[Rd] R 2.11.0 on RHEL5 / RODBC

2010-05-12 Thread Chuck White
I am compiling R 2.11.0 on a RHEL5.3 box using the following settings

./configure --with-readline=yes --enable-R-shlib=yes --with-x=yes 
--with-blas=-llibptf77blas -lpthread -llibatlas --prefix=/usr/local/R-2.11.0 
JAVA_HOME=$JAVA_HOME CPPFLAGS=-I/usr/local/unixODBC-2.3.0/include

I have compiled and installed unixODBC-2.3.0 (64-bit) in /usr/local and can 
connect to and query the database using isql.  When I try to build RODBC, I get 
the following error:

checking for sqlext.h... yes
checking for library containing SQLTables... no
configure: error: no ODBC driver manager found
ERROR: configuration failed for package âRODBCâ
* removing â/usr/local/R-2.11.0/lib64/R/library/RODBCâ

The directory /usr/local/unixODBC-2.3.0/lib has the files libodbc.so and 
libodbc.la, and LD_LIBRARY_PATH has 
/usr/local/lib64:/usr/lib64:/usr/local/unixODBC-2.3.0/lib.  Also, there is no 
other unixODBC version on the box. I would appreciate any help in this regard.  
Thanks.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] R 2.11.0 on RHEL5

2010-05-05 Thread Chuck White
Thanks again for your note.  I am now compiling using:

--with-readline=yes --enable-R-shlib=yes --with-x=yes 
--with-blas=-llibptf77blas -lpthread -llibatlas --prefix=/usr/local/R-2.11.0

The above configuration is creating libRlapack.so as well. Do you specify 
--with-lapack as well? If so, what do you set that to? Section A.3.2 states -- 
Please do bear in mind that using --with-lapack is `definitely not 
recommended': it is provided only because it is necessary on some platforms.

Thanks.

 Uwe Ligges lig...@statistik.tu-dortmund.de wrote: 
 
 Why do you want shlib support, do you really need it?
 For the static case, A.3.1.1 ATLAS described the way to go.
 
 I have never before linked dynamically to ATLAS versions. ATL_dgemm is 
 probably defined in one of the other libraries. I have not looked since 
 I currently do not have any ATLAS left on my machines which I was more 
 used to roughly 3 years ago.
 
 Uwe Ligges

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] R 2.11.0 on RHEL5

2010-05-04 Thread Chuck White
Thank you for your response. I did look in the suggested manual under A.3.1.5 
Shared BLAS.

Here are my steps: compiled R using
--with-readline=yes --enable-R-shlib=yes --enable-BLAS-shlib=yes --with-x=yes 
--prefix=/usr/local/R-2.11.0

Compiled LAPACK first and then ATLAS using 
http://www.scipy.org/Installing_SciPy/Linux#head-89e1f6afaa3314d98a22c79b063cceee2cc6313c
   All the .so files have been copied to /usr/lib64

In /usr/local/R-2.11.0/lib64/R/lib, I renamed the .so files for blas and lapack 
and linked as
libRblas.so - /usr/local/lib64/libf77blas.so
libRlapack.so - /usr/local/lib64/liblapack.so

R does startup OK but when I do the following:
require(nlme)
fm1 - nlme(height ~ SSasymp(age, Asym, R0, lrc), data = Loblolly, fixed = Asym 
+ R0 + lrc ~ 1, random = Asym ~ 1, start = c(Asym = 103, R0 = -8.5, lrc = 
-3.3)) 

I get
/usr/local/R-2.11.0/lib64/R/bin/exec/R: symbol lookup error: 
/usr/local/R-2.11.0/lib64/R/lib/libRblas.so: undefined symbol: ATL_dgemm
This runs just fine without using the .so files which were built with R.

Can you please help?  Thanks.

 Uwe Ligges lig...@statistik.tu-dortmund.de wrote: 
 This is explained in detail in the manual R Installation and 
 Administration.
 
