X-Git-Url: https://gitweb.michael.orlitzky.com/?a=blobdiff_plain;f=src%2FMain.hs;h=1cc4c73aa47d963e56f30a7188f05781f010b76c;hb=2e8aab2a69a8b9887e4f037115b80a2063d3a6b7;hp=fc1b1b18eba02c44d549de693cc8f3d65fd6ea1e;hpb=031638c08a4ce9d7ea156ead71cb8ef02e23fa81;p=email-validator.git diff --git a/src/Main.hs b/src/Main.hs index fc1b1b1..1cc4c73 100644 --- a/src/Main.hs +++ b/src/Main.hs @@ -1,52 +1,55 @@ {-# LANGUAGE DoAndIfThenElse #-} {-# LANGUAGE RecordWildCards #-} -module Main +module Main (main) where -import Control.Concurrent.ParallelIO.Global (parallel, stopGlobalPool) -import Control.Monad (unless) -import qualified Data.ByteString as BS -import qualified Data.ByteString.UTF8 as BSU +import Control.Concurrent.ParallelIO.Global ( + parallelInterleaved, + stopGlobalPool ) +import qualified Data.ByteString.Char8 as BS ( + hGetContents, + hPutStrLn, + lines, + null, + pack ) import Network.DNS ( Domain, Resolver, - ResolvConf(..), + ResolvConf( resolvTimeout ), defaultResolvConf, makeResolvSeed, - withResolver) -import Network.DNS.Lookup (lookupA, lookupMX) -import System.Directory (doesFileExist) -import System.Exit (exitWith, ExitCode(..)) + withResolver ) +import Network.DNS.Lookup ( lookupA, lookupMX ) import System.IO ( - Handle, - IOMode( WriteMode ), - hClose, hFlush, - openFile, stdin, - stdout) + stdout ) -import CommandLine (Args(..), apply_args) -import EmailAddress -import ExitCodes (exit_input_file_doesnt_exist) +import CommandLine ( + Args( Args, accept_a, rfc5322 ), + get_args ) +import EmailAddress( + Address, + parts, + validate_syntax ) --- | Resolver parameters. We increase the default timeout from 3 to 5 +-- | Resolver parameters. We increase the default timeout from 3 to 10 -- seconds. resolv_conf :: ResolvConf -resolv_conf = defaultResolvConf { resolvTimeout = 5 * 1000 * 1000 } +resolv_conf = defaultResolvConf { resolvTimeout = 10 * 1000 * 1000 } -- | A list of common domains, there's no need to waste MX lookups -- on these. common_domains :: [Domain] -common_domains = map BSU.fromString [ "aol.com", - "comcast.net", - "gmail.com", - "msn.com", - "yahoo.com", - "verizon.net" ] +common_domains = map BS.pack [ "aol.com", + "comcast.net", + "gmail.com", + "msn.com", + "yahoo.com", + "verizon.net" ] -- | Check whether the given domain has a valid MX record. @@ -56,8 +59,9 @@ validate_mx resolver domain | otherwise = do result <- lookupMX resolver domain case result of - Nothing -> return False - _ -> return True + -- A list of one or more elements? + Right (_:_) -> return True + _ -> return False -- | Check whether the given domain has a valid A record. @@ -67,16 +71,16 @@ validate_a resolver domain | otherwise = do result <- lookupA resolver domain case result of - Nothing -> return False - _ -> return True + Right (_:_) -> return True + _ -> return False -- | Validate an email address by doing some simple syntax checks and -- (if those fail) an MX lookup. We don't count an A record as a mail -- exchanger. -validate :: Resolver -> Bool -> Address -> IO (Address, Bool) -validate resolver accept_a address = do - let valid_syntax = validate_syntax address +validate :: Resolver -> Bool -> Bool -> Address -> IO (Address, Bool) +validate resolver accept_a rfc5322 address = do + let valid_syntax = validate_syntax rfc5322 address if valid_syntax then do let (_,domain) = parts address mx_result <- validate_mx resolver domain @@ -93,52 +97,35 @@ validate resolver accept_a address = do return (address, False) --- | Append a ByteString to a file Handle, followed by a newline. -append_handle_with_newline :: Handle -> BS.ByteString -> IO () -append_handle_with_newline h bs = do - BS.hPutStr h bs - BS.hPutStr h newline - where - newline = BSU.fromString "\n" - main :: IO () main = do - Args{..} <- apply_args - - -- Get the input from either stdin, or the file given on the command - -- line. - input <- case input_file of - Nothing -> BS.hGetContents stdin - Just path -> do - is_file <- doesFileExist path - unless is_file $ - exitWith (ExitFailure exit_input_file_doesnt_exist) - BS.readFile path - - -- Do the same for the output handle and stdout. - output_handle <- case output_file of - Nothing -> return stdout - Just path -> openFile path WriteMode - - -- Split the input into lines. - let addresses = BSU.lines input + Args{..} <- get_args + + -- Split stdin into lines, which should result in a list of addresses. + input <- BS.hGetContents stdin + let addresses = BS.lines input -- And remove the empty ones. let nonempty_addresses = filter (not . BS.null) addresses rs <- makeResolvSeed resolv_conf - withResolver rs $ \resolver -> do - -- Construst a list of [IO (Address, Bool)] - let actions = map (validate resolver accept_a) nonempty_addresses - -- And compute them in parallel. - results <- parallel actions - stopGlobalPool - -- Find the pairs with a True in the second position. - let good_pairs = filter snd results - -- And output the results. - mapM_ ((append_handle_with_newline output_handle) . fst) good_pairs - - -- Clean up. It's safe to try to close stdout. - hFlush output_handle - hClose output_handle + let validate' addr = withResolver rs $ \resolver -> + validate resolver accept_a rfc5322 addr + + -- Construct a list of [IO (Address, Bool)]. The withResolver calls + -- are the ones that should be run in parallel. + let actions = map validate' nonempty_addresses + + -- Run the lookup actions in parallel. + results <- parallelInterleaved actions + + -- Filter the bad ones. + let valid_results = filter snd results + + -- Output the results. + let valid_addresses = map fst valid_results + mapM_ (BS.hPutStrLn stdout) valid_addresses + + stopGlobalPool + hFlush stdout