depending on minetest_game beds breaks player_in_bed counter
to reproduce a bug, start a server with minetest_game and homedecor, build a red bed from beds, and a white one from homedecor. sleep in white one and press nothing so that the form quits after 2 seconds. get into the red bed: the player counter for players in bed is now broken. generally, sleeping in white beds decrements this counter by one.
this bug may or may not affect the ability to sleep (waiting for majority), since that particular counter may or may not be consulted for majority assertion, which can instead be completed by combing through the beds hash.
the source of this bug is the fact that player_in_bed counter and lay_down function are local in minetest_game beds mod. we don't see an easy way to fix this. homedecor lays players down without incrementing the counter via its own local lay_down. but then the counter gets decremented in at least 2 places: when homedecor calls beds.kick_players, as well as via beds formspecs, which call beds' local version of lay_down (to get up), which decrements the counter.
to fix this in a most elegant way, wethinks, one would have to expose more of the minetest_game beds (we understand that fixing homedecor bug this way may be frowned upon, but actually beds could benefit from more interface exposure, whereas homedecor would have to be aligned with it anyway, so it's a comprehensive all-around fix). exposing player_in_bed (as beds.player_in_bed) as well as lay_down (as beds.lay_down) would probably be enough, even with reusing the formspecs.
since there is no way to increment local player_in_bed counter from outside of the beds mod, any fix not touching minetest_game would have to stop using beds.kick_players and beds formspecs.