It’s too late for today, so the program is untested. I’ll probably test it on my calc the other day when I got more time to waste.
In short, this calculator is extremely limited. Hats off to the contributors at https://algos-casiofx92.com/; you guys are genius.
Note: Access X&Y registers with “Aller a x:y”
The program space is rather limited. Instead of doing I/O the way I’m doing it (by using two registers as the “streams”), you can modify the program to redirect the I/O functions to “Demander valeur” and “Afficher result”. In that case the maximum BF program length is 16*5=80 steps.
Maximum Length: 16*3 Steps X:PC : Program Counter Y:XS : Execution State A:AC : Accumulator B:IS : Input Stream C:OS : Output Stream D:SC : Scratch Pad E:P1 : Program Space 1 F:P2 : Program Space 2 M:P3 : Program Space 3 ; Opcode Encoding: ; 0: NOP 1: > 2: < 3: + 4: - ; 5: . 6: , 7: [ 8: ] ; Limitations: ; 1. Any datum > 9 will corrupt nearby values in the scratch pad ; Fetch Instruction Ent(PC/16)+1->AC Si AC=1 Alors Ent(10*(P1*10^(PC-AC*16)-Ent(P1)*10^(PC-AC*16))->AC Sinon Si AC=2 Alors Ent(10*(P2*10^(PC-AC*16)-Ent(P2)*10^(PC-AC*16))->AC Sinon Ent(10*(P3*10^(PC-AC*16)-Ent(P3)*10^(PC-AC*16))->AC Fin Fin ; A State Machine for [ ] Instructions Si XS=1 Alors PC+1->PC Sinon Si XS=2 Alors PC+1->PC Si AC=8 Alors XS=1 Fin AC=0 Sinon PC-1->PC Si AC=7 Alors XS=1 Fin AC=0 Fin Fin ; Execute the Instruction ; This part is rather ugly. The culprit it the 3-level return limitation of the calculator. Si AC<=4 Alors Si AC<=2 Alors Si AC=1 Alors SC*10->SC ; > Sinon Si AC=2 Alors SC/10->SC ; < Fin Fin Sinon Si AC=3 Alors SC+1->SC ; + Sinon SC+1->SC ; - Fin Fin Sinon Si AC<=6 Alors Si AC=5 Alors ; . Ent(10*((SC/10)-Ent((SC/10))))->AC OS*10+AC->OS Sinon ; , Ent(10*((IS/10)-Ent((IS/10))))->AC ((Ent(SC/10))*10+(SC-Ent(SC)))+AC->SC Fin Sinon Si AC=7 Alors ; [ Si Ent(10*((SC/10)-Ent((SC/10))))==0 2->XS Fin Sinon ; ] Si Ent(10*((SC/10)-Ent((SC/10))))!=0 3->XS Fin Fin Fin Fin