Potential recompilation of certain packages - R-devel

Mon, Jun 17, 2002 9:24 AM #

This applies to people working with the up-to-the-minute source taken
directly or indirectly from the CVS repository. Changes to the
mechanism for registering native (C and Fortran) routines in
DLLs/shared libraries and the package level data structures mean that
you will have to recompile any package that uses this approach to
exporting routines. These include all of the MASS packages, but few
others (yet :-)). No changes to those packages are necessary, just a
simple re-installation.

The changes allow one to specify types for the parameter of the
routines, and also whether they contain useful output values when
returned to R. These have the potential to avoid unecessary copying
and are more robust to errors in types passed to the routines.

Sorry for any inconvenience.

 Duncan

_______________________________________________________________

Duncan Temple Lang                duncan@research.bell-labs.com
Bell Labs, Lucent Technologies    office: (908)582-3217
700 Mountain Avenue, Room 2C-259  fax:    (908)582-3340
Murray Hill, NJ  07974-2070       
         http://cm.bell-labs.com/stat/duncan
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

A.J. Rossini

Mon, Jun 17, 2002 9:29 AM #

And the obvious question, given the timing, 
"does this apply to 1.5.1?"

best,
-tony

duncan> This applies to people working with the up-to-the-minute
    duncan> source taken directly or indirectly from the CVS
    duncan> repository. Changes to the mechanism for registering
    duncan> native (C and Fortran) routines in DLLs/shared libraries
    duncan> and the package level data structures mean that you will
    duncan> have to recompile any package that uses this approach to
    duncan> exporting routines. These include all of the MASS
    duncan> packages, but few others (yet :-)). No changes to those
    duncan> packages are necessary, just a simple re-installation.

    duncan> The changes allow one to specify types for the parameter
    duncan> of the routines, and also whether they contain useful
    duncan> output values when returned to R. These have the potential
    duncan> to avoid unecessary copying and are more robust to errors
    duncan> in types passed to the routines.

A.J. Rossini				Rsrch. Asst. Prof. of Biostatistics
U. of Washington Biostatistics		rossini@u.washington.edu	
FHCRC/SCHARP/HIV Vaccine Trials Net	rossini@scharp.org
-------------- http://software.biostat.washington.edu/ ----------------
FHCRC: M: 206-667-7025 (fax=4812)|Voicemail is pretty sketchy/use Email
UW:   Th: 206-543-1044 (fax=3286)|Change last 4 digits of phone to FAX
(my tuesday/wednesday/friday locations are completely unpredictable.)



-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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

Mon, Jun 17, 2002 9:36 AM #

rossini@blindglobe.net (A.J. Rossini) writes:

No. Bleeding edge, i.e., what might become 1.6.0, only. (Or DTL would
have made himself *very* unpopular.)

O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk)             FAX: (+45) 35327907
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Duncan Temple Lang

Mon, Jun 17, 2002 9:36 AM #

A.J. Rossini wrote:

Good question. No, it should not have anything to do with that.  And,
my guess is that you will notice it immediately on calling library()
on any of the affected packages with a large seg-fault. At least its
dramatic :-)

 D.

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