[R] R-release dynamic load problem on HPUX10.20
partha_bagchi at hgsi.com
Wed Mar 22 20:51:47 CET 2000
Looks like this was the clue along with Douglas Bates comment. Indeed, I
had an older version of ctest that was overwriting the base version of
ctest. Hence the dyn.load problem. I can load ctest just fine now. I
re-downloaded chron and reinstalled R, along with all the packages. Now all
the packages load just fine.
Thanks for all your help as well as Prof. Douglas Bates.
I am finally upgrading to R version 1.0. The problems mentioned below were
show-stoppers for me. I intend to visit CRAN once a month to make sure that
I have updated packages.
<guido at sirio.stat To: Douglas Bates <bates at stat.wisc.edu>
.unipd.it> cc: partha_bagchi at hgsi.com, osman.buyukisik at ae.ge.com,
r-help at stat.math.ethz.ch
03/22/00 02:06 PM Subject: Re: [R] R-release dynamic load problem on
On Wed, Mar 22, 2000 at 11:46:53AM -0600, Douglas Bates wrote:
> partha_bagchi at hgsi.com writes:
> > I have a similar problem with R 1.0.0 (Windows 98) in loading 3
> > ctest, chron, lme. Every other package loads fine. I can live without
> > chron, but ctest and lme are very important to me. Reinstallation of R
> > doesn't help. I can't figure out what the problem. Has anybody else had
> > this problem in Windows? Any suggestions appreciated. TIA.
> Please do not try to install the lme package. That package has been
> superseded by the nlme package, for which a .zip file is available in
> the bin/windows/windows-9x/contrib/ area on CRAN.
just a quick question (due to Douglas comment): are you using the
updated versions of the packages at CRAN (BTW, ctest is now in
the base distribution) ?
Reasons of the question: (i) lme is not in the 'pre-compiled for Windows'
contrib section at CRAN (ii) I can dyn.load the ctest in the base
distribution also under Win98 (indeed, also make pkgcheck-ctest works
for me under Win98) (iii) every change in the R API can result
under Windows in the impossibility to load a particular dll (it happens
if the dll needs a symbol in R.dll now exported with another name);
for this reason, at least under Windows, it has been necessary to upgrade
not only the base distribution but also the packages. Perhaps, now
this would be less mandatory since the plans are to keep the API
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
More information about the R-help