I don't particularly feel that contributions from fluxsort and quadsort are being minimized in Orson's writing/speaking. The talk in May and README are both pretty abbreviated and the places you should expect full attribution are the FOSDEM talk and upcoming paper; first looks good. We all tend to think of our own work as more important than it really is. And to overlook differences and assume it's more similar to other work.
I haven't looked into quadsort deeply and still haven't had any time to seriously read glidesort, so I can't comment on the details of merge implementations. The places I see influence are in merging 4 arrays at a time (which I did read about before quadsort in "Patience is a virtue", but I think neither I nor Orson took that paper very seriously) and in the bidirectional idea which is credited in the FOSDEM talk.