[R-SIG-Mac] OS X builds for R-devel

[David Winsemius]

> On Feb 27, 2018, at 1:45 AM, Alexey Sergushichev <alsergbox at gmail.com> wrote:
> 
> Hi David,
> 
> My question was not about installing r-devel on Mac, but whether there are plans on having the packages available on CRAN web-site. I'm asking this, because Bioconductor build system depends on availability of OS X binaries of some packages on CRAN.

You should specify which packages you feel are not available that should be available. The two references you made were to two efforts at building packages. One of them (re: ChemmineR) was a warning, not a report of failure. And it was not a warning about a missing CRAN hosted package but rather a warning about C code in the package itself:

---
* checking whether package ‘ChemmineR’ can be installed ... WARNING
Found the following significant warnings:
  r_wrap.cc:1689:14: warning: explicitly assigning value of variable of type 'void *' to itself [-Wself-assign]
See ‘/Users/biocbuild/bbs-3.7-bioc/meat/ChemmineR.Rcheck/00install.out’ for details.
---

The other  ( re: cbaf) was a report of error, but the error was the absence on the build machine of the 'xlsx'-package. Why that package is missing on that particular machine is surely not something that the maintainers of the Mac fork can be expected to know. That package has been available on CRAN for many years and continues to be available. Its Package Description file does not indicate any need for compilation.


-- 
David.


> 
> --
> Alexey
> 
> On Fri, Feb 23, 2018 at 8:01 PM, David Winsemius <dwinsemius at comcast.net> wrote:
> 
> > On Feb 19, 2018, at 12:27 PM, Alexey Sergushichev <alsergbox at gmail.com> wrote:
> >
> > Dear all,
> >
> > Are there any plans on having package build for r-devel on CRAN? I
> > understand it's far from the highest priority but it would be good to have
> > any estimates. I'm aware of several packages on Bioconoductor-devel that
> > fail builds due to absence of binary versions of some packages (
> > http://bioconductor.org/checkResults/devel/bioc-LATEST/ChemmineR/merida2-buildsrc.html
> > http://bioconductor.org/checkResults/devel/bioc-LATEST/cbaf/merida2-buildsrc.html
> > ).
> >
> > Thanks,
> > Alexey Sergushichev
> >
> >       [[alternative HTML version deleted]]
> 
> The fact that you post in HTML and make no effort to describe your setup makes me suspect you may also have missed the instructions about installing XCode and configuring R for compiling source packages. This can still be done with the binary distribution versions of R and R.app, and do not require that you build R from sources. See the CRAN-Mac-tools page:
> 
> https://cran.r-project.org/bin/macosx/tools/
> 
> It used to be the case that one needed Xcode and Command Line Tools installed. (I do have them.) I cannot tell from the current Mac-devel page whether that is still the case. I have a current version of R running in El Capitan. This is my ~/.R/Makevars file:
> 
> # created by rstan at Thu Mar 19 11:26:15 2015
> CXXFLAGS = -mtune=core2 -O3 $(LTO)  #set_by_rstan
> R_XTRA_CPPFLAGS =  -I$(R_INCLUDE_DIR)   #set_by_rstan
> CC=/usr/local/clang4/bin/clang
> CXX=/usr/local/clang4/bin/clang++
> LDFLAGS=-L/usr/local/clang4/libI
> 
> I just executed the following commands in the R.app environment.
> 
> source("https://bioconductor.org/biocLite.R")
> biocLite()
> # declined the offer to update packages
>  biocLite(c("ChemmineR"), type="source", dependencies=TRUE)
> 
> It then tried to install all the dependencies (failing for ChemmineOB because the OpenBabel system requirement was not met) and got a  report of success with ChemmineR and the package was capable of loading. So I've successfully tested my theory that the only barriers to successful compilation of ChemmineR from source are lack of knowledge of the R and Bioconductor ecosystem (and lack of one external package.) I'm not that interested in getting a full package installed, so haven't taken the time to install OpenBabel and recompile ChemmineOB.
> 
> 
> Hope this helps;
> David.
> 
> >
> > _______________________________________________
> > R-SIG-Mac mailing list
> > R-SIG-Mac at r-project.org
> > https://stat.ethz.ch/mailman/listinfo/r-sig-mac
> 
> David Winsemius
> Alameda, CA, USA
> 
> 'Any technology distinguishable from magic is insufficiently advanced.'   -Gehm's Corollary to Clarke's Third Law
> 
> 
> 
> 
> 
> 

David Winsemius
Alameda, CA, USA

'Any technology distinguishable from magic is insufficiently advanced.'   -Gehm's Corollary to Clarke's Third Law