Re: [Plplot-devel] Coordinate wrapping problems for the python, tcl, and java front ends

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

OK.  I have fixed this problem for the local (to me) python version of
plshades which I have been working on as a test case.  The trick was to have
a pointer zused which is used for the PLplot library calls rather than z.
For pltr0 and pltr1, and wrap=0 with pltr2, zused = z, but for pltr2 and
wrap=1 or 2, zwrapped is allocated and filled appropriately from z with one
row or column wrapped, and zused = zwrapped. Finally, zwrapped is freed
again for pltr2 and wrap = 1 or 2.

These changes seem to completely fix the bad behaviour noted before. The
python test of wrap=2 (with the last theta = 2pi - delta theta) gives an
identical result to the C example without wrapping but with the last theta =
2 pi.  wrap=1 makes a fairly weird plot (as expected), but the segfault is
gone.

I am satisfied this is the right approach so I will extend it to plcont and
plshade and plshades for the python, tcl, and java front ends.  Hopefully,
I will get all the changes checked in today.

Alan

email: ir...@be...
phone: 250-727-2902	FAX: 250-721-7715
snail-mail:
Dr. Alan W. Irwin
Department of Physics and Astronomy,
University of Victoria, P.O. Box 3055,
Victoria, British Columbia, Canada, V8W 3P6
__________________________

Linux-powered astrophysics
__________________________