是什么导致“Prelude.chr:错误的论点”?

What causes `Prelude.chr: bad argument`?

我编写了以下 Haskell 程序,其目的是像 Caesar cipher:

  1 import System.IO
  2 import System.Environment
  3 import System.Exit
  4 import Data.Char
  5 
  6 shiftRight :: Int -> Char -> Char
  7 shiftRight shift char = do
  8   if isAsciiLower char
  9   then if (toEnum (fromEnum char + shift) :: Char) > 'z'
 10     then shiftRight (shift - 26) char
 11     else toEnum (fromEnum char + shift) :: Char
 12   else if isAsciiUpper char
 13     then if (toEnum (fromEnum char + shift) :: Char) > 'Z'
 14       then shiftRight (shift - 26) char
 15       else toEnum (fromEnum char + shift) :: Char
 16     else char
 17     
 18 shiftLeft :: Int -> Char -> Char
 19 shiftLeft shift char = do
 20   if isAsciiLower char  
 21   then if (toEnum (fromEnum char - shift) :: Char) < 'a'
 22     then shiftLeft (shift + 26) char
 23     else toEnum (fromEnum char - shift) :: Char
 24   else if isAsciiUpper char
 25     then if (toEnum (fromEnum char - shift) :: Char) < 'A'
 26       then shiftLeft (shift + 26) char
 27       else toEnum (fromEnum char - shift) :: Char
 28     else char
 29 
 30 main = do 
 31   args <- getArgs
 32   message <- getLine
 33   case args of
 34     [aString, aInt] -> 
 35       if aString == "-encode"
 36       -- `read` converts aInt from string to int
 37       -- `map` is used to apply `shiftRight` to each char in the string `message`
 38       then putStrLn $ show $ map (shiftRight $ read $ aInt) message
 39       else 
 40         if aString == "-decode"
 41         then putStrLn $ show $ map (shiftLeft $ read $ aInt) message
 42         else do
 43           putStrLn ("Second argument should be either '-decode' or '-encode'!")
 44           exitFailure
 45     _ -> do 
 46       progName <- getProgName
 47       putStrLn ("Usage: " ++ progName ++ " [-encode|-decode] [0-9]")
 48       exitFailure

我的ghc版本如下:

$ ghc --version
The Glorious Glasgow Haskell Compilation System, version 8.6.5

我在 macos (Catalina) 上编译 Haskell:

$ ghc Prog1d.hs -o Prog1d
Loaded package environment from $HOME/.ghc/x86_64-darwin-8.6.5/environments/default
[1 of 1] Compiling Main             ( Prog1d.hs, Prog1d.o )
Linking Prog1d ...

然后我运行我的代码:

$ echo "ABCXYZabcxyz" | ./Prog1d -encode 1
"BCDYZAbcdyza"
$ echo "ABCXYZabcxyz" | ./Prog1d -encode 2
"CDEZABcdezab"
$ echo "ABCXYZabcxyz" | ./Prog1d -encode 4
"EFGBCDefgbcd"
$ echo "ABCXYZabcxyz" | ./Prog1d -encode 100
"WXYTUVwxytuv"
$ echo "ABCXYZabcxyz" | ./Prog1d -decode 1
Prog1d: Prelude.chr: bad argument: (-14)
$ echo "ABCXYZabcxyz" | ./Prog1d -decode 2
Prog1d: Prelude.chr: bad argument: (-15)
$ echo "ABCXYZabcxyz" | ./Prog1d -decode 4
Prog1d: Prelude.chr: bad argument: (-17)
$ echo "ABCXYZabcxyz" | ./Prog1d -decode 100
Prog1d: Prelude.chr: bad argument: (-35)

为什么我得到 Prelude.chr: bad argument?是什么原因造成的,我该怎么做才能解决这个问题?

我读到过其他人遇到此错误,但在他们的情况下,删除 *.hi 文件解决了问题。我删除了 Prog1d.hi(以及 Prog1d.oProg1d),但没有任何效果。我觉得这可能是由我的代码中的某些内容引起的,可能是第 41 行:

then putStrLn $ show $ map (shiftLeft $ read $ aInt) message

但是这一行就像第 38 行一样,它适用于 -encode 用例。我一定遗漏了一些明显的东西。

我是 Haskell 的新手,所以请帮助我。 我主要习惯用C++、python、Java等命令式语言编写代码,对Haskell等函数式语言的思想和语法还不熟悉

感谢阅读本文!

我发现了问题。

在我的代码中,我将一个字符转换成它对应的ASCII码。

例如fromEnum 'A'会给你65。问题是,如果您将值从 ASCII 值的 table 移开,例如-1,你导致 toEnum 的错误参数错误,我用来比较以检查偏移是否过大。

例如:

echo "A" | ./Prog1d -decode 66

这将带我到程序中的第 25 行:

25     then if (toEnum (fromEnum char - shift) :: Char) < 'A'

fromEnum char 将是 fromEnum 'A' 这将给我 65.

然后我从 65 中减去 shift 值:66。所以,65 - 66 = -1.

然后 toEnum -1 导致 Prelude.chr: bad argument: (-1).

