[R-sig-ME] Another case of -1.0 correlation of random effects

[Douglas Bates]

On Tue, Apr 13, 2010 at 10:19 AM, Kevin E. Thorpe
<kevin.thorpe at utoronto.ca> wrote:
> Ben Bolker wrote:
>>
>>  In that case, just go to http://r-forge.r-project.org/scm/?group_id=60
>> and follow the instructions:
>>
>> svn checkout svn://svn.r-forge.r-project.org/svnroot/lme4
>>
>> then navigate to the lme4a directory and build/install it ...
>
> It's not my day.  I go to the directory where the lme4a tree is located and
> type "R CMD build lme4a" and get an error when the process gets to:
>
> g++ -I/usr/local/lib/R/include  -I/usr/local/include
> -I"/usr/local/lib/R/library/Matrix/include"
> -I"/usr/local/lib/R/library/Rcpp/include"   -fpic  -g -O2 -c glmFamily.cpp
> -o glmFamily.o
>
> In file included from glmFamily.cpp:1:
> glmFamily.h:5:18: error: Rcpp.h: No such file or directory
> In file included from glmFamily.cpp:1:
> [errors continue]
>
> I looked at lme4a/src/glmFamily.h and find
>
> #include <Rcpp.h>
> #include <Rmath.h>
>
> Now, I have both of these files in my installation, but not in the locations
> referenced with the -I switches.  My Rcpp.h is found in
> /usr/local/lib/R/library/Rcpp/lib (not include) and my Rmath.h is
> found in /usr/local/lib/R/include which doesn't even show up in an -I
> switch.
>
> So, I don't know how these -I switches get generated, but there is a problem
> somewhere, either with my installation or something else.
>
> My installation was a standard build (configure; make; make install).
>
> Yes, I fully understand what "unstable" means.  Hopefully these errors are
> helpful.

The DESCRIPTION file says that the lme4a package is linking to Matrix
and Rcpp so you need to have both of those packages installed before
you can compile lme4a.