Sorry to be such a nuisance but after asking for help with compiling R-1.1.0 on an Intel-based machine I now have another problem on the Alpha DS-10 server which my students use. Compilation and instalation proceed perfectly, but when I invoke R I get: [etdx01:bin]$ ./R Fatal error: The X11 shared library could not be loaded. The error was /home/etptupaf/R/lib/R/bin/R_X11.so: undefined symbol: R_GlobalEnv I have read the INSTALL, README, and FAQ files and grep'ed the docs for info on R_Global_Env, but found none. Aside: I am using gcc as I have ever done, because I first want a functional executable. However, I have read (in www.alphalinux.org?) that substantial improvements on speed of heavily floating point code can be achieved on Alphas by using the Compaq C compiler (available at their site). If somebody has some experience with that, I would like to hear from him/her. ft. -- Fernando TUSELL e-mail: Departamento de EconometrÃa y EstadÃstica etptupaf at bs.ehu.es Facultad de CC.EE. y Empresariales Tel: (+34)94.601.3733 Avenida Lendakari Aguirre, 83 Fax: (+34)94.601.3754 E-48015 BILBAO (Spain) Secr: (+34)94.601.3740 PGP: finger etptupaf at bsdx01.bs.ehu.es http://agesi.bs.ehu.es ---------------------------------------------------------------------- -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
> Date: Tue, 20 Jun 2000 10:40:32 +0100 (BST) > From: Prof Brian D Ripley <ripley at stats.ox.ac.uk> > Sender: ripley at auk.stats > Cc: r-help at stat.math.ethz.ch > Content-Type: TEXT/PLAIN; charset=US-ASCII > > On Tue, 20 Jun 2000, F.Tusell wrote: > > > Sorry to be such a nuisance but after asking for help with compiling > > R-1.1.0 on an Intel-based machine I now have another problem on the > > Alpha DS-10 server which my students use. > > Um. Is this running Linux then? If so, what distribution? It might > matter. >Yes, Red Hat 6.1.> > Compilation and instalation proceed perfectly, but when I invoke R I > > get: > > > > [etdx01:bin]$ ./R > > Fatal error: The X11 shared library could not be loaded. > > The error was /home/etptupaf/R/lib/R/bin/R_X11.so: undefined symbol: R_GlobalEnv > > > > I have read the INSTALL, README, and FAQ files and grep'ed the > > docs for info on R_Global_Env, but found none. > > R_Global_Env is an exported constant from R.bin. I get (Solaris, but i686 > Linux is similar) > > auk% nm -g R_X11.so > R_X11.so: > 0000000000 U .div > 0000000000 U .rem > 0000000000 U .udiv > 0000000000 U .umul > 0000000000 U C_alloc > 0000000000 U C_free > 0000039416 T GetX11Pixel > 0000014944 T RX11_dataentry > 0000000000 U R_ExpandFileName > 0000000000 U R_GlobalEnv > ... > > auk% nm -g R.bin > R.bin: > ... > 0002008832 B R_GlobalEnv > > So what should happen is that R_GlobalEnv is exported from R.bin > (specifically, from src/main/main.o, I think) and when R_X11.so is > dynamically loaded, the symbol is resolved. > > Martyn Plummer has pointed out that on some Linux systems (RH6.2 > specifically) the `undefined symbol' messages are unreliable. > So you might want to bear that in mind. >I wouldn't mind about the message only, but execution aborts (or perhaps does not even start).> BTW, where does make check get to? (The same error I presume.) > You could also try > > R --gui=none > > and then try loading some libraries. That would indicate if there > was a general problem with dlopen or something specific to R_X11.so. >I will do that and let you know if I find something. In the mean time, if any of the developers can think of some relevant change in the libraries involved, that might be a clue. What puzzles me is that after the failure with R-1.1.0, I backtracked and installed again R-1.0.1 with absolutely no problem (and there is also a R_X11.so there). Thank you for your help. ft.> -- > Brian D. Ripley, ripley at stats.ox.ac.uk > Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ > University of Oxford, Tel: +44 1865 272861 (self) > 1 South Parks Road, +44 1865 272860 (secr) > Oxford OX1 3TG, UK Fax: +44 1865 272595 >-- Fernando TUSELL e-mail: Departamento de Econometr?a y Estad?stica etptupaf at bs.ehu.es Facultad de CC.EE. y Empresariales Tel: (+34)94.601.3733 Avenida Lendakari Aguirre, 83 Fax: (+34)94.601.3754 E-48015 BILBAO (Spain) Secr: (+34)94.601.3740 PGP: finger etptupaf at bsdx01.bs.ehu.es http://agesi.bs.ehu.es ---------------------------------------------------------------------- -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Invoking R with --gui=none (as suggested by Prof. Ripley) does at least
start execution. But i never get to the prompt ">", and after a
while
R crashes. Pasted below is what I get.