那是因为 -1 没有 ASCII 字符,该值超出范围。

这意味着我只需比较 int 值而不是 chars 来检查我的班次是否越界。

这是更正后的代码:

import System.IO
import System.Environment
import System.Exit
import Data.Char
 
shiftRight :: Int -> Char -> Char
shiftRight shift char = do
  if isAsciiLower char 
  then if (fromEnum char + shift) > 122 -- 122 is 'z' in ASCII
    then shiftRight (shift - 26) char
    else toEnum (fromEnum char + shift) :: Char
  else if isAsciiUpper char
    then if (fromEnum char + shift) > 90 -- 90 is `Z` in ASCII
      then shiftRight (shift - 26) char
      else toEnum (fromEnum char + shift) :: Char
    else char
 
shiftLeft :: Int -> Char -> Char
shiftLeft shift char = do
  if isAsciiLower char 
  then if (fromEnum char - shift) < 97 -- 97 is 'a' in ASCII
    then shiftLeft (shift - 26) char
    else toEnum (fromEnum char - shift) :: Char
  else if isAsciiUpper char
    then if (fromEnum char - shift) < 65 -- 65 is 'A' in ASCII
      then shiftLeft (shift - 26) char
      else toEnum (fromEnum char - shift) :: Char
    else char
 
main = do
  args <- getArgs
  message <- getLine
  case args of
    [aString, aInt] -> 
      if aString == "-encode"
      -- `read` converts aInt from string to int
      -- `map` is used to apply `shiftRight` to each char in the string `message`
      then putStrLn $ map (shiftRight $ read $ aInt) message
      else 
        if aString == "-decode"
        then putStrLn $ map (shiftLeft $ read $ aInt) message
        else do
          putStrLn ("Second argument should be either '-decode' or '-encode'!")
          exitFailure
    _ -> do 
      progName <- getProgName
      putStrLn ("Usage: " ++ progName ++ " [-encode|-decode] [0-9]")
      exitFailure

这里是正确的、期望的输出:

$ echo "ABCXYZabcxyz" | ./Prog1d -encode 1
BCDYZAbcdyza
$ echo "ABCXYZabcxyz" | ./Prog1d -encode 2
CDEZABcdezab
$ echo "ABCXYZabcxyz" | ./Prog1d -encode 4
EFGBCDefgbcd
$ echo "ABCXYZabcxyz" | ./Prog1d -encode 100
WXYTUVwxytuv
$ echo "ABCXYZabcxyz" | ./Prog1d -decode 1
ZABWXYzabwxy
$ echo "ABCXYZabcxyz" | ./Prog1d -decode 2
YZAVWXyzavwx
$ echo "ABCXYZabcxyz" | ./Prog1d -decode 4
WXYTUVwxytuv
$ echo "ABCXYZabcxyz" | ./Prog1d -decode 100
EFGBCDefgbcd

编辑 (2020-09-01):John Purdy 给了我一些反馈,这促使我从根本上重构我的代码。这是改进后的版本:

import System.IO
import System.Environment
import System.Exit
import Data.Char

shift :: Int -> Char -> Char
shift amount char
  -- `ord` is `fromEnum` but only for `Char` types; converts a Char to an Int, 
  -- which gives us the ASCII number for that character
  -- see: https://hackage.haskell.org/package/base-4.14.0.0/docs/Data-Char.html#v:ord
  | isAsciiLower char && (shifted < ord 'a' || shifted > ord 'z') = shift cycled char
  -- `chr` is `toEnum` but only for `Char` types; it converts from an Int to Char
  -- see: https://hackage.haskell.org/package/base-4.14.0.0/docs/Data-Char.html#v:chr
  | isAsciiLower char = chr shifted
  | isAsciiUpper char && (shifted < ord 'A' || shifted > ord 'Z') = shift cycled char
  | isAsciiUpper char = chr shifted 
  | otherwise = char
  where shifted = ord char + amount -- e.g. 'A' (ASCII: 66) shifted 1 = 'B' (ASCII: 67), so shifted would be 67. 
        -- cycled: the shift amount, but cycled by 26 to start over the alphabet, 
        -- e.g. 'Z' (by ASCII: 90) shifted 1 is 91, so -26 to get 65, which is 'A'
        -- the amount `div` amount makes sure we add or subtract 26 as needed to cycle 
        -- e.g. 'A' (65) shifted -1 is 64, but if we subtracted 26 from 64, it wouldn't cycle us to Z, 
        -- it is in the wrong direction
        -- so instead we multiply by the sign of the amount: -1 - (26 * (-1 / |-1|) to get +25, 
        -- so 65 ('A') + 25 = 90 ('Z')
        cycled = amount - ((amount `div` abs (amount)) * 26) 

main = do
  args <- getArgs
  message <- getLine
  case args of
      -- `read` converts aInt from string to int
      -- `map` is used to apply `shiftRight` to each char in the string `message`
    ["-encode", aInt] -> putStrLn $ map (shift $ read $ aInt) message
    ["-decode", aInt] -> putStrLn $ map (shift $ negate $ read $ aInt) message
    _ -> do 
      progName <- getProgName
      putStrLn ("Usage: " ++ progName ++ " [-encode|-decode] [0-9]")
      exitFailure