patchm vertex ordering?

Here I have two triangles with vertices resting on the unit sphere, specified in clockwise order looking at the outer surface of the sphere. Plotting vgood with patchm gives the expected result, a triangle in the projected map. However, vbad is interpreted as the inverse, a polygon covering the entire globe, leaving the triangle as a hole.

Given that both triangles have the same vertex ordering, I would expect them to have the same behavior. What's the problem here?

vbad = [ ...
    -0.6375993113323332 -0.20403949778403738 -0.7428559763053484; ...
    -0.8500194502416963  0.09199320574455130 -0.5186561330088028; ...
    -0.8113166555881062 -0.28694517808882358 -0.5093405041196936];

vgood = [ ...
    -0.6375993113323332 -0.20403949778403738 -0.7428559763053484; ...
    -0.6518705906365251 0.19086859353491015 -0.7339168298010295; ...
    -0.8500194502416963 0.09199320574455130 -0.5186561330088028];

% convert to lat/long
llbad = [-acos(vbad(:,3)) atan2(vbad(:,2),vbad(:,1))]*(360/(2*pi));
llbad(:,1) = llbad(:,1)+90;
llgood = [-acos(vgood(:,3)) atan2(vgood(:,2),vgood(:,1))]*(360/(2*pi));
llgood(:,1) = llgood(:,1)+90;

axesm pcarree;
patchm(llbad(:,1), llbad(:,2), 'FaceColor', 'r', 'EdgeColor', 'k');
patchm(llgood(:,1), llgood(:,2), 'FaceColor', 'g', 'EdgeColor', 'k');

[William]

Looks like this is a bug, there is a bug report on this same issue that is claimed to be fixed in 2010b, however I'm using 2011a and I am having the same problems.

http://www.mathworks.com/support/bugreports/641390