Indeed it explains everything.
First, I just committed (to trunk only) the apt maps with various validation errors fixed (custom types for saved beds, layer conflicts) and while I was there turned the flags to face the right way.
But second, due to the way unique maps (ie, apts) are stored, existing ones override whatever the arches currently say is the right face for any given object. So although the test server wmask_bflag_fire1_1 arch says the faces are flag9.u.131/151 the unique maps saved of old in the players dir references wmask_bflag_fire1.u.131/151, hence the mapbug.
I see three solutions:
1. I do 'rpl -R "wmask_bflag_fire1.u.131" "flag9.u.131" testserv/server/data/players/*' on the testserv (and whatever else is needed for othe objects). I am reluctant to do this because while it should fix the immediate problem if it works properly, if it doesn't or I get confused, it is likely to make a tiny problem infinitely worse.
2. We just ignore the warnings until a char reset/wipe when it will no longer be a problem (although this also needs the trunk arches to be in branches/on main so iit'll be a while yet.
3. As above (to remedy the current problem) and someone fixes the unique map storage so that unique maps use the current arches like everything else.
But on the whole ISTM to be a very minor cosmetic issue with a known cause and only affecting test. So as the cure is (potentially) worse than the disease, just ignore it for now.