在R中执行库（程序包）会报告该程序包是为i386构建的，是否可以将其安装在x86_64系统上？

Question

I'm attempting to use an add-on R package UBbipl but am unable to add it as a library from within R once installed. Executing library(UBbipl) returns

Error: package ‘UBbipl’ was built for i386-pc-mingw32

The README (see here ) for the package states that i386 and x86_64 architectures are supported. I've consulted tutorials on creating packages (amongst other things) but I don't know enough about R (or its packages) to diagnose the problem (I think).

R.Version()$arch returns x86_64 , I'm running R version 2.14.1 and Sys.info() confirms this architecture for my machine, as well as the fact that I'm running Linux. To install it I created a tar.gz file from the zip file and ran R CMD INSTALL UBbipl_1.0.tar.gz as root. This completes successfully, reporting

* installing to library ‘/usr/local/lib/R/site-library’
* installing *binary* package ‘UBbipl’ ...
* DONE (UBbipl)

I used the --no-multiarch and --arch x86_64 options but this has not effect on the error message. Additionally, the README states that one should

Download and save the appropriate version of UBbipl to your computer. Then Install the package from within R.

but there is only one file available for download (see previous link) so I'm unsure as to whether I'm doing something wrong or if I require a different package. It would be great if someone could point me in the right direction.

Many thanks.

Answer 1

I'm going to answer my own question to close it. themel technically provided the answer in the comments. The R package is a Windows binary package built for Windows. The source package was not provided and I was thus unable to use it in R on a linux machine.

Answer 2

The file

http://www.wiley.com/legacy/wileychi/gower/supp/UBbipl_3.0.4.tar.gz

seems to be a standard R source file, with no complied elements. It has the first Fortran program source I've seen in twenty years, but ought to work. It certainly installed fine with R CMD INSTALL UBbipl_3.0.4.tar.gz once the two dependencies I did'nt already have - 'rgl' and 'numDeriv' were installed.

Regards, Anthony Staines