On the included portal object in namida_systemtest, the result is the receive point being 1 pixel further to the right than it should be. (Workaround for this in the meantime is to move all portals where this positioning is critical, eg. portals that receive lemmings who are climbing / sliding, one pixel to the left.)
Untested how flipped portals are affected and what the exact cause / impacts are, beyond the observed effects on the sample object.
