-{-# LANGUAGE ScopedTypeVariables, RecordWildCards, DoAndIfThenElse #-}
+{-# LANGUAGE DoAndIfThenElse #-}
module Main
where
+import Control.Concurrent (threadDelay)
import Control.Monad (when)
-import Data.Maybe (fromJust)
+import Data.Maybe (fromJust, isJust)
import Prelude hiding (readFile)
import System.Directory (doesFileExist)
import System.IO (
import Test.Framework (Test, testGroup)
import Test.Framework.Providers.HUnit (testCase)
import Text.Regex.Posix ((=~))
-import Text.XML.HXT.Core hiding (when)
-
+import Text.XML.HXT.Core (
+ IOStateArrow,
+ SysConfigList,
+ XmlTree,
+ no,
+ readString,
+ withParseHTML,
+ withValidate,
+ withWarnings,
+ yes
+ )
import CommandLine (show_help)
import Configuration (Cfg(..), get_cfg, use_account)
import LWN.HTTP (get_page, log_in, make_cookie_jar)
import LWN.Page (epublish, parse)
-import LWN.URI (is_lwn_url, make_absolute_url, make_https)
+import LWN.URI (add_trailing_slash, is_lwn_url, make_absolute_url, make_https)
import Misc (contains)
my_read :: String -> IOStateArrow s b XmlTree
my_read = readString my_read_opts
+
+-- |A wrapper around threadDelay which takes seconds instead of
+-- microseconds as its argument.
+thread_sleep :: Int -> IO ()
+thread_sleep seconds = do
+ let microseconds = seconds * (10 ^ (6 :: Int))
+ threadDelay microseconds
+
+
-- | Try to parse the given article using HXT. We try a few different
-- methods; if none of them work, we return 'Nothing'.
get_xml_from_article :: Cfg -> IO (Maybe (IOStateArrow s b XmlTree))
li_result <- log_in cj
(fromJust $ username cfg)
(fromJust $ password cfg)
- when (not li_result) $ do
- hPutStrLn stderr "Failed to log in."
+
+ -- Without this, the cookie file is empty during
+ -- get_page. Whaaat?
+ thread_sleep 1
+
+ when (isJust li_result) $ do
+ let msg = "Failed to log in. " ++ (fromJust li_result)
+ hPutStrLn stderr msg
html <- get_page (Just cj) my_article
- return $
- case html of
- Nothing -> Nothing
- Just h -> Just $ my_read h
+
+ case html of
+ Left err -> do
+ let msg = "Failed to retrieve page. " ++ err
+ hPutStrLn stderr msg
+ return Nothing
+ Right h -> return $ Just $ my_read h
else do
html <- get_page Nothing my_article
- return $
- case html of
- Nothing -> Nothing
- Just h -> Just $ my_read h
+ case html of
+ Left err -> do
+ let msg = "Failed to retrieve page. " ++ err
+ hPutStrLn stderr msg
+ return Nothing
+ Right h -> return $ Just $ my_read h
-- | If we're given an empty path, return a handle to
-- 'stdout'. Otherwise, open the given file and return a read/write
real_article_path :: String -> IO String
real_article_path s = do
is_file <- doesFileExist s
- return $ if is_file then s else check_cases
+ return $ if is_file then s else add_trailing_slash check_cases
where
abs_current =
case make_absolute_url "current" of
cfg <- get_cfg
output_handle <- get_output_handle (output cfg)
+ when (use_account cfg) $ do
+ putStrLn "Using account."
+
maybe_html <- get_xml_from_article cfg
case maybe_html of
test_current_article_path :: Assertion
test_current_article_path = do
- let expected = "https://lwn.net/current"
+ let expected = "https://lwn.net/current/"
actual <- real_article_path "current"
assertEqual "Current article path constructed" expected actual
test_numbered_article_path :: Assertion
test_numbered_article_path = do
- let expected = "https://lwn.net/Articles/69"
+ let expected = "https://lwn.net/Articles/69/"
actual <- real_article_path "69" -- I'm twelve
assertEqual "Numbered article path constructed" expected actual