======================================================================Simple
CRAN packages which do not compile without modifications (all
others do)
====================================================================== --
akima
/usr/bin/ld: multiple definitions of symbol _idlc_
-- fracdiff
/usr/bin/ld: multiple definitions of symbol _gammfd_
(and others)
-- odesolve
-- princurve
lamix:
Error on line 116: Declaration error for v: adjustable dimension on
non-argument
-- splancs
/usr/bin/ld: multiple definitions of symbol _bounds_
-- wle
mleaic:
Error on line 229: Declaration error for work: adjustable dimension on
non-argument
-- subselect
configure: error: R was configured without a FORTRAN compiler
ERROR: configuration failed for package `subselect'
The multiple defined symbols are a well known OS X problem. There are ad
hoc fixes
using linker flags, but the best way to fix this is to use extern in the
source code (although
almost all these packages use f2c to translate the fortran into c). The
"adjustable dimension"
error is possibly an f2c problem. Ans subselect does not recognize f2c
as a fortran
compiler.
======================================================================Complicated/interface
packages which compile (almost) out-of-the-box
======================================================================RPgSQL
RmSQL
RMySQL
gtk
RGtk (although we need to add a typedef for ulong to gtkAccessors.c)
RSPerl
XML (this gives a multiple defined symbol for xmlParserError, which is
defined in Utils.c but also in libxml.dylib -- I commented out the
routine
in Utils.c, but this may be a bit heavy handed).
Rpy
rpvm
RObjectTables
======================================================================CRAN
packages I have not tried yet (because I don't have/use the things
they interface to)
======================================================================RODBC
ROracle
RQuantlib
RSQLite
======================================================================CRAN
packages which compile and install, but do not work
> library("Matrix")
Error in dyn.load(x, as.logical(local), as.logical(now)) :
unable to load shared library
"/sw/lib/R/library/Matrix/libs/Matrix.so":
dlcompat: dyld: /sw/lib/R/bin/R.bin Undefined symbols:
_ATL_cGetNB
_ATL_dGetNB
_ATL_sGetNB
_ATL_zGetNB
_cblas_dgemm
_cblas_dscal
_cblas_dswap
_cblas_dsyrk
_cblas_dtrsm
_cblas_idamax
_dsymv_
Error in library("Matrix") : .First.lib failed
If I use nm on Matrix.so, all these symbols are undefined.
==Jan de Leeuw; Professor and Chair, UCLA Department of Statistics;
US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554
phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu
homepage: http://www.stat.ucla.edu/~deleeuw
======================================================= No matter
where you go, there you are. --- Buckaroo Banzai
http://www.stat.ucla.edu/~deleeuw/sounds/nomatter.au
=======================================================
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel 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-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
I compiled and installed fracdiff, akima, splancs and odesolve by adding the -m linker flag when foo.so is build. I compiled and installed princurve and wle by removing nasty end-of-file characters from some of the fortran sources There are no solutions for the subselect and Matrix problems yet. On Saturday, April 6, 2002, at 11:28 AM, Jan de Leeuw wrote:> ======================================================================> Simple CRAN packages which do not compile without modifications (all > others do) > ======================================================================> -- akima > /usr/bin/ld: multiple definitions of symbol _idlc_ > > -- fracdiff > /usr/bin/ld: multiple definitions of symbol _gammfd_ > (and others) > > -- odesolve > > -- princurve > lamix: > Error on line 116: Declaration error for v: adjustable dimension on > non-argument > > -- splancs > /usr/bin/ld: multiple definitions of symbol _bounds_ > > -- wle > mleaic: > Error on line 229: Declaration error for work: adjustable dimension on > non-argument > > -- subselect > configure: error: R was configured without a FORTRAN compiler > ERROR: configuration failed for package `subselect' > > The multiple defined symbols are a well known OS X problem. There are > ad hoc fixes > using linker flags, but the best way to fix this is to use extern in > the source code (although > almost all these packages use f2c to translate the fortran into c). The > "adjustable dimension" > error is possibly an f2c problem. Ans subselect does not recognize f2c > as a