如何摆脱SBV中-0.0的解决方案

Question

以下代码(使用SBV)：

{-# LANGUAGE ScopedTypeVariables #-}
import Data.SBV
main :: IO ()
main = do
    res <- allSat zeros
    putStrLn $ show res

zeros :: Predicate
zeros = do
    z1 <- sDouble "Z1"
    constrain $ z1 .== 0.0
    return true

生成两个解决方案：

Solution #1:
  Z1 = 0.0 :: SDouble
Solution #2:
  Z1 = -0.0 :: SDouble
Found 2 different solutions.

如何消除无趣的-0.0解决方案？我不能使用z1 ./= -0.0，因为当z1是0.0时也是如此。

Answer 1

[在提供适当支持的Hackage (http://hackage.haskell.org/package/sbv)上发布sbv4.4之后更新。]

假设您拥有SBV >= 4.4，现在可以这样做：

Prelude Data.SBV> allSat $ \z -> z .== (0::SDouble)
Solution #1:
  s0 = 0.0 :: Double
Solution #2:
  s0 = -0.0 :: Double
Found 2 different solutions.
Prelude Data.SBV> allSat $ \z -> z .== (0::SDouble) &&& bnot (isNegativeZeroFP z)
Solution #1:
  s0 = 0.0 :: Double
This is the only solution.

Answer 2

根据这些注释，以下代码只生成一个解决方案：

zeros :: Predicate
zeros = do
    z1 <- sDouble "Z1"
    constrain $ z1 .== 0.0 &&& ((1 / z1) .> 0)
    return true

输出：

Solution #1:
  Z1 = 0.0 :: SDouble
This is the only solution.