Finally got some time to look at this again. I'm doing side-by-side debugging between FSCI and SSCI's debuggers.
I start the game, then teleport to room 25 by setting global variable 13. I then set a breakpoint on PQ::newRoom. Here is the first different I see: pc=2869 acc=b92e o=44a2 fp=5332 sp=533c prev=1 sbase=530c globls=2380 &restmod=0 Step #683302 2869: [B] callk NextNode[38] 02 Kernel params: (b92e) > pc=286c acc=b938 o=44a2 fp=5332 sp=5338 prev=1 sbase=530c globls=2380 &restmod=0 Step #683303 286c: [B] sat 01 > pc=286e acc=b938 o=44a2 fp=5332 sp=5338 prev=1 sbase=530c globls=2380 &restmod=0 Step #683304 286e: [W] push1 > pc=286f acc=b938 o=44a2 fp=5332 sp=533a prev=1 sbase=530c globls=2380 &restmod=0 Step #683305 286f: [W] push1 > In SSCI, after NextNode is called, the accumulator is 0. I'm assuming this means the end of the list. FSCI has one more item in the list for some reason, but it doesn't appear to be a heap object. There are other differences after this (breakpoint on PQ::startRoom to notice one immediately), but this is the first one I noticed. If someone would like to meet me on IRC tomorrow to help me with the next step of this investigation, I would appreciate it. -- http://www.clock.org/~matt
