Code clean-up using convertible-text.

[Lucu.git] / Network / HTTP / Lucu / Utils.hs

diff --git a/Network/HTTP/Lucu/Utils.hs b/Network/HTTP/Lucu/Utils.hs
index a488aafbb64161feee9e67404c05f29a065236ec..297ea3a762ef64992466469a3845053ffb86d7af 100644 (file)
--- a/Network/HTTP/Lucu/Utils.hs
+++ b/Network/HTTP/Lucu/Utils.hs
@@ -1,77 +1,97 @@
 {-# LANGUAGE
-    BangPatterns
+    FlexibleContexts
   , OverloadedStrings
   , UnicodeSyntax
   #-}
--- |Utility functions used internally in the Lucu httpd. These
--- functions may be useful too for something else.
+-- |Utility functions used internally in this package.
 module Network.HTTP.Lucu.Utils
-    ( splitBy
-    , joinWith
+    ( Host
+    , PathSegment
+    , Path
+
+    , splitBy
     , quoteStr
     , parseWWWFormURLEncoded
-    , splitPathInfo
-    , show3
+    , uriHost
+    , uriPathSegments
+    , trim
+
+    , (⊲)
+    , (⊳)
+    , (⋈)
+    , mapM
+
+    , getLastModified
     )
     where
-import Blaze.ByteString.Builder.ByteString as B
-import Blaze.Text.Int as BT
-import Control.Monad
+import Control.Applicative hiding (empty)
+import Control.Monad hiding (mapM)
 import Data.Ascii (Ascii, AsciiBuilder)
 import qualified Data.Ascii as A
 import Data.ByteString (ByteString)
 import qualified Data.ByteString.Char8 as BS
-import Data.List hiding (last)
+import Data.CaseInsensitive (CI)
+import qualified Data.CaseInsensitive as CI
+import Data.Char
+import Data.Collections
+import Data.Collections.BaseInstances ()
+import Data.Convertible.Base
+import Data.Convertible.Instances.Time ()
+import Data.Maybe
 import Data.Monoid.Unicode
 import Data.Text (Text)
-import Data.Text.Encoding as T
+import qualified Data.Text as T
+import Data.Time
 import Network.URI
-import Prelude hiding (last)
+import Prelude hiding (last, mapM, null, reverse)
 import Prelude.Unicode
+import System.Directory
+
+-- |'Host' represents an IP address or a host name in an URI
+-- authority.
+type Host = CI Text
+
+-- |'PathSegment' represents an URI path segment, split by slashes and
+-- percent-decoded.
+type PathSegment = ByteString
 
--- |> splitBy (== ':') "ab:c:def"
---  > ==> ["ab", "c", "def"]
+-- |'Path' is a list of URI path segments.
+type Path = [PathSegment]
+
+-- |>>> splitBy (== ':') "ab:c:def"
+-- ["ab", "c", "def"]
 splitBy ∷ (a → Bool) → [a] → [[a]]
+{-# INLINEABLE splitBy #-}
 splitBy isSep src
-    = case break isSep src
-      of (last , []       ) → [last]
-         (first, _sep:rest) → first : splitBy isSep rest
-
--- |> joinWith ":" ["ab", "c", "def"]
---  > ==> "ab:c:def"
-joinWith ∷ Ascii → [AsciiBuilder] → AsciiBuilder
-{-# INLINEABLE joinWith #-}
-joinWith sep = flip go (∅)
-    where
-      go ∷ [AsciiBuilder] → AsciiBuilder → AsciiBuilder
-      {-# INLINE go #-}
-      go []     ab = ab
-      go (x:[]) ab = ab ⊕ x
-      go (x:xs) ab = go xs (ab ⊕ A.toAsciiBuilder sep ⊕ x)
-
--- |> quoteStr "abc"
---  > ==> "\"abc\""
+    = case break isSep src of
+        (last , []       ) → [last]
+        (first, _sep:rest) → first : splitBy isSep rest
+
+-- |>>> quoteStr "abc"
+-- "\"abc\""
 --
---  > quoteStr "ab\"c"
---  > ==> "\"ab\\\"c\""
+-- >>> quoteStr "ab\"c"
+-- "\"ab\\\"c\""
 quoteStr ∷ Ascii → AsciiBuilder
 quoteStr str = A.toAsciiBuilder "\"" ⊕
                go (A.toByteString str) (∅) ⊕
                A.toAsciiBuilder "\""
     where
-      go ∷ BS.ByteString → AsciiBuilder → AsciiBuilder
+      go ∷ ByteString → AsciiBuilder → AsciiBuilder
       go bs ab
           = case BS.break (≡ '"') bs of
               (x, y)
-                  | BS.null y → ab ⊕ b2ab x
-                  | otherwise → go (BS.tail y) (ab ⊕ b2ab x
-                                                   ⊕ A.toAsciiBuilder "\\\"")
+                  | BS.null y
+                      → ab ⊕ b2ab x
+                  | otherwise
+                      → go (BS.tail y)
+                           (ab ⊕ b2ab x ⊕ A.toAsciiBuilder "\\\"")
 
