[Rd] R 1.4.0 build fails on AIX (PR#1289)

Thomas Vogels Thomas Vogels <tov@ece.cmu.edu>
04 Feb 2002 00:00:09 -0500

Thomas Lumley <tlumley@u.washington.edu> writes:

> On Wed, 30 Jan 2002, Dan Million wrote:
> > Thanks for the reply.  I tried building without the modules package, and
> > got past that point.  Now the build fails trying to link tcltk.so.  Seems
> > the tclConfig.sh on my host refers to a file that does not exist
> > (/opt/freeware/src/packages/BUILD/tcltk-8.3.3/tcl8.3.3/unix/lib.exp).
> >
> > I'll freely admit to being clueless about most of this.  I don't know
> > anything about R; I'm just a sysadmin trying to build/install it for a
> > user who needs it.
> Our autodetection of tcl/tk is less reliable than most stuff, partly
> because the tclConfig.sh and similar files are sometimes inaccurate.  If
> your users don't need tcl/tk (which is a fairly low-level set of
> facilities for writing GUIs) you can again remove tcltk from the R_PKGS
> line of Makeconf, or reconfigure
>    ./configure --without-tcltk
> If your users do need this package then you (or they) need to track down
> where tcl and tk lurk and provide this information to R. Probably the
> easiest way would be to install a new version of Tcl and Tk, which should
> produce an honest set of configuration files.

hmm, that used to be the case.  Up to around 8.0, 8.1? tcl/tk was
great in its installation.  Now it, eh, sucks.  The installation of
8.3 and 8.4 has more bugs than the one Dan pointed out.

Dan: you can look at tclConfig.sh and tkConfig.sh.  Some files are
referenced incorrectly--you have to change the filenames to the
installed version instead of the build version.  You'll also have to
get rid of comments in compiler flags.

mailto:tov@ece.cmu.edu (Tom Vogels)   Tel: (412) 268-6638   FAX: -3204
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