Haskell 中的 SHA1 编码

Question

I have a list of filepaths and want all these files to store as sha1 encoded hash in a list again. It should be as general as possible, so the files could be text as well as binary files. And now my questions are:

1. What packages should be used and why?
2. How consistent is the approach? With that I mean: if there could be different results with different programs using sha1 for encoding itself (eg sha1sum)

Answer 1

The cryptohash package is probably the simplest to use. Just read your input into a lazy ¹ ByteString and use the hashlazy function to get a ByteString with the resulting hash. Here's a small sample program which you can use to compare the output with that of sha1sum .

import Crypto.Hash.SHA1 (hashlazy)
import qualified Data.ByteString as Strict
import qualified Data.ByteString.Lazy as Lazy
import System.Process (system)
import Text.Printf (printf)

hashFile :: FilePath -> IO Strict.ByteString
hashFile = fmap hashlazy . Lazy.readFile 

toHex :: Strict.ByteString -> String
toHex bytes = Strict.unpack bytes >>= printf "%02x"

test :: FilePath -> IO ()
test path = do
  hashFile path >>= putStrLn . toHex
  system $ "sha1sum " ++ path
  return ()

Since this reads plain bytes, not characters, there should be no encoding issues and it should always give the same result as sha1sum :

> test "/usr/share/dict/words"
d6e483cb67d6de3b8cfe8f4952eb55453bb99116
d6e483cb67d6de3b8cfe8f4952eb55453bb99116  /usr/share/dict/words

This also works for any of the hashes supported by the cryptohash package. Just change the import to eg Crypto.Hash.SHA256 to use a different hash.

_{¹ Using lazy ByteStrings avoids loading the entire file into memory at once, which is important when working with large files.}

Answer 2

As to @hammar's answer, it is excellent but you can use Base16 library instead of making your own toHex .

import qualified Data.ByteString.Base16 as B16
hashFile path >>= putStrLn . B16.encode