Skip to content

problems in installing packages with R version 2.4.0 alpha (2006-09-05 r39134)

6 messages · Richard M. Heiberger, Brian Ripley, Uwe Ligges +2 more

Original
Richard M. Heiberger
# 
I just downloaded the windows version 
 R version 2.4.0 alpha (2006-09-05 r39134)

1. When I downloaded the packages, the following two were not found.
--- Please select a CRAN mirror for use in this session ---
dependency ''fCalendar'' is not available

dependency ''SparseM'' is not available


I am not sure which other packages depend on these.


2. Of the others, several didn't install cleanly

package 'mlbench' successfully unpacked and MD5 sums checked
Warning: unable to move temporary installation 'C:\Program Files\R\R-
2.4.0alpha\library\file41bb5af1\mlbench' to 'C:\Program Files\R\R-2.4.0alpha\library\mlbench'
package 'tseries' successfully unpacked and MD5 sums checked
Warning: unable to move temporary installation 'C:\Program Files\R\R-2.4.0alpha\library\file305e124
\tseries' to 'C:\Program Files\R\R-2.4.0alpha\library\tseries'
package 'DAAG' successfully unpacked and MD5 sums checked
Warning: unable to move temporary installation 'C:\Program Files\R\R-
2.4.0alpha\library\file491c440d\DAAG' to 'C:\Program Files\R\R-2.4.0alpha\library\DAAG'
package 'acepack' successfully unpacked and MD5 sums checked
Warning: unable to move temporary installation 'C:\Program Files\R\R-
2.4.0alpha\library\file767d7a5a\acepack' to 'C:\Program Files\R\R-2.4.0alpha\library\acepack'
package 'rcom' successfully unpacked and MD5 sums checked
Warning: unable to move temporary installation 'C:\Program Files\R\R-
2.4.0alpha\library\file58784b40\rcom' to 'C:\Program Files\R\R-2.4.0alpha\library\rcom'


I reinstalled all five from the local zip files that were downloaded as part of the above
installation, and four worked and one didn't.
I needed to reinstall mlbench a third time and this time it worked.

This type of problem frequently happens when I download a new version of R.


3. rgl is still not working in Windows

package 'rgl' successfully unpacked and MD5 sums checked
Error in dyn.load(x, as.logical(local), as.logical(now)) : 
        unable to load shared library 'C:/PROGRA~1/R/R-24~1.0AL/library/rgl/libs/rgl.dll':
  LoadLibrary failure:  The specified procedure could not be found.

Error: package/namespace load failed for 'rgl'
Therefore the Rcmdr 3-d graphics isn't working.


Rich
Brian Ripley
# 
At least some of this will go away if you use a current version of 2.4.0 
alpha rather than one that is a week old (as the posting guide does ask).  
We are now at r39258, and some of those binary packages were built against 
a substantially later version of 2.4.0 alpha.

If you look at the list of available packages at

http://cran.r-project.org/bin/windows/contrib/checkSummaryWin.htm

you will see which are unavailable as Windows binaries and why.  More can 
be installed from the sources but do not pass R CMD check for Uwe (or in 
most cases for anyone else: see the other CRAN summaries).

This is the 'alpha' period, and we are waiting for updates on quite a 
number of packages (especially ones that use S4 methods).  However, R 
2.4.0 is not even in feature freeze yet, and it is quite possible that 
packages will need to be reinstalled again prior to release.
On Mon, 11 Sep 2006, Richard M. Heiberger wrote:

            

      
      
      
    

        
This suggests problems with your local file system: it is definitely a 
Windows rather than an R problem.

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Martin Maechler
    

    

      
      #
    

  


  

          

      
      
      
    

        
BDR> At least some of this will go away if you use a current
    BDR> version of 2.4.0 alpha rather than one that is a week
    BDR> old (as the posting guide does ask).  We are now at
    BDR> r39258, and some of those binary packages were built
    BDR> against a substantially later version of 2.4.0 alpha.

    [............]

Still "Thank you, Rich!" for running R-alpha and your report.
We are grateful for everyone who does use R-alpha and reports
"unusual" findings, even if some of them are already known to
the R developers.

Martin

  
  


  


      

    

    

      
      

        


  

        

      Uwe Ligges
    

    

      
      #
    

  


  

      
Prof Brian Ripley wrote:

            

      
      
      
    

        
Currently, the main problems with many packages are missing *recent* 
binaries (after the latest S4 changes) of BioC packages, hence many 
dependencies may not be working. I do not want to compile the whole BioC 
repository as well.

Uwe

            

      
      
      
    

        

      
      
    

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Martin Maechler
    

    

      
      #
    

  


  

          

      
      
      
    

        
>>>>> "BDR" == Prof Brian Ripley <ripley at stats.ox.ac.uk>
  >>>>>     on Mon, 11 Sep 2006 22:25:58 +0100 (BST) writes:

       BDR> At least some of this will go away if you use a current
       BDR> version of 2.4.0 alpha rather than one that is a week
       BDR> old (as the posting guide does ask).  We are now at
       BDR> r39258, and some of those binary packages were built
       BDR> against a substantially later version of 2.4.0 alpha.

       BDR> [............]

    MM> Still "Thank you, Rich!" for running R-alpha and your report.
    MM> We are grateful for everyone who does use R-alpha and reports
    MM> "unusual" findings, even if some of them are already known to
    MM> the R developers.

I have really failed to add the following:

The fact that you get *binary* windows packages for R-alpha **at all**
is already remarkable and only the result of extensive, persistent and
diligent work by Uwe Ligges and Brian Ripley.
They shall be thanked thoroughly hereby!

[ For alpha-testing, I'd typically expect anyone to build
  anything from the sources -- which of course is quite a
  requirement for users of R for windows. ]

Martin

  
  


  


      

    

    

      
      

        


  

        

      Seth Falcon
    

    

      
      #
    

  


  

      
Uwe Ligges <ligges at statistik.uni-dortmund.de> writes:

            

      
      
      
    

        
The Bioconductor build systems have been updated to a recent R 2.4.0
alpha and binary packages should be available Real Soon Now.

+ seth