Yes, all of the single lines having second statements that are always executed look a bit funky to me too, but I had taken Jez to be pointing us specifically to the interesting bit where the code generated for 'area-=(yy+1)' is not correct independent of what any subsequent statements do or don't do:

mov a,#0x01
add a,_yy
mov r7,a ; <-- Missing
clr c
mov a,_area
subb a,a ; <-- Wrong! Should be subb a,r7
mov area,a