 Best,
 Uwe Ligges
 
 
 On 03.05.2010 06:02, Chuck White wrote:
  I have built lapack-3.2 and atlas3.8.3 for a RHEL5.3 VM.  The following 
  files have been created -- libatlas (.a  .so), libcblas (.a  .so), 
  libf77blas (.a  .so), liblapack (.a  .so), libptcblas (.a  .so), 
  libptf77blas (.a  .so) and libstatatlas.a.
 
  When building R 2.11.0, I would like to be able to specify --with-blas and 
  --with-lapack (and not use the ones which come with R). What is the syntax 
  for that? Which of these libraries should I specify?
 
  Thanks.
 
  __
  R-devel@r-project.org mailing list
  https://stat.ethz.ch/mailman/listinfo/r-devel

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


[Rd] R 2.11.0 on RHEL5

2010-05-02 Thread Chuck White
I have built lapack-3.2 and atlas3.8.3 for a RHEL5.3 VM.  The following files 
have been created -- libatlas (.a  .so), libcblas (.a  .so), libf77blas (.a  
.so), liblapack (.a  .so), libptcblas (.a  .so), libptf77blas (.a  .so) and 
libstatatlas.a.

When building R 2.11.0, I would like to be able to specify --with-blas and 
--with-lapack (and not use the ones which come with R). What is the syntax for 
that? Which of these libraries should I specify?

Thanks.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


[Rd] R 2.11.0 on RHEL5

2010-05-02 Thread Chuck White
I have built lapack-3.2 and atlas3.8.3 for a RHEL5.3 VM.  The following files 
have been created -- libatlas (.a  .so), libcblas (.a  .so), libf77blas (.a  
.so), liblapack (.a  .so), libptcblas (.a  .so), libptf77blas (.a  .so) and 
libstatatlas.a.

When building R 2.11.0, I would like to be able to specify --with-blas and 
--with-lapack (and not use the ones which come with R). What is the syntax for 
that? Which of these libraries should I specify?

Thanks.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


[Rd] R292 and tcl/tk

2009-11-02 Thread Chuck White
BACKGROUND
  R2.9.2 built on  power5 aix5.3 using gcc 4.2.4   (also installed: tcl and 
tk 8.5.7)
 Access from a Windows XP machine using Exceed v9.0, aixterm (DISPLAY etc. 
setup)

export LDFLAGS=-L/usr/lib64 -L/opt/freeware/lib64 -L/opt/freeware/64/lib 
-L/usr/X11R6/lib -L/opt/freeware/lib
export CPPFLAGS=-I/opt/freeware/include -I/usr/lpp/X11/include/X11


CONFIGURATION 1 (with tcl/tk)
 ../R-2.9.2/configure --enable-R-shlib --enable-BLAS-shlib 
--with-tcl-config=/opt/freeware/lib64/tclConfig.sh 
--with-tk-config=/opt/freeware/lib64/tkConfig.sh --prefix=/usr/local/R-2.9.2  
config_292.log 21

 update.packages()
--- Please select a CRAN mirror for use in this session ---
Loading Tcl/Tk interface ...
 *** caught segfault ***
address 21188073e, cause 'memory not mapped'
 
Traceback:
 1: .C(tcltk_init, PACKAGE = tcltk)
 2: fun(...)
 3: doTryCatch(return(expr), name, parentenv, handler)
...
19: tryCatch(asNamespace(pkg), hasNoNamespaceError = function(e) NULL)
20: tcltk::tk_select.list
21: menu(m[, 1L], graphics, CRAN mirror)
22: chooseCRANmirror()
23: contrib.url(repos, type)
24: available.packages(contriburl = contriburl, method = method)
25: update.packages()


CONFIGURATION 2 (without tcl/tk)

../R-2.9.2/configure --enable-R-shlib --enable-BLAS-shlib --with-tcltk=no 
--prefix=/usr/local/R-2.9.2  config_292.log 21 ; make  make_292.log 21

 update.packages()
--- Please select a CRAN mirror for use in this session ---
CRAN mirror
 
 1: Argentina (Buenos Aires)2: Australia   
 3: Austria 4: Belarus 
 5: Belgium 6: Brazil (PR)


QUESTION
1.  Are there any known issues with tcl/tk? Should I be doing anything 
different in the configuration?
2. When I try a simple plot command in either configuration   (eg.  plot(sin, 
-pi, pi)),  the plot window shows up immediately on my machine but the plot 
itself takes all of two minutes to draw.  I am not sure if the two issues are 
related.