fortran > compiler. > ======================================================================> Complicated/interface packages which compile (almost) out-of-the-box > ======================================================================> RPgSQL > RmSQL > RMySQL > gtk > RGtk (although we need to add a typedef for ulong to gtkAccessors.c) > RSPerl > XML (this gives a multiple defined symbol for xmlParserError, which is > defined in Utils.c but also in libxml.dylib -- I commented out the > routine > in Utils.c, but this may be a bit heavy handed). > Rpy > rpvm > RObjectTables > ======================================================================> CRAN packages I have not tried yet (because I don't have/use the things > they interface to) > ======================================================================> RODBC > ROracle > RQuantlib > RSQLite > ======================================================================> CRAN packages which compile and install, but do not work > > > library("Matrix") > Error in dyn.load(x, as.logical(local), as.logical(now)) : > unable to load shared library > "/sw/lib/R/library/Matrix/libs/Matrix.so": > dlcompat: dyld: /sw/lib/R/bin/R.bin Undefined symbols: > _ATL_cGetNB > _ATL_dGetNB > _ATL_sGetNB > _ATL_zGetNB > _cblas_dgemm > _cblas_dscal > _cblas_dswap > _cblas_dsyrk > _cblas_dtrsm > _cblas_idamax > _dsymv_ > Error in library("Matrix") : .First.lib failed > > If I use nm on Matrix.so, all these symbols are undefined. >==Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu homepage: http://www.stat.ucla.edu/~deleeuw ======================================================= No matter where you go, there you are. --- Buckaroo Banzai http://www.stat.ucla.edu/~deleeuw/sounds/nomatter.au =======================================================>> -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- > .-.-.- > r-devel 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-devel- > request@stat.math.ethz.ch > _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._ > ._._._ >-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel 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-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
>>>>> Jan de Leeuw writes:> ======================================================================> Simple CRAN packages which do not compile without modifications (all > others do) > ======================================================================> -- akima > /usr/bin/ld: multiple definitions of symbol _idlc_> -- fracdiff > /usr/bin/ld: multiple definitions of symbol _gammfd_ > (and others)> -- odesolve> -- princurve > lamix: > Error on line 116: Declaration error for v: adjustable dimension on > non-argument> -- splancs > /usr/bin/ld: multiple definitions of symbol _bounds_> -- wle > mleaic: > Error on line 229: Declaration error for work: adjustable dimension on > non-argument> -- subselect > configure: error: R was configured without a FORTRAN compiler > ERROR: configuration failed for package `subselect'> The multiple defined symbols are a well known OS X problem. There are > ad hoc fixes using linker flags, but the best way to fix this is to > use extern in the source code (although almost all these packages use > f2c to translate the fortran into c). The "adjustable dimension" error > is possibly an f2c problem. Ans subselect does not recognize f2c as a > fortran compiler.I recently added a shell script wrapper to compile/link FORTRAN 77 code based on f2c, which for the purpose of Autoconf can be used as the FORTRAN 77 compiler. The configure script in subselect will need to be modified to use R CMD config F77 (rather than grep/sed on etc/Makeconf) for getting this. But this will only make the configuration error go away: the package will still not compile because the code, as wle and princurve, use variable-sized arrays.> CRAN packages which compile and install, but do not work>> library("Matrix") > Error in dyn.load(x, as.logical(local), as.logical(now)) : > unable to load shared library > "/sw/lib/R/library/Matrix/libs/Matrix.so": > dlcompat: dyld: /sw/lib/R/bin/R.bin Undefined symbols: > _ATL_cGetNB > _ATL_dGetNB > _ATL_sGetNB > _ATL_zGetNB > _cblas_dgemm > _cblas_dscal > _cblas_dswap > _cblas_dsyrk > _cblas_dtrsm > _cblas_idamax > _dsymv_ > Error in library("Matrix") : .First.lib failed> If I use nm on Matrix.so, all these symbols are undefined.[I assume there are in your Atlas and cblas libs? What does ldd on Matrix.so say?] -k -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel 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-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._