Interesting. Is it behaving differently from the glider falling down an extremely narrow (ie. 1-pixel wide) vertical column of air (with walls on both sides)? I guess in this case, maybe the blocker actually turns the falling column into effectively 0-width?
I think we can at least agree that the glider probably should at least still be falling down vertically some nonzero amount in any one frame update, in situations where it is not landing but also restricted by nearby walls from moving any of the full amount it could've moved in one frame on open air. Handling of facing direction is interesting--especially in this blocker case, I guess it makes sense the lemming is forced facing right while it falls through the blocker field.
Kinda makes me wonder now what happens in Lix with tumblers (or other similar movements like jumping or getting batted) and blockers, that kind of situation there would be similar to the glider-blocker situation here. I seem to recall (maybe) seeing examples that show game will ensure vertical movement so that the lix won't get stuck indefinitely in the same location, but haven't really tested extensively there.