I don't think R developers should spend time on that. Dumping core is
not unprecedented with RH6.1. On my system it happens with gnorpm, which
is part of the distribution, I believe. I only report this so R developers
have all the facts. For the time being, I am entirely happy vith R-1.0.1
which works wonderfully for me.
ft.
----------------------------------------------------------
[etdx01:bin]$ ./R --gui=none
R : Copyright 2000, The R Development Core Team
Version 1.1.0 (June 15, 2000)
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.
R is a collaborative project with many contributors.
Type "?contributors" for a list.
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.
Floating point exception (core dumped)
[etdx01:bin]$
------------------------------------------------------------
--
Fernando TUSELL e-mail:
Departamento de Econometr?a y Estad?stica etptupaf at bs.ehu.es
Facultad de CC.EE. y Empresariales Tel: (+34)94.601.3733
Avenida Lendakari Aguirre, 83 Fax: (+34)94.601.3754
E-48015 BILBAO (Spain) Secr: (+34)94.601.3740
PGP: finger etptupaf at bsdx01.bs.ehu.es http://agesi.bs.ehu.es
----------------------------------------------------------------------
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Hello, After you do ./configure, check if it is properly configured. In my case, I was getting the problem you mentioned because the configure script did not recognize my machine (Alpha 21164PC). This problem is already reported to r-devel. I guess DS-10 is 21264?? Then, it should be recognized as ev6 or ev67? Check the very begining of the output of ./configure and it should say something like: checking host system type... alphapca56-unknown-linux-gnu If it says "failed to recognized ev67", try ./configure --host=alpha. I assume you are using Linux on the alpha. If your system is rpm based, there is a prebuit rpm of most recent R. ftp://cran.r-project.org/pub/R/bin/linux/redhat/6.x/alpha/ Naoki Naoki Takebayashi <ntakebay at bio.indiana.edu> --- Dept. of Biology, Indiana University, Jordan Hall 142 --- 1001 E. Third St. Bloomington, IN 47405-3700 USA On Tue, 20 Jun 2000, F.Tusell wrote:> Sorry to be such a nuisance but after asking for help with compiling > R-1.1.0 on an Intel-based machine I now have another problem on the > Alpha DS-10 server which my students use. > > Compilation and instalation proceed perfectly, but when I invoke R I > get: > > [etdx01:bin]$ ./R > Fatal error: The X11 shared library could not be loaded. > The error was /home/etptupaf/R/lib/R/bin/R_X11.so: undefined symbol: R_GlobalEnv-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
On Tue, 20 Jun 2000, F.Tusell wrote:> Aside: I am using gcc as I have ever done, because I first want a > functional executable. However, I have read (in www.alphalinux.org?) > that substantial improvements on speed of heavily floating point code > can be achieved on Alphas by using the Compaq C compiler (available at > their site). If somebody has some experience with that, I would like to > hear from him/her.This would be a worthwhile experiment. If the Compaq compiler is superior only for floating point there would probably not be much difference, though. R is not 'heavily floating point code' in most cases, as is shown by the lack of improvement from using optimised BLAS code and the relative performance on x86 and SPARC architectures. -thomas Thomas Lumley Assistant Professor, Biostatistics University of Washington, Seattle -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
> Aside: I am using gcc as I have ever done, because I first want a > functional executable. However, I have read (in www.alphalinux.org?) > that substantial improvements on speed of heavily floating point code > can be achieved on Alphas by using the Compaq C compiler (available at > their site). If somebody has some experience with that, I would like to > hear from him/her. >I've had trouble using R compiled with the C compiler that comes with Compaq's Tru64 OS (not sure how this compares to the Linux version you mention above). It seems that during R's garbage collection some variables that are in use get set to NULL. Don't know why -- maybe a compiler bug? I'm using gcc now. Too bad since I have some C code that runs about twice as fast with the DEC/Compaq compiler than with gcc. Although as Thomas has said a similar improvement might not be expected with R. Brad -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._