dirk | note that I don't have the admin rights on the computer. | In terms of setup, I have followed the procedure : | => installed Rtools 2.14 | => changed my path : | PATH=C:\Trading\R\RTools\bin;C:\Trading\R\RTools\MinGW\bin;C:\Trading\R\R-2.14.1\bin\i386;C:\WINDOWS\system32;%OTHER_PATHS% | | But when I try R CMD SHLIB *.* I have litterally nothing, no output at all. | (and R CMD INSTALL debug tells me, of course that the operation failed while | calling R CMD SHLIB) | I tried the command R CMD SHLIB -n, which gives me : | make -f "C:/Trading/R/R-2.14.1/etc/i386/Makeconf" -f | "C:/Trading/R/R-2.14.1/share/make/winshlib.mk" SHLIB= OBJECTS= | | So I had a closer look at these two files, and found some strange things in | Makeconf, for example a variable LOCAL_SOFT with a d: path which has nothing | to do with my configuration : | LOCAL_SOFT = d:/RCompile/CRANpkg/extralibs/local | | Ok, the thing is, my Makeconf is strange, but when I actually use the | """ make -f "C:/Trading/R/R-2.14.1/etc/i386/Makeconf" -f | "C:/Trading/R/R-2.14.1/share/make/winshlib.mk" SHLIB= OBJECTS= """ | directly in the command prompt, it works more or less. | I managed to build a simple dll based on C code (convolve2.dll, from the | example), but couldn't build any dll based on c++ code. | | Is my SHLIB properly set up ? Why do I have a strange Makeconf ? Should I | try to find a proper version of the file ? or reinstall R ? | | Does anyone have any idea how to fix that ? | Been trying to fix it for a long time now ! | | Thanks in advance ! | | Lorenzo | | -- | View this message in context: http://r.789695.n4.nabble.com/R-CMD-SHLIB-in-Windows-XP-No-output-at-all-tp4385797p4385797.html | Sent from the R devel mailing list archive at Nabble.com. | | ______________________________________________ | R-devel at r-project.org mailing list | https://stat.ethz.ch/mailman/listinfo/r-devel -- "Outside of a dog, a book is a man's best friend. Inside of a dog, it is too dark to read." -- Groucho Marx
On 13/02/2012 7:09 PM, Lorenzo wrote:> Hello all, > > I'm trying hard to make R CMD SHLIB work on Windows XP (32-bit) - please > note that I don't have the admin rights on the computer. > In terms of setup, I have followed the procedure : > => installed Rtools 2.14 > => changed my path : > PATH=C:\Trading\R\RTools\bin;C:\Trading\R\RTools\MinGW\bin;C:\Trading\R\R-2.14.1\bin\i386;C:\WINDOWS\system32;%OTHER_PATHS% > > But when I try R CMD SHLIB *.* I have litterally nothing, no output at all. > (and R CMD INSTALL debug tells me, of course that the operation failed while > calling R CMD SHLIB) > I tried the command R CMD SHLIB -n, which gives me : > make -f "C:/Trading/R/R-2.14.1/etc/i386/Makeconf" -f > "C:/Trading/R/R-2.14.1/share/make/winshlib.mk" SHLIB= OBJECTS> > So I had a closer look at these two files, and found some strange things in > Makeconf, for example a variable LOCAL_SOFT with a d: path which has nothing > to do with my configuration : > LOCAL_SOFT = d:/RCompile/CRANpkg/extralibs/localThat's a path on the machine which builds the R binary distributions. It shouldn't have leaked out, but it should be harmless. I think the problem might be with the "*.*". On a Unix system, the shell expands wildcards, but on Windows, the individual commands are supposed to do it. But our commands generally follow the Unix conventions, so perhaps nobody is expanding that, and nothing is being requested. (When I try it in R-patched, it works. But I'm not using that version of Rtools, and you don't say what version of R you're using.) If you use a bash shell (e.g. from Cygwin) then it will follow the Unix convention, and that will probably work. That's what I normally do. Alternatively, you can just list the files that you want to compile. If it's a long list, I think there are Windows ways to put the whole list into a shell variable and put the shell variable on that line, but I forget what they are. I would certainly try that for debugging purposes. Duncan Murdoch> Ok, the thing is, my Makeconf is strange, but when I actually use the > """ make -f "C:/Trading/R/R-2.14.1/etc/i386/Makeconf" -f > "C:/Trading/R/R-2.14.1/share/make/winshlib.mk" SHLIB= OBJECTS= """ > directly in the command prompt, it works more or less. > I managed to build a simple dll based on C code (convolve2.dll, from the > example), but couldn't build any dll based on c++ code. > > Is my SHLIB properly set up ? Why do I have a strange Makeconf ? Should I > try to find a proper version of the file ? or reinstall R ? > > Does anyone have any idea how to fix that ? > Been trying to fix it for a long time now ! > > Thanks in advance ! > > Lorenzo > > -- > View this message in context: http://r.789695.n4.nabble.com/R-CMD-SHLIB-in-Windows-XP-No-output-at-all-tp4385797p4385797.html > Sent from the R devel mailing list archive at Nabble.com. > > ______________________________________________ > R-devel at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-devel
Reasonably Related Threads
- RCMD SHLIB problem in 64 bits
- LDFLAGS defined in R_MAKEVARS_USER file is ignored for R CMD SHLIB on Windows
- R Compilation gets stuck on Windows 64
- Setting up a windows system for rcpp
- Problems with installing R packages from source and running C++ in R, even on fresh R installation