Re: [R-pkg-devel] Help with Windows build failure

2017-04-07 Thread Alexandre Sieira
So actually 'Biarch: false’ didn’t help at all, same problem: https://win-builder.r-project.org/KhXyGby5Yhjm/ Does anyone know if the win builder machines have both 32-bit and 64-bit Python installed and in the PATH? -- Alexandre Sieira CISA, CISSP, ISO 27001 Lead Auditor "The truth is rarely p

Re: [R-pkg-devel] Help with Windows build failure

2017-04-07 Thread Alexandre Sieira
Hi, Dirk. Thank you for responding. I do sympathize with your frustrations, due to the exact complexity you are mentioning and also my ignorance on how R packages compile on Windows, I had SnakeCharmR signal ‘OS_type: Unix’ for the longest time. The thing is that AFAIK this works on Windows now,

Re: [R-pkg-devel] R CMD check not finding my vignettes

2017-04-07 Thread Uwe Ligges
Can you send the tarball please, I can take a look, Uwe On 07.04.2017 12:09, David Hugh-Jones wrote: Hi Uwe, Indeed yes, but my built tarball has got the files in inst/doc as shown. If I then run 'R CMD check' on that tarball, I get the error quoted. Is there something happening that I don't u

Re: [R-pkg-devel] R CMD check not finding my vignettes

2017-04-07 Thread David Hugh-Jones
Hi Uwe, Indeed yes, but my built tarball has got the files in inst/doc as shown. If I then run 'R CMD check' on that tarball, I get the error quoted. Is there something happening that I don't understand? Incidentally, the tarball also has files in the vignettes directory: (a) the original .Rmd fi

Re: [R-pkg-devel] maintainer built vignette

2017-04-07 Thread Uwe Ligges
Please note that packages with a FOSS licences cannot have BuildVignettes: no Such FOSS Packages with a vignette (typically as pdf) need the sources that produced the vignette. Hence, if you really want to ship a vignette including code that is hardly executable on CRAN systems, reply to the au