On 2006-04-27 01:42+1000 Andrew Roach wrote:
> At 08:58 AM 26/04/2006 +0200, you wrote:
> Cygwin/wingcc:
>> 0. I got some messages about the Python version I have installed -
>> something about needing a development version? Given the
>> other reports on the subject of Python builds, and my lack of
>> knowledge on Python versions/installations, I have not looked
>> into it further.
>
> I get that error on MinGW too, but I don't even have python installed !
That's the whole point. The default configuration is --enable-python. Thus,
it looks for a valid python and if you don't have it, it warns you. To stop
that warning either (a) --disable-python (which testers should not do if
they want a complete test of the tarball on their platform) or (b) install
the python and Numeric packages.
Arjen, if you would like to complete your Cygwin tests for python, you
should install the combination of
http://cygwin.com/packages/python/python-2.4.1-1 and
http://cygwin.com/packages/Numeric/Numeric-24.2-1.
Andrew, if you would like to complete your MinGW tests for python, I think
you must use a native windows python (which I assume is available from
python.org) or else build your own following
http://jove.prohosting.com/iwave/ipython/pyMinGW.html. For Numeric, a native
windows build is available from
http://sourceforge.net/project/showfiles.php?group_id=1369 or you can build
your own.
I know pre-release testing is time consuming, but python is an important
platform for us so I would appreciate anything you both could do following
the above suggestions for getting access to python/Numeric for your
platforms. If our python interface still doesn't work after you have
installed Python/Numeric, then please send _all_ the details....
configure.out, make.out, make_install.out, make_examples.out,
plplot-test.out (or however far you get in the configure, build, install,
build installed examples, and run examples sequence). I will take the time
to read through those *.out results, and try and find a solution before
release.
Also, thanks to both of you for trying --disable-dyndrivers. Sorry it
hasn't worked out on Cygwin and MinGW. My working hypothesis at the moment
from what you two have reported is that the general PLplot-related
dyn-drivers build is working fine on Cygwin, but there is something
specifically wrong with the libtool libltdlc library for the Cygwin platform
so the dynamic load does not work correctly, and a specific bug report to
the libtool group about that might solve the matter. In contrast, the MinGW
problem is more severe, and there is some general problem (perhaps similar
to the previous sed issue) keeping Andrew from even building the
PLplot-related part of the dyn-drivers build.
From my point of view, these dynamic driver issues are not as high a
priority as the python question (since --disable-dyndrivers works for all
devices except the psttf and wxwidgets devices that are disabled in that
case) so given pre-release time constraints we should let it slide for now.
However, if either of you would like to sort this out post-release (like we
previously did together for shared libraries), then I will try to find the
time to help you out. But I don't have that time now.
Alan
__________________________
Alan W. Irwin
Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).
Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); 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
__________________________
|