The original DA 40/2.8 has better resolution and performs much better regarding CA. Yet, the new XS version gets a better DxOMark Score.
Who cares about T-Stops (a 3.1 vs 2.9 difference anyhow) or minute differences in vignetting?
I applaud your sensor testing and white papers, but your lens testing seems broken.
All Pentax lenses seem to be penalised for comparatively low T-Stops but I doubt that with today's high-ISO performances, this aspect of lenses really matters.