[Rd] MAINLD and SHLIBLD

Albrecht Gebhardt albrecht.gebhardt@uni-klu.ac.at
Tue, 27 Jun 2000 22:10:00 +0200 (MET DST)


On Fri, 23 Jun 2000, Kurt Hornik wrote:

> I would like to always use ${CC} for linking the main binary and the
> shlibs.
> 
> Inspection of configure.in reveals that we currently deviate in the
> following cases:
> 
> 			mainld				shlibld
> 	aix[2-9]*	ld				ld
> 	alpha*osf*	${FC} -g -nofor_main -fpe3	${FC}

I'm only speaking for alpha*osf*:

I think these settings have their roots in very old R versions, where we
had no means to determine the system dependent Fortran libraries (-lUfor
-lots ... for DEC compilers). Now it works without these settings, i.e. 
with mainld=shlibld=${CC}. 

With DEC compilers it still fails in a "make check" with 
  ../../bin/R --vanilla < base-Ex.R > base-Ex.Rout
  /bin/sh: 19080 Memory fault
but this is unrelated to shlibld and mainld settings, as it occurs the
same way with current settings. 

The combination gcc/g77 (version 2.95.2) works as before (successfull
"make check"), the nofor_main and fpe3 options where meaningless (and even
generated some warnings) for GNU compilers.

Here is what I did:  
agebhard@delta[R-1.2.0]$ diff configure.in.bak configure.in
630,633c630,635
<     mainld="${FC} -g -nofor_main -fpe3 "
<     cpicflags=
<     cxxpicflags=
<     shlibld=${FC}
---
>     if test "${GCC}" != yes; then
>   #    mainld="${FC} -g -nofor_main -fpe3 "
>       cpicflags=
>       cxxpicflags=
>   #    shlibld=${FC}
>     fi

DEC compilers have no PIC flags, thats the reason for setting them to "", 
but may be configure would guess this automatically.

This was with a current r-devel version, on alpha-dec-osf4.0e

Albrecht


> 	hpux*		${CC}				ld
> 
> Does anyone recall why we do not use ${CC} in these cases?  Did anyone
> successfully build R on these platforms with ${CC} for mainld/shlibld?
> (Inspection of ltconfig from libtool seems to indicate that ${CC} should
> be possible ...)
> 
> -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
> _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
> 

......................................................................
| Albrecht Gebhardt          Tel.: (++43 463) 2700/832               |
| Institut fuer Mathematik   Fax : (++43 463) 2700/834               |
| Universitaet Klagenfurt    mailto:albrecht.gebhardt@uni-klu.ac.at  |
| Villacher Str. 161         http://www-stat.uni-klu.ac.at/~agebhard |
| A-9020 Klagenfurt, Austria                                         |
`--------------------------------------------------------------------'



-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._