[Plib-devel] Quaternion multiplication bug in sg in plib v.1.2.0

[Sam S. <sa...@sp...>]

Hi,

The quaternions multiplication routines in v1.2.0 appear to be broken.  I knocked up some code to demonstrate this, that's attached.

Basically it does this:

i  - identity quaternion
a - (x, y, z, w) = (0, cos(1.25), sin(1.25)
r - result

Doing: sgMult(&r, &a, &i) in 1.1.11 gives:

x:    0.315322
y:   -0.000000
z:    0.948985
w:   0.000000

and doing: sgMult(r, a, i) in 1.2.0 gives:

x:    0.948985
y:    0.000000
z:    0.315322
w:  -0.000000

so it appears x <-> z and y<-> w.  I haven't had the time to check the plib code to find the error yet - I just hacked the transform
into my code :/

I'm surprised this one slipped by (surely people must use quaternions) - so maybe it's my code?

(Btw, I did notice that the quaternion layout is different between these versions)

Sam