I would appreciate any help with this.  Thanks.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] unable to compile RODBC (R 2.9.2, AIX)

2009-11-01 Thread Chuck White
I could compile RODBC on AIX 5.3/R-2.9.2. I set LDFLAGS to:

export LDFLAGS=-L/usr/lib64 -L/opt/freeware/lib64 -L/usr/X11R6/lib

Earlier, I had some other paths which had 32-bit libraries as well. Obviously, 
I re-compiled R and all the other libraries using this flag.

Prof. Ripley and Marc, thanks for your response.


 Marc Schwartz marc_schwa...@me.com wrote: 
 On Oct 31, 2009, at 1:13 PM, Chuck White wrote:
 
  Hello -- I am trying to compile (linker error, not compiler error)  
  RODBC for R 2.9.2 on AIX 5.3 using gcc. I get the following error:

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] unable to compile mgcv

2009-10-31 Thread Chuck White
I have been able to compile mgcv 1.5-6 *successfully* on the AIX box with gcc.  
 I replaced all the TRUE/FALSE in magic.c and gdi.c with mgcvTRUE/mgcvFALSE.

Thanks to William Dunlap for his help.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


[Rd] unable to compile RODBC (R 2.9.2, AIX)

2009-10-31 Thread Chuck White
Hello -- I am trying to compile (linker error, not compiler error) RODBC for R 
2.9.2 on AIX 5.3 using gcc. I get the following error:

gcc -maix64 -pthread -std=gnu99 -shared -Wl,-brtl -Wl,-G -Wl,-bexpall 
-Wl,-bnoentry -lc -L/usr/lib -L/usr/X
11R6/lib -L/usr/local/lib -L/opt/freeware/lib -L/opt/freeware/lib64 -o RODBC.so 
RODBC.o -lodbc -lm -L/usr/l
ocal/builddir/lib -lR
ld: 0711-738 ERROR: Input file /usr/lib/libodbc.so:
XCOFF32 object files are not allowed in 64-bit mode.
collect2: ld returned 8 exit status
gmake: *** [RODBC.so] Error 1
ERROR: compilation failed for package 'RODBC'

I would appreciate any help I can get in this regard.  Thanks.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] unable to compile mgcv

2009-10-29 Thread Chuck White
Thanks for your response. In the past I have build R for RHEL and it was very 
simple. I did not have to do anything specific for certain libraries.

I do see the .tgz file (actually it is a link to the original file which has 
version number too). Questions
[1] do I need to unzip the tgz file, make changes, and zip it back again? Or 
would the makefile be able to handle the unzipped directory?
[2] is there a way to just compile that library only?
[3] what do I do to resume the compilation of the rest of the libraries. 
Re-running make would likely try to build the original version of mgcv again.

Thanks -- I would appreciate any document/instructions you can point me to.


 William Dunlap wdun...@tibco.com wrote: 
 
  -Original Message-
  From: Chuck White [mailto:chuckwhi...@charter.net] 
  Sent: Wednesday, October 28, 2009 3:47 PM
  To: r-devel@r-project.org; William Dunlap
  Subject: RE: [Rd] unable to compile mgcv
  
  Thanks for your response.  Can you please point me to the 
  make file where I would make this change. Thanks.
 
 If you untar src/library/Recommended/mgcv.tgz (sp?)
 this file will be in mgcv/src.  I'm not sure of the
 best way to take care of it on AIX.  It may be possible
 to #undef TRUE in include/R.h.  The AIX system includes
 define it in many places, the most prominent of which
 is in types.h.
 
 If you cannot generally #undef TRUE in a global
 R include file then you will have to edit the C or h files
 in mgcv.


