R-beta: scripting with R? (and also some news on RSXNT)

[guido@sirio.stat.unipd.it]

Peter Dalgaard BSA writes:
> 
> zk@biocem.univ-bpclermont.fr (Zivan Karaman) writes:
> 
> > Does it work (and how) under Windows? (I'm using the version compiled by
> > Guido Masarotto on an NT4 machine).
> 
> Nope. With Win32 (95/NT) you have to decide whether to write a
> "console application" or a "windowed" one. Guido's is of the latter
> variety and you need the former in order to do redirections etc. 
   
   A simple solution would be to introduce a command-line option, e.g.,
   R -save/nosave -source FileName1 FileName2 .....
   with a obvious meaning, or maybe
   R -save/nosave FileName FileName1 FileName2 .....
   This seems easy to do and can be a interesting option also under
   Unix since, it is only my opinion, it provides a cleaner interface.
   Perhaps, under MSWindows or other "windowed" version, we should
   provide also a -iconic option.
   
   
> It is in fact quite easy to do a console R with the mingw32 tools and
> (probably) also with RSXNT. Start with the Unix sources and throw out
> anything that refuses to compile, such as the X11 stuff and things
> related to system() (the latter might be fakeable, but won't work out
> of the box). 

  I can confirm that it easy. This was my first attempt and in half an
  hour I have a console R working (no dynamic linking, but I 
  suppose that you can include the dynload.c from Windows).
  Anyway, I didn't keep the modified source.
  The problems with the system() stuff under MsWindows is not the
  C system() function that works but the R shell scripts (which
  of course require a Unix shell). 
  I don't know if it is already in the TODO list, but a low
  priority suggestion is to insert somewhere (CFLAGS, Platform.h
  I don't know where) a NOX11 flag.
  
  BTW, this morning, I had some time and I tried to recompile R
  (the last "stable" version, i.e., 0.61.3) with RSXNT.
  It seems to work (but I just tryed a couple of demos
  under Win95). 
  I just used my mingw32 port and the mingw32 winapi header
  and didn't try to use g77 (Ok, I tryed but it didn't work,
  I don't know why).
  I hope to put something available the next week.
  In particular, I want to try to use R.exe as a dll so that
  dynamic libraries can use R_alloc, callS ....
  I was not able to do this with egcs-mingw32 (I tryed all
  the rules that I found to build a relocable exe with the
  gnu binutils but  no one worked; suggestion?) but
  I made some tests with RSXNT (not with R, however) with
  good results. A minor problem is that the standard C stack
  is not large enough to execute all the demos. I can
  increase it but not dinamically. How can I determine a reasonable
  value?
  
  guido masarotto

  

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