diff --git a/CHANGELOG.md b/CHANGELOG.md new file mode 100644 --- /dev/null +++ b/CHANGELOG.md @@ -0,0 +1,11 @@ +## Version 0.2.0.0 (15 Sep 2013) + +- Added option -v for version information + +- Added option -f and -t to filter files based on modification time + +- Changed option -i (exclude based to regex) to -e + +## Version 0.1.0.0 (14 Dec 2012) + +- Initial release. diff --git a/LICENSE b/LICENSE --- a/LICENSE +++ b/LICENSE @@ -1,4 +1,4 @@ -Copyright (c) 2012 Mauro Taraborelli +Copyright (c) 2013 Mauro Taraborelli All rights reserved. diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -14,15 +14,18 @@ user@host:~$ haskdeep - Usage: haskdeep COMMAND [-c|--computation MODE] [-r|--root DIRNAME] [-k|--known FILENAME] [-i|--ignore RULE] + Usage: haskdeep COMMAND [-c|--computation MODE] [-r|--root DIRNAME] [-k|--known FILENAME] Computes hashes and audit a set of files Available options: + -h,--help Show this help text + -v,--version Show version information -c,--computation MODE md5 | sha1 | sha256 | skein512 - default md5 -r,--root DIRNAME Root directory - default current directory -k,--known FILENAME Known hashes file - default known.haskdeep - -i,--ignore RULE Regex to ignore files or directories - -h,--help Show this help text + -e,--excl-regex REGEX Exclude files or directories based on regex + -f,--incl-mod-from DATE Include files modified from yyyy-mm-ddThh:mm:ssZ + -t,--incl-mod-upto DATE Include files modified up to yyyy-mm-ddThh:mm:ssZ Available commands: compute Computes file hashes and saves them to known hashes file @@ -32,11 +35,11 @@ 1. create known hashes of files contained in a root directory (traversed recursively) - user@host:~$ haskdeep compute -c md5 -r myimportantfiles/ -k knownhashes.txt -i "tmp|\.log" + user@host:~$ haskdeep compute -c md5 -r myimportantfiles/ -k knownhashes.txt -e "tmp|\.log" 2. verify a copy of the same files comparing them with known hashes - user@host:~$ haskdeep audit -c md5 -r copyofmyimportantfiles/ -k knownhashes.txt -i "tmp|\.log" + user@host:~$ haskdeep audit -c md5 -r copyofmyimportantfiles/ -k knownhashes.txt -e "tmp|\.log" Licensing --------- diff --git a/haskdeep.cabal b/haskdeep.cabal --- a/haskdeep.cabal +++ b/haskdeep.cabal @@ -1,10 +1,10 @@ name: haskdeep -version: 0.1.0.0 +version: 0.2.0.0 cabal-version: >=1.8 build-type: Simple license: BSD3 license-file: LICENSE -copyright: © 2012 Mauro Taraborelli +copyright: © 2013 Mauro Taraborelli author: Mauro Taraborelli maintainer: maurotaraborelli@gmail.com stability: experimental @@ -22,15 +22,18 @@ . > user@host:~$ haskdeep > - > Usage: haskdeep COMMAND [-c|--computation MODE] [-r|--root DIRNAME] [-k|--known FILENAME] [-i|--ignore RULE] + > Usage: haskdeep COMMAND [-c|--computation MODE] [-r|--root DIRNAME] [-k|--known FILENAME] > Computes hashes and audit a set of files > > Available options: + > -h,--help Show this help text + > -v,--version Show version information > -c,--computation MODE md5 | sha1 | sha256 | skein512 - default md5 > -r,--root DIRNAME Root directory - default current directory > -k,--known FILENAME Known hashes file - default known.haskdeep - > -i,--ignore RULE Regex to ignore files or directories - > -h,--help Show this help text + > -e,--excl-regex REGEX Exclude files or directories based on regex + > -f,--incl-mod-from DATE Include files modified from yyyy-mm-ddThh:mm:ssZ + > -t,--incl-mod-upto DATE Include files modified up to yyyy-mm-ddThh:mm:ssZ > > Available commands: > compute Computes file hashes and saves them to known hashes file @@ -40,16 +43,19 @@ . 1. create known hashes of files contained in a root directory (traversed recursively) . - > user@host:~$ haskdeep compute -c md5 -r myimportantfiles/ -k knownhashes.txt -i "tmp|\.log" + > user@host:~$ haskdeep compute -c md5 -r myimportantfiles/ -k knownhashes.txt -e "tmp|\.log" . 