My updated patches

[David Bateman]

I've made some changes to the previous patches I sent and added a new
patch. Find the patches attached.

The patch5, adds the pie, surfc, and quiver functions, adds these to
Makefile.in(SOURCES) together with surf, etc that were missing and
documents them in plot.txi (in a minimalist fashion). The reason to
include surfc, pie and quiver are that then these can be removed from
octave-forge and the chance of getting strange effects due to calling
gnuplot directly can be eliminated at least from octave-forge. All three
of these functions are new implementations, though surfc is based
heavily on meshc. The 'filled' argument of quiver doesn't work
correctly, not because its incorrectly implemented, but rather because
gnuplot can't fill markers as far as I can tell as so __go_draw_axes__
doesn't respect the makerfacecolor field. The quiver function doesn't
implement quiverseries objectis but rather gives a list of handles for
the body, head and marker of the arrows in a similar manner to matlab
v6, and so should be easy to adapt to a quiverseries in the future by
moving these objects to be children of the quiverseries. The change
since the last version of this patch is the addition of quiver and demos
in the pie function

patch7 is the changes necessary to __go_draw_axes__ to allow surf, et al
to work with gnuplot 4.0 and a fix for meshc that was completely broken
by the changes to surf. This second change require that the nd=4 case
of  __go_draw_axes__ to be folded back into the nd=3 case with a flag
for the presence of the cdata. The change since the last version is that
if a line object has linestyle and marker fields set to "none" then it
prevents gnuplot plotting anything. This is what matlab appears to do so
that the handles returned from quiver and be manipulated with the set
command to make parts of the quiver plot appear and disappear (typically
either the arrow or the marker but not both).

patch9 is a fix to the gradient function that doesn't work correctly if
dx or dy are scalar and the Z matrix is not square. I noticed this issue
while running the example for quiver from the matlab website. This is a
new patch.

Regards
David

patches.tar.bz2
Description: BZip2 compressed data