Hello,
Thanks for the feedback, that looks like a bug in the new preview of the V3 version. It's detecting it as a singularity because there is a large amount of axis movement in a small amount of time. However here that seems like a bug with the interpolation between the opposing planes. I'll look into it, but it will be a short while.
PTP works because it interpolates on an axis level. LIN should either take the time from the linear speed or axis speed, whatever is less. In that case LIN (and SPL) are failing because they take the time of the 0mm distance (i.e. they "jump"), rather than the X degrees rotation on the slowest axis. I haven't ever used SPL movements for just reorientation (i.e. TCP stays at the same point, while ABC change), but I guess it should work like LIN.
I appreciate the feedback, sorry for the inconvenience!
Best,
Johannes