New Rtools sent to CRAN
Thanks Dan, this is helpful. Just to confirm: these are packages that did not have issues with the regular r-devel (3.3) on windows? For example the 'xps' only has some warnings which seem unrelated to the toolchain.
On Sun, Sep 27, 2015 at 8:59 PM, Dan Tenenbaum <dtenenba at fredhutch.org> wrote:
----- Original Message -----
From: "Dan Tenenbaum" <dtenenba at fredhutch.org> To: "Jeroen Ooms" <jeroenooms at gmail.com> Cc: r-sig-windows at r-project.org Sent: Sunday, September 27, 2015 11:57:13 AM Subject: Re: [R-SIG-Win] New Rtools sent to CRAN ----- Original Message -----
From: "Jeroen Ooms" <jeroenooms at gmail.com> To: "Dan Tenenbaum" <dtenenba at fredhutch.org> Cc: r-sig-windows at r-project.org Sent: Thursday, September 24, 2015 3:13:03 PM Subject: Re: [R-SIG-Win] New Rtools sent to CRAN Hi Dan, Currently the r-devel builds from CRAN still use the old tool chain. A copy of r-devel which has been configured with the new tool chain is available from http://www.stat.ucla.edu/~jeroen/mingw-w64/R-experimental-r69418.exe This build has BINPREF (the path to the compiler) hardcoded to the default location of the new tool chains. To change it you would need to edit this variable in C:\Program Files\R\R-experimental\etc\x64\Makeconf and C:\Program Files\R\R-experimental\etc\i386\Makeconf. We are currently discussing how to design it such that you can set the path to the compiler using an environment variable.
I ran the Bioconductor builds overnight using this toolchain and recent R-devel and did not notice anything odd with packages containing native code, which means either the new toolchain is working perfectly (which is wonderful) or gcc-4.6.3 is still being used and maybe I need to do something to faciliate the use of gcc-4.9.3. Can you advise?
We're talking to CRAN on how to coordinate the migration to the new tool chain. If you have bioc packages that link to external c++ libraries, those need a rebuild. The most current information is available from: https://github.com/rwinlib/r-base#readme
OK, I ran the Bioconductor builds with the new toolchain and the experimental version of R-devel. A number of packages fail to install/compile which worked before with regular R-devel. Some of these are fairly important packages which have a lot of reverse dependencies. The build report is here: http://bioconductor.org/checkResults/3.3/bioc-LATEST/windows2.bioconductor.org-index.html#show=timeout,error The list of packages with compilation errors not appearing in the normal build are: affxparser AffyTiling arrayQualityMetrics bamsignals ChemmineOB deepSNV flowMatch flowWorkspace GeneNetworkBuilder HilbertVisGUI mzR Rdisop xps You can view the build report for one of these packages by replacing PACKAGENAME with one of the packages above in this link: http://bioconductor.org/checkResults/3.3/bioc-LATEST/PACKAGENAME/windows2.bioconductor.org-install.html For example, for affxparser: http://bioconductor.org/checkResults/3.3/bioc-LATEST/affxparser/windows2.bioconductor.org-install.html
Actually, the error will not always be in the INSTALL phase so instead use a url like this: http://bioconductor.org/checkResults/3.3/bioc-LATEST/PACKAGENAME/ And click on the red ERROR icon(s). Dan
FYI Thanks, Dan
_______________________________________________ R-SIG-windows mailing list R-SIG-windows at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-windows