Move the pure-xml functions into the LWN.XHTML module.

[dead/lwn-epub.git] / src / LWN / HTTP.hs

diff --git a/src/LWN/HTTP.hs b/src/LWN/HTTP.hs
index 4cfcb04bfbc389c6c75f539516249ee91145b882..7da6ed1488d97114247091476c26b57ace5cc7d8 100644 (file)
--- a/src/LWN/HTTP.hs
+++ b/src/LWN/HTTP.hs
@@ -5,6 +5,7 @@ where
 
 import qualified Data.ByteString as B (hPut)
 import qualified Data.Map as Map (Map, empty, insert)
+import Data.Maybe (fromJust)
 import Network.Curl (
   CurlCode(..),
   CurlOption(..),
@@ -18,8 +19,11 @@ import Network.Curl (
 import Network.Curl.Download (openURI)
 import System.Directory (doesFileExist, getTemporaryDirectory)
 import System.IO (hClose, hPutStrLn, stderr)
+import qualified System.IO.UTF8 as Utf8 (readFile)
 import System.IO.Temp (openBinaryTempFile, openTempFile)
 
+import qualified Configuration as C (Cfg, password, use_account, username)
+import LWN.Article (real_article_path)
 import LWN.URI (URL, filename)
 
 login_url :: URL
@@ -182,3 +186,47 @@ download_image_urls image_urls = do
     my_insert :: ImageMap -> (URL, Maybe FilePath) -> ImageMap
     my_insert dict (_, Nothing)  = dict
     my_insert dict (k, Just v) = Map.insert k v dict
+
+
+-- | Try to parse the given article using HXT. We try a few different
+--   methods; if none of them work, we return 'Nothing'.
+get_article_contents :: C.Cfg -> URL -> IO (Maybe String)
+get_article_contents cfg article_name = do
+  my_article <- real_article_path article_name
+  is_file <- doesFileExist my_article
+  case is_file of
+    True -> do
+      contents <- Utf8.readFile my_article
+      return $ Just $ contents
+    False -> do
+      -- Download the URL and try to parse it.
+      if C.use_account cfg then do
+        -- use_account would be false if these fromJusts would fail.
+        cj <- make_cookie_jar
+        li_result <- log_in cj
+                      (fromJust $ C.username cfg)
+                      (fromJust $ C.password cfg)
+
+        case li_result of
+          Left err -> do
+            let msg = "Failed to log in. " ++ err
+            hPutStrLn stderr msg
+          Right response_body -> do
+            hPutStrLn stderr response_body
+
+        html <- get_page (Just cj) my_article
+
+        case html of
+          Left err -> do
+            let msg = "Failed to retrieve page. " ++ err
+            hPutStrLn stderr msg
+            return Nothing
+          Right h -> return $ Just h
+      else do
+        html <- get_page Nothing my_article
+        case html of
+          Left err -> do
+            let msg = "Failed to retrieve page. " ++ err
+            hPutStrLn stderr msg
+            return Nothing
+          Right h -> return $ Just h