Is ResolveSplitPenetration supposed to use resultUpperLess?

[TrianglesPCT]

(In file btSequentialImpulseConstraintSolver.cpp)

I noticed in gResolveSplitPenetrationImpulse_sse2 that it reads c.m_upperLimit and then does a comparison against it here:

resultUpperLess = _mm_cmplt_ps(sum, upperLimit1);

But resultUpperLess is never used, is this intentional?

I do see that the scalar reference does not use the upper limit at all.