[R-SIG-Mac] R Versioning and Packages

Simon Urbanek simon.urbanek at r-project.org
Fri Apr 17 20:39:07 CEST 2015


> On Apr 17, 2015, at 10:19 AM, peter dalgaard <pdalgd at gmail.com> wrote:
> 
> 
> On 17 Apr 2015, at 15:23 , Duncan Murdoch <murdoch.duncan at gmail.com> wrote:
> 
>> On 17/04/2015 5:54 AM, Roy Mendelssohn - NOAA Federal wrote:
>>> Hi All:
>>> 
>>> Since R 3.2 has been released and the Mac version should be available shortly, I always forget how the R versioning system and updates relate to installed packages. Some updates work with what you have installed, some do not.  I know if the update was to R version 4.x I would likely have to re-install packages.  Will this require reinstallation, and in general at what level of update is this the case.
>> 
>> In version x.y.z, you probably won't need to update if only z changes, but you will if x or y changes, so you should update in this case.
>> 
>> An easy way to do this is to copy all the packages to the new installation, then run
>> 
>> update.packages(checkBuilt = TRUE)
>> 
>> This will reinstall any packages that were built under an earlier version of R. (I believe it ignores z in this choice.) The only disadvantage to doing this is that you will go to the latest version of all packages, and you might have been postponing that due to incompatible changes.
>> 
> 
> On Mac it seems to be a bit less convenient because the default setup allows installs to the system library. E.g.,  I have
> 
> $ ls /Library/Frameworks/R.framework/Versions/3.1/Resources/library/
> ADGofTest	codetools	graphics	nortest		stats4
> Ecdat		colorspace	grid		numDeriv	stringr
> Ecfun		compiler	gsl		parallel	survival
> ISwR		copula		gss		plyr		tcltk
> KernSmooth	datasets	gtable		proto		timeDate
> Lahman		dichromat	labeling	pspline		timeSeries
> MASS		digest		lattice		quadprog	tools
> Matrix		evir		lmom		reshape2	translations
> RColorBrewer	fBasics		manipulate	rpart		tseries
> Rcpp		fEcofin		methods		rstudio		urca
> XML		fGarch		mgcv		scales		utils
> base		forecast	mnormt		sn		vcd
> boot		foreign		munsell		spatial		zoo
> bootstrap	fracdiff	mvtnorm		splines
> class		ggplot2		nlme		stabledist
> cluster		grDevices	nnet		stats
> 
> and once I get around to installing the binaries for 3.2.0, it would be problematic to copy all that stuff (NB: including "base", for instance) over the default install.
> 
> Something like this may be a better idea:
> 
>> lib2 <- list.files("/Library/Frameworks/R.framework/Versions/3.0/Resources/library/")
>> lib1 <- list.files("/Library/Frameworks/R.framework/Versions/3.1/Resources/library/")
>> setdiff(lib2,lib1)
> [1] "Hmisc"          "lme4"           "rms"            "SparseM"       
> [5] "stockPortfolio" "zipfR"         
>> install.packages(setdiff(lib2,lib1))
> 
> or, of course, I could have changed the permissions on ..../library to exclude group write permission by "admin" in the first place (but frankly, I have forgotten whether one does anything related to this during standard installation.)
> 


Or for Mac users that don't want to do things by hand there is a GUI for that - just use "Select packages from R 3.1" in the Package Installer.

The separation of packages by version is intentional since they rarely work across major versions - we had too many headaches and silent segfaults.

Cheers,
Simon


FWIW: The Mac 3.2.0 binary release is slightly delayed precisely because we're trying to have a more complete set of binary packages than previously so they are all getting build against the release. For those impatient you can always get the R binaries from the Mac devel page.


> Duncan Murdoch
>> 
>> _______________________________________________
>> R-SIG-Mac mailing list
>> R-SIG-Mac at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
> 
> -- 
> Peter Dalgaard, Professor,
> Center for Statistics, Copenhagen Business School
> Solbjerg Plads 3, 2000 Frederiksberg, Denmark
> Phone: (+45)38153501
> Office: A 4.23
> Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com
> 
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
> 



More information about the R-SIG-Mac mailing list