-      b2ab ∷ BS.ByteString → AsciiBuilder
+      b2ab ∷ ByteString → AsciiBuilder
       b2ab = A.toAsciiBuilder ∘ A.unsafeFromByteString
 
--- |> parseWWWFormURLEncoded "aaa=bbb&ccc=ddd"
---  > ==> [("aaa", "bbb"), ("ccc", "ddd")]
+-- |>>> parseWWWFormURLEncoded "aaa=bbb&ccc=ddd"
+-- [("aaa", "bbb"), ("ccc", "ddd")]
 parseWWWFormURLEncoded ∷ Ascii → [(ByteString, ByteString)]
 parseWWWFormURLEncoded src
     -- THINKME: We could gain some performance by using attoparsec
@@ -86,28 +106,63 @@ parseWWWFormURLEncoded src
                             )
     where
       unescape ∷ String → ByteString
-      unescape = BS.pack ∘ unEscapeString ∘ map plusToSpace
+      unescape = BS.pack ∘ unEscapeString ∘ (plusToSpace <$>)
 
       plusToSpace ∷ Char → Char
       plusToSpace '+' = ' '
       plusToSpace c   = c
 
--- |> splitPathInfo "http://example.com/foo/bar"
---  > ==> ["foo", "bar"]
-splitPathInfo ∷ URI → [Text]
-splitPathInfo uri
+-- |>>> uriHost "http://example.com/foo/bar"
+-- "example.com"
+uriHost ∷ URI → Host
+{-# INLINE uriHost #-}
+uriHost = CI.mk ∘ T.pack ∘ uriRegName ∘ fromJust ∘ uriAuthority
+
+-- |>>> uriPathSegments "http://example.com/foo/bar"
+-- ["foo", "bar"]
+uriPathSegments ∷ URI → Path
+uriPathSegments uri
     = let reqPathStr = uriPath uri
-          reqPath    = [unEscapeString x | x ← splitBy (≡ '/') reqPathStr, (¬) (null x)]
+          reqPath    = [ unEscapeString x
+                         | x ← splitBy (≡ '/') reqPathStr, (¬) (null x) ]
       in
-        map (T.decodeUtf8 ∘ BS.pack) reqPath
+        BS.pack <$> reqPath
 
--- |> show3 5
---  > ==> "005"
-show3 ∷ Integral n ⇒ n → AsciiBuilder
-{-# INLINEABLE show3 #-}
-show3 = A.unsafeFromBuilder ∘ go
+-- |>>> trim "  ab c d "
+-- "ab c d"
+trim ∷ String → String
+trim = reverse ∘ f ∘ reverse ∘ f
     where
-      go i | i ≥ 0 ∧ i < 10   = B.fromByteString "00" ⊕ BT.digit    i
-           | i ≥ 0 ∧ i < 100  = B.fromByteString "0"  ⊕ BT.integral i
-           | i ≥ 0 ∧ i < 1000 =                         BT.integral i
-           | otherwise        = error ("show3: the integer i must satisfy 0 <= i < 1000: " ⧺ show i)
+      f = dropWhile isSpace
+
+infixr 5 ⊲
+-- | (&#x22B2;) = ('<|')
+--
+-- U+22B2, NORMAL SUBGROUP OF
+(⊲) ∷ Sequence α a ⇒ a → α → α
+(⊲) = (<|)
+
+infixl 5 ⊳
+-- | (&#x22B3;) = ('|>')
+--
+-- U+22B3, CONTAINS AS NORMAL SUBGROUP
+(⊳) ∷ Sequence α a ⇒ α → a → α
+(⊳) = (|>)
+
+infixr 5 ⋈
+-- | (&#x22C8;) = ('><')
+--
+-- U+22C8, BOWTIE
+(⋈) ∷ Sequence α a ⇒ α → α → α
+(⋈) = (><)
+
+-- |Generalised @mapM@ from any 'Foldable' to 'Unfoldable'. Why isn't
+-- this in the @collections-api@?
+mapM ∷ (Foldable α a, Unfoldable β b, Functor m, Monad m)
+     ⇒ (a → m b) → α → m β
+{-# INLINE mapM #-}
+mapM = flip foldrM empty ∘ (flip ((<$>) ∘ flip insert) ∘)
+
+-- |Get the modification time of a given file.
+getLastModified ∷ FilePath → IO UTCTime
+getLastModified = (cs <$>) ∘ getModificationTime