__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


[Rd] unable to compile mgcv

2009-10-28 Thread Chuck White
Hello -- I am trying to compile R-2.9.2 on IBM Power5 machine with AIX 5.3.  I 
have posted three previous posts. Over time I have made significant progress 
towards getting a successful build. I am now getting the following error:

==
begin installing recommended package mgcv
* Installing *source* package 'mgcv' ...
** libs
gmake[2]: Entering directory `/tmp/RtmpOtsRnd/R.INSTALL2781446b/mgcv/src'
gcc -maix64 -pthread -std=gnu99 -I/usr/local/builddir/include  
-I/usr/local/include  -O2 -g -mcpu=power5 -c gcv.c -o gcv.o
gcc -maix64 -pthread -std=gnu99 -I/usr/local/builddir/include  
-I/usr/local/include  -O2 -g -mcpu=power5 -c gdi.c -o gdi.o
gdi.c: In function 'qr_ldet_inv':
gdi.c:294: error: expected identifier or '(' before numeric constant
gdi.c:308: error: lvalue required as unary '' operand
gdi.c:308: error: lvalue required as unary '' operand
gdi.c:316: error: 'j' undeclared (first use in this function)
gdi.c:316: error: (Each undeclared identifier is reported only once
gdi.c:316: error: for each function it appears in.)
gdi.c: In function 'get_detS2a':
gdi.c:617: error: expected identifier or '(' before numeric constant
gdi.c:620: error: 'Mf' undeclared (first use in this function)
gdi.c:636: error: 'max_col' undeclared (first use in this function)
.
ERROR: compilation failed for package 'mgcv'
* Removing '/usr/local/builddir/library/mgcv'
gmake[1]: *** [mgcv.ts] Error 1
gmake[1]: Leaving directory `/usr/local/builddir/src/library/Recommended'
gmake: *** [recommended-packages] Error 2
make: The error code from the last command is 2.

Stop.
==

Initially I had mgcv 1.5-5 and I downloaded mgcv 1.5-6 and tried again. I get 
the same error.

I would really appreciate any help with this.

Thanks.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] unable to compile mgcv

2009-10-28 Thread Chuck White
Thanks for your response.  Can you please point me to the make file where I 
would make this change. Thanks.

 William Dunlap wdun...@tibco.com wrote: 
 The first offending line in gdc.c is
 294   int *pivot,i,TRUE=1,j;
 and TRUE must be #defined in some AIX
 include file.  Change it to true or
 add #undef TRUE prior to its use and it
 may compile.
 
 Bill Dunlap
 Spotfire, TIBCO Software
 wdunlap tibco.com

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] R292 on AIX53 using gcc

2009-10-21 Thread Chuck White
Hello -- I haven't received any respones to my previous post on compiling 
R2.9.2 for AIX 5.3. Meanwhile, I have looked at the following post by Ei-ji 
Nakama [http://tolstoy.newcastle.edu.au/R/e4/devel/08/07/2171.html] and it 
appears that all the changes suggested in the patch are already in R2.9.2 
Makefiles.

I have also looked at the documentation under R-Forge/AIX page 
[http://r-forge.r-project.org/projects/aix/]. That does not seem to help either.

Bottomline, I am trying to compile R2.9.2 on a AIX 5.3 Power5 box with gcc 4.3 
and get the following error:

gcc -maix64 -pthread -std=gnu99 -Wl,-brtl -Wl,-bexpall 
-Wl,-bpT:0x1 -Wl,-bpD:0x11000 -lc  -L/usr/local/lib -o R.bin 
Rmain.o -L../../lib -lR -lRblas
ld: 0711-224 WARNING: Duplicate symbol: .memcpy, memcpy, .bcopy, bcopy, 
.memmove, .memset, memset, .fres, fres
ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more information.
ld: 0711-317 ERROR: Undefined symbol: R_running_as_main_program

Command used: ./configure --enable-R-shlib --enable-BLAS-shlib --without-iconv 
; make  make.log 21

I would really appreciate any help. 

