1* Version 1.9.2:
2
3- Lifted hardcoded upper limit on the number of warriors.
4
5* Version 1.9:
6
7- Flags in instructions aren't masked when decoding instructions
8  during simulation anymore.  You should either clear the flags fields
9  before loading, or use the new function sim_load_warrior() to load
10  your warriors.  If that's not feasible, define the preprocessor
11  variable SIM_STRIP_FLAGS to be 0 when compiling.
12
13  As far as I can tell nobody was using the flags field anyway.  Sorry
14  if this breaks anyones code.
15
16- P-space!  See the README for details.
17
18- You need to link the file pspace.o whenever you are linking sim.o
19  regardless of whether p-space is in use or not.
20
21- To accomodate p-space the encoding of instructions has changed.  One
22  flag bit has evaporated to make room for LDP/STP, and the order of
23  enumeration of opcodes differs from version 1.7.  If you've been
24  using the macros in insn.h or the assembler, then there's nothing to
25  worry about.  Sorry if this breaks anyones code.
26
27- Now placed into the Public Domain.
28
29- asm.h/asm_line() can now return -3 == ASMLINE_PIN to identify the
30  PIN pseudo-op.
31
32- The warrior_st struct in exhaust.h has new fields `have_pin' and `pin'.
33
34- Other changes to fix some compiler warnings/sign oddities.  Some
35  arguments of public prototypes have changed in signedness.
36
37- No more ABORT_TIES_EARLY, but that's OK as 1.8 was never released.
38
39- The example program exhaust.c now supports multiwarrior battles and
40  p-space.  Slightly cleaned up as well.
41
42* Version 1.8:
43
44- New compile-time option ABORT_TIES_EARLY to detect unwinnable
45  configurations where there are no deadly instructions in core.
46