This is already checked in to master, as you know.
I checked in the files that are produced when the demo is run without any alterations to the config. I can't really test it, but it should produce the same results as buildBot.
I did test the sdss demo using the old pipeline and configuring it so that GaussianFlux has fixed=True. In this configuration, we match against the old pipeline to within .5%.
As noted in the
DM-1071 ticket changes are expected because of the change to GaussianFlux, a sub-percentage change to PsfFlux, and because the EDGE condition throw is handled differently in the new framework. Otherwise, the two pipelines match.