Skip to content

Error in autoloader

2 messages · Peter Ho, Brian Ripley

#
Hi,

After installing the latest versions of lme4, Matrix, VR on a Debian 
Box, I have run into a problem. When I use library(lme4) on R Version 
2.1.0  (2005-04-18) I get the following error message:

 > library(lme4)
Loading required package: Matrix
Error in autoloader(name = "confint", package = "MASS") :
        autoloader did not find 'confint' in 'MASS'

An example of the complete session is given at the end of this email. A 
search in the docs.
At present I have 3 directories where packages are kept, as I use 
apt-get install "r-cran-package_ name" when one is available in Debian 
and when it is not R, R CMD INSTALL package_name. This gives me the 
following directories when use help.start():

(1) Packages in /usr/local/lib/R/site-library  (packages installed by R 
CMD INSTALL)

(2) Packages in /usr/lib/R/site-library (packages installed with apt-get 
install)

(3) Packages in /usr/lib/R/library (base install with apt-get install)


Matrix, lme4 and other suggested packages were first installed with R 
CMD INSTALL.  and can be found in /usr/local/lib/R/site-library.  After 
the error occured, I updated VR with apt-get install (found in 
/usr/lib/R/library) and got the same error message.

Installing a second version of lme4 and matrix with apt-get install puts 
these packages in /usr/lib/R/site-library and gives the same error as 
before.

Looking at the MASS no function 'confint' exists, only confint.glm, 
confint.nls,confint.profile.glm and confint.profile.nls

Can anyone suggest a solution to the problem or why I am unable to 
access lme4 or as the error statement suggests, a problem with the 
Matrix package and Mass?



Thanks

Peter



#############################################
R : Copyright 2005, The R Foundation for Statistical Computing
Version 2.1.0  (2005-04-18), ISBN 3-900051-07-0

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

  Natural language support but running in an English locale

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for a HTML browser interface to help.
Type 'q()' to quit R.

Loading Tcl/Tk interface ... done
Loading required package: tcltk
Loading required package: rgl
Loading required package: zoo
Loading required package: strucchange
Loading required package: sandwich
Loading required package: relimp
Loading required package: nnet
Loading required package: graphics
Loading required package: grDevices
Loading required package: stats
Loading required package: nlme

Attaching package: 'nlme'


        The following object(s) are masked from package:stats :

         contr.SAS

Loading required package: mvtnorm
Loading required package: multcomp
Loading required package: mgcv
This is mgcv 1.2-4
Loading required package: MASS
Loading required package: lmtest
Loading required package: lattice
Loading required package: grid
Loading required package: foreign
Loading required package: effects
Loading required package: car
Loading required package: abind
[Previously saved workspace restored]

 > library(lme4)
Loading required package: Matrix
Error in autoloader(name = "confint", package = "MASS") :
        autoloader did not find 'confint' in 'MASS'
 >
#
The problem is **NOT** in MASS.  confint is in stats, not MASS, and has 
been (in base then stats) since R 1.7.0.  You could (and should) have 
checked that from the (O)NEWS files of R and MASS.

Nor is there an error in autoloader, as your title stated.

Some other piece of software on your system is seriously out of date.
Running traceback() might tell you what it is.  My suspicion is that
you have somewhere a long-outdated version of lme4.

Starting each R session with that long list of packages is surely not a 
good idea.  Part of the homework before posting should be to run R 
--vanilla and see what happens there.

If this were me I would scrap all of the outdated R installation you have 
and start again with R 2.1.1 (or R-patched).
On Thu, 11 Aug 2005, Peter Ho wrote: