[Plplot-devel] Re: New CVS distribution tarball available

[Alan W. I. <ir...@be...>]

Don, I am forwarding my response to your e-mail to the plplot-devel list
because it is interesting to the Plplot developers that hang out there.  If you
want to respond on list that would be great (subscribe at
http://sourceforge.net/mail/?group_id=2915).

On 2004-01-16 11:51-0500 Don Spong wrote:

> Alan,
>
> I've tested this on three different systems that I work on:
>
> Mac OSX (10.3.2):
>    Here (as I think others have already confirmed) I still have to
> disable f77 and dyndrivers to get it to work.  ./plplot-test.sh
> produced 35 postscript files.  Examples 14 and 21 failed I guess
> because I don't have all of the tk/tcl available and due to problems
> with something called Qhull.

If you get a not-implemented message for example 14, 17, or 21 that is quite
normal. 14 and 17 are interactive examples that are not appropriate for those
file device results.  example 21 is suitable for file devices, but it has
different results each time which depend on computer load and speed of the
computer in question so it is not suitable for differencing with prior
results.

In most cases all these examples are explicitly skipped.  In a few cases the
examples are actually executed knowing full well that only a not-implemented
message comes back from that particular language interface.

Note, if my description doesn't describe the messages that you are seeing,
please let us know exactly what they are.

> I would continue to hope that the problems with Fortran on Mac OSX
> can be resolved.  I have a lot of legacy Fortran code that I use.
> While in my office I can always run this on my Linux system, but it
> would be nice if I could be able to also run this on my Mac portable
> when I'm on travel.

We have had two very recent reports of success with fink/g77 on MacOS X with
the latest tarball.  Those reports were on the plplot-devel list.
Unfortunately the archives seem to be down now so I cannot refer you to
those posts.

In short, the trick seems to be to set FLIBS.  But it turns out the correct
values were different on two different systems!

export FLIBS='-L/sw/lib -lfrtbegin -lSystem'
./configure ...

succeeded on one system, but extra flags were required on another system
to get fortran to work, e.g.,

export FLIBS='-L/sw/lib -lfrtbegin -lSystem -lgcc -lcc_dynamic'
./configure ....

> Most of the discussion I read seems centered on
> the g77 compiler.  I have also been using IBM's xlf Fortran compiler;
> is it possible Plplot could be made to work with this as well?

I frankly don't know.  But you could try setting the F77 make file variable,
e.g.,

make F77=xlf

and

make F77=xlf install

You will probably want to set FFLAGS as well, and you will probably have to
fool with FLIBS as well.

> RedHat Linux 7.2:
>    Everything worked okay here.  Again, examples 14 and 21 failed I
> guess because I don't have all of the tk/tcl stuff available and
> because of problems with Qhull.  ./plplot-test.sh  produced 51
> postscript files.

Same comment on examples 14 and 21. That looks like a clean test, for
that platform unless I haven't satisfactorily explained the messages
you are seeing.  Thanks for that information.

>
> IBM AIX:
>    I was not able to successfully install here.  I tried what I
> thought had worked last time I tried AIX, which was: --disable-cxx
> --disable-f77 --disable-tcl, but even that didn't fix the problems.

An IBM AIX user is a valuable resource to us (!) because we don't have many
of those to test our tarball.

Could you give us a more detailed report (preferably to plplot-devel, and
note you can subscribe at http://sourceforge.net/mail/?group_id=2915)?

What's needed here is the complete results of
your ./configure, make, and make install steps.  If you get that far capturing
the results of the installed examples make command would be helpful as well.

I usually capture output from the appropriate commands using
e.g., ./configure --myoptions >& configure.out, etc.

>
> -Don
>
> Donald A. Spong, Fusion Energy Theory, ORNL
> **NOTE ADDRESS CHANGE**: P. O. Box 2008, Bldg. 5700, Room R-F312
> Oak Ridge, Tennessee  37831-6169
> Phone: (865) 574-1304   FAX: (865) 576-7926
> E-mail: sp...@or...
> web page: http://www.ornl.gov/fed/Theory/stci/stellarator_theory.html

Thanks very much for your report, Don.

Alan
__________________________
Alan W. Irwin
email: ir...@be...
phone: 250-727-2902

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the PLplot scientific plotting software
package (plplot.org), the Yorick front-end to PLplot (yplot.sf.net), the
Loads of Linux Links project (loll.sf.net), and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________