Skip to content

Failed installWithVers from local tar.gz

3 messages · kkarr, David Winsemius, Brian Ripley

#
Hi,
I am having trouble installing a previous version of a package, and multiple
version of the package in R 2.12 or 2.15.  I have download the  tar.gz
packages from the CRAN archive, and created an folder holding all the files.

install.packages("changepoint_0.3.tar.gz",
lib="~/Rtools/dir/of/pkgs/",
repos=NULL,
dependencies=FALSE,
contriburl="~/Rtools/dir/of/pkgs/changepoint_0.3.tar.gz",
installWithVers= TRUE)


I have received several different errors, depending on which version of R I
am using
Version 2.12: 
warning: unable to access index for repository
'lib = "~/Documents/Rtools/dir/of/pkgs/"' is not writable


Thanks for any thought.
Cheers



--
View this message in context: http://r.789695.n4.nabble.com/Failed-installWithVers-from-local-tar-gz-tp4644937.html
Sent from the R help mailing list archive at Nabble.com.
#
On Oct 3, 2012, at 12:42 PM, kkarr wrote:

            
The tilde in the path makes me think this is a Mac OS question, and if so should have been explicitly stated, but the warning seems strange. Do you actually have a copy of "changepoint_0.3.tar.gz" in that location? What does this return:

list.files( path = "~/Rtools/dir/of/pkgs/")


You should also remember that type="mac.binary" is the default for install.packages on a Mac (if that is what we are dealing with) and that you would need type ="source" if installing on older versions of R. There is a binary version of pkg:changepoint at CRAN. S why not:

install.packages("changepoint") and just let the magic take hold?
#
On 03/10/2012 20:42, kkarr wrote:
installWithVers was made obsolete years ago.  And you really do not want 
to install to the directory containing the tarballs.
This list does not support ancient versions of R (and as the posting 
guide says, there was no '2.12').  Please update as you were asked to do 
before posting.
Whatever instructions you are following are misleading you.  Use the 
official manuals instead.

        
PLEASE do.