Thanks in advance.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] R292 on AIX53 using gcc

2009-10-18 Thread Chuck White
Hello -- after my previous post, I tried to compile R 2.9.2 using xlc.  The 
config.site changes are:
 
==
OBJECT_MODE=64
AR=ar -X64
R_SHELL=/usr/bin/bash
CONFIG_SHELL=/usr/bin/bash
 
CC=xlc_r -q64
CFLAGS=-O -qstrict
CPPFLAGS=-I/usr/include -I/usr/local/include -I/opt/freeware/include 
-I/usr/X11R6/include
CXX=xlC_r -q64
CXXFLAGS=-O -qstrict
LDFLAGS=-L/opt/freeware/lib64 -L/opt/freeware/64/lib -L/opt/freeware/lib 
-L/usr/local/lib -L/usr/lib
 
F77=xlf_r -q64
FFLAGS=-O -qstrict
FC=xlf95 -q64
FCFLAGS=-O -qstrict
 
MAKE=/opt/freeware/bin/make
==
 
Error --
 
mkdir ../../../../library/grDevices/libs
Warning in solve.default(rgb) :
  unable to load shared library '/usr/local/R-2.9.2/modules//lapack.so':
  rtld: 0712-001 Symbol idamax was referenced
  from module /usr/local/R-2.9.2/lib/libRlapack.so(), but a runtime 
definition
 of the symbol was not found.
rtld: 0712-001 Symbol dger was referenced
  from module /usr/local/R-2.9.2/lib/libRlapack.so(), but a runtime 
definition
 of the symbol was not found.
..
Error in solve.default(rgb) : lapack routines cannot be loaded
Error: unable to load R code in package 'grDevices'
Execution halted
make: The error code from the last command is 1.
 
Thanks in advance. I would really appreciate any help.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


[Rd] R292 on AIX53 using gcc

2009-10-17 Thread Chuck White
I apologize for cross posting this message in the R-help group as well. Having 
posted it there a couple of hours ago, I felt this may be a more appropriate 
forum for a question of this type.


Hello -- I am unable to build R 2.9.2 on IBM PowerPC AIX5.3.  I would 
appreciate 
any help in this matter. 
 
===details== 
Machine:  IBM PowerPC_POWER5 / 4 proc, 1499 MHz 64-bit / AIX 5.3.0.0 
 
Building R 2.9.2 using gcc/g++/gfortran 4.2.4 
 
Config.site changes are: 
  OBJECT_MODE=64 
  CONFIG_SHELL=/usr/bin/bash 
  LIBICONV=/opt/freeware/lib 
  AR=ar -X64 
  CC=gcc -maix64 
  F77=gfortran -maix64 
  LDFLAGS=-L/opt/freeware/lib64 -L/opt/freeware/64/lib -L/opt/freeware/lib 
-L/usr/local/lib 
  CXX=g++ -maix64 
  FC=gfortran -maix64 
  R_BROWSER=/opt/freeware/bin/firefox 
  MAKE=/opt/freeware/bin/make 
 
 
Command for configure:  ./configure --enable-R-shlib 
 
Looks like I have to set OBJECT_MODE explicitly on the command line. So, I 
enter: set OBJECT_MODE=64.  It does not seem to help that I have it in 
config.site. 
 
make stops with the following error: 
 
 gcc -maix64 -std=gnu99 -Wl,-brtl -Wl,-bexpall -Wl,-bpT:0x1 
-Wl,-bpD:0x11000 -lc  -L/opt/freeware/lib64 -L/opt/freeware/64/lib 
-L/opt/freeware/lib -L/usr/local/lib -o R.bin Rmain.o -L../../lib -lR 
ld: 0711-224 WARNING: Duplicate symbol: .memcpy 
...duplicate symbol warning for: memcpy,bcopy,memset,fres 
ld: 0711-317 ERROR: Undefined symbol: R_running_as_main_program 
collect2: ld returned 8 exit status 
make: The error code from the last command is 1.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel