Skip to content

Porting R

10 messages · Peter Dalgaard, stefano iacus, G. Sawitzki +1 more

Original
G. Sawitzki
# 
I am trying to get R 1.0 running on the Mac.

The main target is MacX. Anyone else working on that?

For the recent Macintosh system, I am trying to compile R using MachTen.
The R core compiles and runs without any changes, using the Unix make files.

The libraries give some problems, presumably due to handling of shared
libraries. It seems I am missing something to link. Before I spend my time
hunting it down, maybe someone has an immediate idea.
  gs.

 ==
my output:

make[4]: Entering directory `//local.src/R-1.0.0/src/library/ctest/src'
gcc -shared -Xlsharedlibrary -Xlexpall  -o ctest.so alnorm.o ansari.o
fexact.o kendall.o
 ks.o prho.o swilk.o  -L/usr/lib/gcc-lib/powerpc-apple-machten4/2.8.1 -lg2c -lm
ansari.o: In function `errmsg':
ansari.c:13: undefined reference to `Rf_error'
ansari.o: In function `w_init':
ansari.c:21: undefined reference to `R_chk_calloc'
...
ansari.c:84: undefined reference to `Rf_choose'
ansari.o: In function `pansari':
ansari.c:110: undefined reference to `Rf_choose'
fexact.o: In function `fexact':
...
kendall.c:24: undefined reference to `Rf_sign'
kendall.o: In function `ckendall':
kendall.c:99: undefined reference to `Rf_gammafn'
swilk.o: In function `swilk':
swilk.c:92: undefined reference to `Rf_qnorm5'
swilk.c:166: undefined reference to `Rf_sign'
...


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Peter Dalgaard
# 
"G. Sawitzki" <gs@statlab.uni-heidelberg.de> writes:
.......
......
Hi Gunther,

I don't know whether it helps, but those unresolved references are
supposed to be there (they point to items in the main R binary) and be
resolved at dyn.load time. That is usually allowed when using the
-shared switch to gcc, although some systems (notably AIX and Win32)
require you use an explicit import library generated from the R
binary.

	-p
stefano iacus
# 
Dear Sawiztki,

I'm also working on the porting of R 1.0 to MacOS. I'm in contact Ross
Ihaka and he thinks (and I agree) that we should use a cross platform
interface like  wxWindows (http://wxwindows.org). 

Let us know what you think about it.

Stefano
-lg2c -lm
.-.-
._._
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
G. Sawitzki
# 
.. so there may be already a prepared rule and config flag somewhere to
create this library? Which flag/option do I need?

 g.


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Peter Dalgaard
# 
"G. Sawitzki" <gs@statlab.uni-heidelberg.de> writes:
Yes, that is the question, but you need to either find someone who
knows machTen, or go through the man pages. Hmm, the following post
from comp.unix.machten doesn't bode too well (Mac-style dynload is
also discussed in
http://www.stat.umn.edu/~luke/xls/projects/dlbasics/dlbasics.html )

-----
 > in article 3852EB96.7A957BAA@uswest.net, Rob & Nancy Wade at rnwade@uswest.net
 wrote on 12/11/99 6:26 PM:
  
 > Anyone had any luck getting Tk installed on Machten?  I'm working with
 > Perl 5.005 and I can't compile with dynamic loading.
  
 Dynamic loading (in the usual Unix sense) doesn't work under MachTen.
 The closest to this is to use the shared libraries Mac-side; see pp.
 219-220 in the manual.
G. Sawitzki
#
At 16:49 +0200 29.03.2000, Peter Dalgaard BSA wrote:
#
Oh. I think I did both.

Hmm, the following post
The relevant lines seem to be:
The Mac does not seem to allow a shared library to have unresolved symbols,
but you can handle references into an executable by linking the library
against the
executable. You also need to take some action to export external symbols;
fortunately this is easy. I have not yet explored what happens when files
are moved
around, search pat issues, and the like ****.

As a work around, I included the standard dynamic library in the core.
Seems to work.

Anyway there should not go too much effort into this. For now, MachTen is
only an intermediate step to go to MachX, which is the really interesting
thing.
 g.


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Robert Gentleman
# 
Hi Guenther,
  Are you going to be home in Sept? Can I still come and visit?
 I finally have made some decision about where I am going (and withdrew
 from the crazy thing at ETH). I'm off to Harvard some time in the fall,

  regards to you and Heidi,
    robert
Robert Gentleman
# 
Well, boy do I feel stupid; I was sure that I had set that to 
mail only to Guenther.
Please disregard my impertinent comments regarding ETH, I apologize
especially to Martin for making them.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
 I will not put both feet in my mouth.
 I will not trust *&%^$& mailers.
Peter Dalgaard
# 
Robert Gentleman <rgentlem@stat.auckland.ac.nz> writes:
<grin> Good thing you didn't put `!!!' after that or Martin would have
had to approve it manually....
Peter Dalgaard
# 
Peter Dalgaard BSA <p.dalgaard@biostat.ku.dk> writes:

[Well, never mind]

Aargh! Another one bites the dust... Please disregard note intended
for Rob and intended to be humourous. Failed on both counts,
obviously. Talking about putting your feet, hands, arms, and legs in
your mouth...