Fixed bug that Saturating operations on Index 1 were unsafe (#682).

[rowanG077]

I have added constant folding test for the specific case Index 1 case.

The parser doesn't detect yet whether it is actually constant folded since it only will trigger an error when it finds a specific number range. It's not possible to easily extend this with the correct number since for Index 1 all operations take and result in 0.

Creating a parser to check that no real operations are taking place would be quite an undertaking I think. What do you think @martijnbastiaan?

[martijnbastiaan]

Hey @rowanG077, thanks for your second PR :)!

I think the following would suffice:

  case (satAdd @(Index 1) SatWrap (lit 0) (lit 0)) of {0 -> 0; 1 -> 22270 :: Int}
  case (satSub @(Index 1) SatWrap (lit 0) (lit 0)) of {0 -> 0; 1 -> 22271 :: Int}
  ...

If Clash successfully constant folds this, you shouldn't see 22270 in the output. It might be that it doesn't actually constant fold this, but that it strips out the case entirely due to Index 1 being zero bits. I'm a bit fuzzy on the details, as it has been a while since I looked at this, but either way should be fine.

[martijnbastiaan]

LGTM, thanks!