2. verify a copy of the same files comparing them with known hashes . - > user@host:~$ haskdeep audit -c md5 -r copyofmyimportantfiles/ -k knownhashes.txt -i "tmp|\.log" + > user@host:~$ haskdeep audit -c md5 -r copyofmyimportantfiles/ -k knownhashes.txt -e "tmp|\.log" . Heavily inspired by @hashdeep@: category: Console, Cryptography, System extra-source-files: README.md + CHANGELOG.md +homepage: https://github.com/maurotrb/haskdeep +bug-reports: https://github.com/maurotrb/haskdeep/issues executable haskdeep hs-source-dirs: src @@ -65,24 +71,32 @@ HaskDeep.KnownHash.Writer build-depends: attoparsec ==0.10.* - , attoparsec-conduit ==0.5.* - , base ==4.5.* + , attoparsec-conduit ==1.0.* + , base ==4.* + , base16-bytestring ==0.1.* , bytestring ==0.9.* , cereal ==0.3.* - , conduit ==0.5.* + , conduit ==1.0.* , containers ==0.4.* - , cryptohash ==0.7.* - , crypto-api ==0.10.* - , crypto-conduit ==0.4.* - , filesystem-conduit ==0.5.* - , optparse-applicative ==0.4.* + , cryptohash ==0.10.* + , cryptohash-cryptoapi ==0.1.* + , crypto-api ==0.12.* + , crypto-conduit ==0.5.* + , filesystem-conduit ==1.0.* + , old-locale ==1.0.* + , optparse-applicative ==0.5.* , system-fileio ==0.3.* , system-filepath ==0.4.* , text ==0.11.* + , time ==1.4.* , regex-tdfa ==1.1.* , regex-tdfa-text ==1.0.* ghc-options: -Wall source-repository head type: git - location: git://github.com/maurotrb/haskdeep.git + location: https://github.com/maurotrb/haskdeep.git + +source-repository head + type: git + location: https://maurotrb@bitbucket.org/maurotrb/haskdeep.git diff --git a/src/HaskDeep/Computation.hs b/src/HaskDeep/Computation.hs --- a/src/HaskDeep/Computation.hs +++ b/src/HaskDeep/Computation.hs @@ -32,6 +32,7 @@ import qualified Data.Conduit.Filesystem as CF import qualified Data.Conduit.List as CL import qualified Data.Text as T +import Data.Time (UTCTime) import qualified Filesystem as FS import Filesystem.Path.CurrentOS (FilePath) import qualified Filesystem.Path.CurrentOS as FSC @@ -50,12 +51,15 @@ compute conf cm = CF.traverse False root $$ CL.foldM insert_hash empty_with_symbol where root = rootDirectory conf - rule = ignoreRule conf + regex = excludeRegex conf + mod_from = includeModFrom conf + mod_upto = includeModUpTo conf empty_with_symbol = HS.setSymbol (symbol cm) HS.empty insert_hash :: HashSet -> FilePath -> IO HashSet insert_hash hs fp = do let fpt = fpToText $ relativize root fp - if ignore fpt rule + fmt <- FS.getModified fp + if (excRegex fpt regex) || (excMod fmt mod_from mod_upto) then return hs else do s <- FS.getSize fp h <- liftM (runComputation cm) $ C.runResourceT @@ -68,6 +72,12 @@ fpToText :: FilePath -> T.Text fpToText = either id id . FSC.toText -ignore :: T.Text -> Maybe T.Text -> Bool -ignore fpt (Just rule) = fpt =~ rule -ignore _ Nothing = False +excRegex :: T.Text -> Maybe T.Text -> Bool +excRegex fpt (Just rule) = fpt =~ rule +excRegex _ Nothing = False + +excMod :: UTCTime -> Maybe UTCTime -> Maybe UTCTime -> Bool +excMod fmt (Just modFrom) (Just modUpTo) = (fmt < modFrom) || (fmt > modUpTo) +excMod fmt (Just modFrom) Nothing = fmt < modFrom +excMod fmt Nothing (Just modUpTo) = fmt > modUpTo +excMod _ Nothing Nothing = False diff --git a/src/HaskDeep/ComputationMode.hs b/src/HaskDeep/ComputationMode.hs --- a/src/HaskDeep/ComputationMode.hs +++ b/src/HaskDeep/ComputationMode.hs @@ -24,14 +24,9 @@ ) where -import Data.Bits ((.