I have the following Haskell program I wrote, the purpose of which is to function like a 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
My ghc version is as follows:
$ ghc --version
The Glorious Glasgow Haskell Compilation System, version 8.6.5
I compile the Haskell on macos (Catalina):
$ 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 ...
Then I run my code:
$ 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)
Why do I get Prelude.chr: bad argument? What causes this, and what can I do to fix the problem?
I have read about others who have had this error, but in their case, deleting the *.hi files solved the problem. I have deleted Prog1d.hi (as well as Prog1d.o and Prog1d), but to no effect. I feel this may be caused by something in my code, maybe with line 41:
then putStrLn $ show $ map (shiftLeft $ read $ aInt) message
But this line is just like line 38, which works just fine for the -encode use-case. I must be missing something obvious.
I am new to Haskell, so please help me along. I am mostly used to writing code in imperative languages like C++, python, Java, etc. and I am not yet familiar with the ideas and syntax of Haskell and other functional languages.
Thanks for reading this!
I found the problem.
In my code, I convert a character into its corresponding ASCII code.
For example, fromEnum 'A' will give you 65. The problem is that, if you shift the value off the table of ASCII values, e.g. -1, you cause a bad argument error for the toEnum I use to compare for checking if the shift went too far.
For example:
echo "A" | ./Prog1d -decode 66
This will take me to line 25 in my program:
25 then if (toEnum (fromEnum char - shift) :: Char) < 'A'
fromEnum char will be fromEnum 'A' which will give me 65.
Then I subtract from 65 the shift value: 66. So, 65 - 66 = -1.
Then toEnum -1 causes Prelude.chr: bad argument: (-1).
That's because there is no ASCII character for -1 and that value is out of bounds.
This means I simply have to compare int values instead of chars to check whether my shift is out of bounds or not.
Here is the corrected code:
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
And here is the proper, desired output:
$ 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
EDIT (2020-09-01): John Purdy gave me some feedback, which prompted me to radically refactor my code. Here is the improved version:
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