RAM $11E4 and adding Gau to a party?

haha, i've never coded anything involving this..  but remember being confused about a hypothetical related to it, and asking Lenophis for help making sense of it.

there's a dummied event in Albrook where if Gau is in the party when Terra and Locke are about to go to Thamasa, he'll decline to accompany them, and then Gau gets removed from the party, un-enlisted, and Byte $1F46, bit 7 (which iirc governs his presence in Vector Cafe) gets set.

normally, this can't happen, as he'd already been removed from the party and put walking around in Vector.  however, the question stemmed from: what if you had leapt Gau previous to the Vector deal, and had the airship now to retrieve him for Albrook?

would he magically be teleported to Vector anyway after the dummied convo?  and would he simultaneously be pacing around there *and* recruitable on the Veldt?

i forget the exact questions and whether there were more, and even the answers.  but it was interesting, and a bit mind-racking with the contradictions and trying to figure out what takes precedence.

------

as for your current deal, i think it's #3 and that you'll be fine.

see the code starting at C2/2FB0.  looks like the recruitment/enlistment flag in $1EDF wins out.

there are a number of things that can clear $11E4 bit 0:
- that C2/2F2F function (also due to party count and enemy formation [as in group]), called at battle start.
- Function C2/2E68 (due to encounter formation [type]), called at battle start and on formation switches.
- C2/4845 (reached upon ending/winning a battle), oddly even when you randomly branch and Gau *doesn't* rejoin.  can't say i understand the reasoning.

with all those things able to clear $11E4 bit 0 so readily, there's gotta be *something* to restore it entering or between battles.  i'm not seeing Bank C2 battle code offhand to initialize the variable or fill its vicinity with FFhs.  so i wanna say it's defaulted to enabled (by Bank EE or C0 or something) when you're on the Veldt.  EDIT: yeah, it and the other 3 bottom bits of $11E4 are set for Veldt steps at C0/C17B.  and it (and most others) are cleared right before a non-Veldt battle at C0/C270.

all this back-and-forth, with broad setting sometimes followed by worry-free clearing, reflects that there's a more persistent variable to keep track of Gau's enlisttment: Variable $1EDF Bit 3.  iow, $11E4 is a bit imprecise.  fwiw, i think the answer to my "and would he simultaneously be pacing around there *and* recruitable on the Veldt?" question was "yes", but would have to double-check.

to avoid nonsense like hypothetical Vector, guess you'd have to recalculate things like a room's possible denizens every time it's entered?

also, i think Square was probably wrong to un-enlist Gau in the dummied event (just party removal should have sufficed), because i don't believe there's a way to keep him off the Veldt once that's happened.  but i can't fault Square, because it's ...well... dummied.