&.), shiftR) - -import Crypto.Hash.MD5 (MD5) -import Crypto.Hash.SHA1 (SHA1) -import Crypto.Hash.SHA256 (SHA256) -import Crypto.Hash.Skein512 (Skein512) +import Crypto.Hash.CryptoAPI (MD5, SHA1, SHA256, Skein512_512) import Data.ByteString (ByteString) -import qualified Data.ByteString as BS +import qualified Data.ByteString.Base16 as BB16 import qualified Data.Serialize as S import qualified Data.Text as T @@ -43,41 +38,16 @@ -- | MD5 computation. md5hash :: ComputationMode MD5 -md5hash = ComputationMode "md5" (toHex . S.encode) +md5hash = ComputationMode "md5" (BB16.encode . S.encode) -- | SHA1 computation. sha1hash :: ComputationMode SHA1 -sha1hash = ComputationMode "sha1" (toHex . S.encode) +sha1hash = ComputationMode "sha1" (BB16.encode . S.encode) -- | SHA256 computation. sha256hash :: ComputationMode SHA256 -sha256hash = ComputationMode "sha256" (toHex . S.encode) +sha256hash = ComputationMode "sha256" (BB16.encode . S.encode) -- | Skein512 computation. -skein512hash :: ComputationMode Skein512 -skein512hash = ComputationMode "skein512" (toHex . S.encode) - --- Only with ByteString 0.10 ---toHex :: ByteString -> ByteString ---toHex = BS.concat . BL.toChunks . BB.toLazyByteString . BA.byteStringHexFixed - --- Taken from --- http://stackoverflow.com/questions/10099921/efficiently-turn-a-bytestring-into-a-hex-representation -toHex :: ByteString -> ByteString -toHex bs0 = - fst $ BS.unfoldrN (BS.length bs0 * 2) go (Left bs0) - where - go (Left bs) = - case BS.uncons bs of - Nothing -> Nothing - Just (w, bs') -> - let w1 = w `shiftR` 4 - w2 = w .&. 15 - c1 = toC w1 - c2 = toC w2 - in Just (c1, Right (c2, bs')) - go (Right (c, bs)) = Just (c, Left bs) - - toC w - | w < 10 = w + 48 - | otherwise = w + 87 +skein512hash :: ComputationMode Skein512_512 +skein512hash = ComputationMode "skein512" (BB16.encode . S.encode) diff --git a/src/HaskDeep/Configuration.hs b/src/HaskDeep/Configuration.hs --- a/src/HaskDeep/Configuration.hs +++ b/src/HaskDeep/Configuration.hs @@ -23,14 +23,17 @@ import Prelude hiding (FilePath) import Data.Text (Text) +import Data.Time (UTCTime) import Filesystem.Path (FilePath) import qualified Filesystem.Path.CurrentOS as FSC -- | HaskDeep configuration. data HaskDeepConfiguration = HaskDeepConfiguration - { rootDirectory :: FilePath -- ^ Root directory - , knownHashes :: FilePath -- ^ Known hashes file - , ignoreRule :: Maybe Text -- ^ Ignore rule (Regex) + { rootDirectory :: FilePath -- ^ Root directory + , knownHashes :: FilePath -- ^ Known hashes file + , excludeRegex :: Maybe Text -- ^ Exclude rule (Regex) + , includeModFrom :: Maybe UTCTime -- ^ Include file modified from + , includeModUpTo :: Maybe UTCTime -- ^ Include file modified up to } -- | HaskDeep default configuration. @@ -38,5 +41,7 @@ defaultHaskDeepConfiguration = HaskDeepConfiguration { rootDirectory = FSC.decodeString "." , knownHashes = FSC.decodeString "known.haskdeep" - , ignoreRule = Nothing + , excludeRegex = Nothing + , includeModFrom = Nothing + , includeModUpTo = Nothing } diff --git a/src/HaskDeep/KnownHash/Writer.hs b/src/HaskDeep/KnownHash/Writer.hs --- a/src/HaskDeep/KnownHash/Writer.hs +++ b/src/HaskDeep/KnownHash/Writer.hs @@ -29,10 +29,12 @@ import qualified Data.Conduit as C import qualified Data.Conduit.Filesystem as CF import qualified Data.Conduit.List as CL +import qualified Data.Time as DT import Data.Text () import qualified Data.Text.Encoding as TE import Filesystem.Path () import qualified Filesystem.Path.CurrentOS as FSC +import qualified System.Locale as SL import HaskDeep.Configuration import HaskDeep.HashSet (HashSet) @@ -44,24 +46,32 @@ -> IO () writeHashes conf hs = C.runResourceT $ CL.sourceList bs_known $$ CF.sinkFile (knownHashes conf) where - newline = "\n" :: ByteString - root = TE.encodeUtf8 $ either id id $ FSC.toText $ rootDirectory conf - ignore = case ignoreRule conf of - (Just rule) -> TE.encodeUtf8 rule - Nothing -> BS.empty - files_count = B8.pack $ show $ HS.filesCount hs - size_sum = B8.pack $ show $ HS.sizeSum hs - known_header1 = "%%%% HASHDEEP-1.0" - `BS.append` newline - known_header2 = "%%%% size," - `BS.append` TE.encodeUtf8 (HS.compSymbol hs) - `BS.append` ",filename" - `BS.append` newline - bs_comments = map (flip BS.append newline . BS.append "## ") - [ "Root directory: " `BS.append` root - , "Ignore rule : " `BS.append` ignore - , "Files count : " `BS.append` files_count - , "Files size : " `BS.append` size_sum - , "" ] - bs_hash_set = map (flip BS.append newline . HS.toByteString) $ HS.toAscList hs - bs_known = [known_header1, known_header2] ++ bs_comments ++ bs_hash_set + newline = "\n" :: ByteString + root = TE.encodeUtf8 $ either id id $ FSC.toText $ rootDirectory conf + exclude_regex = case excludeRegex conf of + (Just regex) -> TE.encodeUtf8 regex + Nothing -> BS.empty + include_mod_from = case includeModFrom conf of + (Just modFrom) -> B8.pack $ DT.formatTime SL.defaultTimeLocale "%FT%TZ" modFrom + Nothing -> BS.empty + include_mod_upto = case includeModUpTo conf of + (Just modUpTo) -> B8.pack $ DT.formatTime SL.defaultTimeLocale "%FT%TZ" modUpTo + Nothing -> BS.empty + files_count = B8.pack $ show $ HS.filesCount hs + size_sum = B8.pack $ show $ HS.sizeSum hs + known_header1 = "%%%% HASHDEEP-1.0" + `BS.append` newline + known_header2 = "%%%% size," + `BS.append` TE.encodeUtf8 (HS.compSymbol hs) + `BS.append` ",filename" + `BS.append` newline + bs_comments = map (flip BS.append newline . BS.append "## ") + [ "Root directory : " `BS.append` root + , "Exclude regex : " `BS.append` exclude_regex + , "Include mod. from : " `BS.append` include_mod_from + , "Include mod. upto : " `BS.append` include_mod_upto + , "Files count : " `BS.append` files_count + , "Files size : " `BS.append` size_sum + , "" ] + bs_hash_set = map (flip BS.append newline . HS.toByteString) $ HS.toAscList hs + bs_known = [known_header1, known_header2] ++ bs_comments ++ bs_hash_set diff --git a/src/Main.hs b/src/Main.hs --- a/src/Main.hs +++ b/src/Main.hs @@ -25,9 +25,10 @@ import Options main :: IO () -main = OA.execParser optionsPI >>= haskdeep +main = OA.customExecParser (OA.prefs OA.showHelpOnError) optionsPI >>= haskdeep haskdeep :: Options -> IO () +haskdeep Version = putStrLn haskdeepVersion haskdeep (Options OptComputation OptMD5 conf) = execComputation conf md5hash haskdeep (Options OptComputation OptSHA1 conf) = execComputation conf sha1hash haskdeep (Options OptComputation OptSHA256 conf) = execComputation conf sha256hash diff --git a/src/Options.hs b/src/Options.hs --- a/src/Options.hs +++ b/src/Options.hs @@ -16,6 +16,7 @@ ,OptExecution(..) ,OptCompMode(..) ,optionsPI + ,haskdeepVersion ) where @@ -23,14 +24,21 @@ import Data.Text (Text) import qualified Data.Text as T +import Data.Time (UTCTime) +import qualified Data.Time as DT import Filesystem.Path.CurrentOS (FilePath) import qualified Filesystem.Path.CurrentOS as FSC import Options.Applicative +import qualified System.Locale as SL import HaskDeep -data Options = Options OptExecution OptCompMode HaskDeepConfiguration +haskdeepVersion :: String +haskdeepVersion = "haskdeep 0.2.0.0 - file hashing and audit" +data Options = Version + | Options OptExecution OptCompMode HaskDeepConfiguration + data OptExecution = OptComputation | OptAudit @@ -40,62 +48,91 @@ | OptSkein512 optionsPI :: ParserInfo Options -optionsPI = info (optionsP <**> helper) +optionsPI = info (helper <*> optionsP) ( fullDesc - & progDesc "Computes hashes and audit a set of files" - & header "haskdeep - file hashing and audit" ) + <> progDesc "Computes hashes and audit a set of files" + <> header haskdeepVersion) configurationP :: Parser