SPIN gripes

• Error messages are inane.
• Preprocessor is hardcoded
• Doesn’t understand hex (e.g., 0xFF12)
• d_step compiler constant too small
• non-deterministic if is hard to translate
• Needs better assert() and panic()
• Can’t nest d_step and/or atomic
• Generated traces too far from source
  • X = (1 - 1) instead of X = 0
  • skip shouldn’t show up as (1)