labels - failing when trying to load a label file with no noise components.

xform. Transform module has new flirtMatrixToSform function which
turns a FLIRT matrix into a src -> struc sform transformation.

returning/assuming transposed affines (column-major). This is because that's
what OpenGL uses.

when it is. New functions in transform module.

target space is 'voxel'), and will optionally not clip the
result. VolumeOpts.getVoxel allows coordinates to be passed
in. Transform.axisBounds allows an offset to be applied to voxel
coordinates. New method, Image.axisMapping, which basically wraps
nib.aff2axcodes.

function from transform module, as it is dodgy, and was not being used
anywhere. I think I wrote it before writing the axisBounds function.

displayed with the bottom left voxel at display location (0, 0). This
means that calculating voxel location from display coordinates is
different depending on the space that the image is displayed in. I think
I've figured out the location panel, but need to double check, and need
to take a closer look at the editor.

previously assuming that voxel axes equaled display space axes. Also,
bugfix in transform.scaleOffsetXform

vertex shader.

2. Bounds test used in gl21/[tensor/volume]_frag moved to separate file.

3. Transformation matrix offset by 0.5, in Display class, removed, as
I've decided that it should be unnecessary.

called. Fixed. GL21/glvolume no longer using PyOpenGL.vbo module, as it
is rubbish.

calculated in one call.

numpy.dot). And a few tweaks elsewhere

between each of the three spaces (voxel, display, and world)

matrix), LightBoxPanel (not passing displayCtx to LightBoxCanvas),
fslview_parseargs (using imageList.bounds instead of displayCtx.bounds),
ImageList (didn't remove bounds-related code), ImageDisplay (attempting
to offset transformation matrices for affine, when only supposed to do
it for id/pixdim)

to use transform module instead of image.voxToWorld/worldToVox methods

coordinate systemss, with the ultimate aim of having the location panel
always reporting image world coordinates, regardless of the way in which
images are being displayed.