[R-sig-Debian] libRlapack.so ... : No such file or directory

[Dirk Eddelbuettel]

On Fri, May 30, 2008 at 12:51:18PM -0300, tyler wrote:
> On Fri, May 30, 2008 at 02:56:37PM +0000, Dirk Eddelbuettel wrote:
> > On Fri, May 30, 2008 at 11:44:34AM -0300, tyler wrote:
> > > I tried to install a few packages and got hung up when Matrix couldn't
> > > be lazy loaded. The actual error follows:
> > > 
> > > > library(Matrix)
> > > Loading required package: lattice
> > > Error in dyn.load(file, DLLpath = DLLpath, ...) : 
> > >   unable to load shared library 
> > >          '/usr/local/lib/R/site-library/Matrix/libs/Matrix.so':
> > >   libRlapack.so: cannot open shared object file: No such file or directory
> > > Error: package/namespace load failed for 'Matrix'
> > > > 
> > > 
> > > I googled a bit, and found some related bug reports, but I can't
> > > understand them. Any suggestions? Could this be related to switching to
> > > the optimized atlas libraries, which I did earlier in the week?
> > 
> > Let me guess: you mixed and matched installations from Debian and
> > directly from CRAN?  Everything from Debian should be fine.
> > 
> 
> Well, yes, but I had it on good authority (yours!) that this was ok ;)

Things change over time. What may have been ok when I said may no
longer be true. That;s where the 'transitory' part in transition comes
from.

> r-cran-matrix is not installed on my system, but r-cran-lattice is.
> Maybe lattice is looking for the .deb location of Matrix, and I've got
> matrix in the cran location? 

Huh? 
 
> The great majority of packages on my system are pulled in as 'depends'
> either from the r-cran-* debs or from CRAN, so I'm not entirely
> certain what gets installed from the two different repositories. In
> general I install with install.packages(), but when that fails due to
> external depends (i.e., non-R graphics libraries or whatever) I use
> aptitude to sort it out.
> 
> So, in short, I'm pretty sure I only have one Matrix library, but
> maybe it's in the wrong place?

I asked you to verify, and you come back saying 'I dunno but feel good'.
That doesn't really help, does it?

So:  Run 'library()' in R and see how often you get Matrix. You can assign the value from calling library() and subset or you can just eyeball.

Or;  From a shell prompt, run 'locate Matrix/DESCRIPTION'

Dirk

-- 
Three out of two people have difficulties with fractions.