HaskDeepConfiguration configurationP = HaskDeepConfiguration <$> nullOption ( long "root" - & short 'r' - & metavar "DIRNAME" - & help "Root directory - default current directory" - & reader fpReader - & value (rootDirectory defaultHaskDeepConfiguration)) + <> short 'r' + <> metavar "DIRNAME" + <> help "Root directory - default current directory" + <> reader fpReader + <> value (rootDirectory defaultHaskDeepConfiguration)) <*> nullOption ( long "known" - & short 'k' - & metavar "FILENAME" - & help "Known hashes file - default known.haskdeep" - & reader fpReader - & value (knownHashes defaultHaskDeepConfiguration)) + <> short 'k' + <> metavar "FILENAME" + <> help "Known hashes file - default known.haskdeep" + <> reader fpReader + <> value (knownHashes defaultHaskDeepConfiguration)) <*> nullOption - ( long "ignore" - & short 'i' - & metavar "RULE" - & help "Regex to ignore files or directories" - & reader ignReader - & value (ignoreRule defaultHaskDeepConfiguration)) + ( long "excl-regex" + <> short 'e' + <> metavar "REGEX" + <> help "Exclude files or directories based on regex" + <> reader regexReader + <> value (excludeRegex defaultHaskDeepConfiguration) + <> hidden ) + <*> nullOption + ( long "incl-mod-from" + <> short 'f' + <> metavar "DATE" + <> help "Include files modified from yyyy-mm-ddThh:mm:ssZ" + <> reader timeReader + <> value (includeModFrom defaultHaskDeepConfiguration) + <> hidden ) + <*> nullOption + ( long "incl-mod-upto" + <> short 't' + <> metavar "DATE" + <> help "Include files modified up to yyyy-mm-ddThh:mm:ssZ" + <> reader timeReader + <> value (includeModUpTo defaultHaskDeepConfiguration) + <> hidden ) optionsP :: Parser Options -optionsP = Options - <$> subparser - ( command "compute" - (info (pure OptComputation) - (progDesc "Computes file hashes and saves them to known hashes file")) - & command "audit" - (info (pure OptAudit) - (progDesc "Audit files comparing them to known hashes"))) - <*> nullOption - ( long "computation" - & short 'c' - & metavar "MODE" - & help "md5 | sha1 | sha256 | skein512 - default md5" - & reader compReader - & value OptMD5) - <*> configurationP +optionsP = flag' Version + ( long "version" + <> short 'v' + <> help "Show version information" + <> hidden ) + <|> ( Options + <$> subparser + ( command "compute" + (info (pure OptComputation) + (progDesc "Computes file hashes and saves them to known hashes file")) + <> command "audit" + (info (pure OptAudit) + (progDesc "Audit files comparing them to known hashes"))) + <*> nullOption + ( long "computation" + <> short 'c' + <> metavar "MODE" + <> help "md5 | sha1 | sha256 | skein512 - default md5" + <> reader compReader + <> value OptMD5) + <*> configurationP ) -fpReader :: String -> Maybe FilePath -fpReader fp = Just $ FSC.decodeString fp +compReader :: String -> Either ParseError OptCompMode +compReader "md5" = Right OptMD5 +compReader "sha1" = Right OptSHA1 +compReader "sha256" = Right OptSHA256 +compReader "skein512" = Right OptSkein512 +compReader _ = Left ShowHelpText -ignReader :: String -> Maybe (Maybe Text) -ignReader = Just . Just . T.pack +fpReader :: String -> Either ParseError FilePath +fpReader fp = Right $ FSC.decodeString fp -compReader :: String -> Maybe OptCompMode -compReader "md5" = Just OptMD5 -compReader "sha1" = Just OptSHA1 -compReader "sha256" = Just OptSHA256 -compReader "skein512" = Just OptSkein512 -compReader _ = Nothing +regexReader :: String -> Either ParseError (Maybe Text) +regexReader = Right . Just . T.pack + +timeReader :: String -> Either ParseError (Maybe UTCTime) +timeReader dt = case parsedTime of + (Just _) -> Right parsedTime + Nothing -> Left ShowHelpText + where + parsedTime = DT.parseTime SL.defaultTimeLocale "%FT%TZ" dt