html-parse 0.1.0.0 → 0.2.0.0

raw patch · 3 files changed

+39/−11 lines, 3 files

Files

• Benchmark.hs +2/−2
• Text/HTML/Parser.hs +36/−8
• html-parse.cabal +1/−1

@@ -11,12 +11,12 @@         [ bgroup "Forced"             [ bench "tagsoup fast Text" $ nf (Soup.parseTagsOptions Soup.parseOptionsFast) t             , bench "tagsoup normal Text" $ nf (Soup.parseTagsOptions Soup.parseOptions) t-            , bench "html-parser" $ nf Me.tagStream t+            , bench "html-parser" $ nf Me.parseTokens t             ]         , bgroup "length"             [ bench "tagsoup fast Text" $ whnf (length . Soup.parseTagsOptions Soup.parseOptionsFast) t             , bench "tagsoup normal Text" $ whnf (length . Soup.parseTagsOptions Soup.parseOptions) t-            , bench "html-parser" $ whnf (length . Me.tagStream) t+            , bench "html-parser" $ whnf (length . Me.parseTokens) t             ]         ] 

@@ -6,11 +6,14 @@ -- so it behaves reasonable well on ill-formed documents from the open Web. -- module Text.HTML.Parser-    ( TagName, AttrName, AttrValue+    ( -- * Parsing+      parseTokens+    , parseTokensLazy+    , token+      -- * Types     , Token(..)+    , TagName, AttrName, AttrValue     , Attr(..)-    , token-    , tagStream     ) where  import Data.Char hiding (isSpace)@@ -22,25 +25,33 @@ import Control.DeepSeq  import Data.Attoparsec.Text+import qualified Data.Attoparsec.Text.Lazy as AL import Data.Text (Text) import qualified Data.Text as T+import qualified Data.Text.Lazy as TL import Data.Text.Lazy.Builder (Builder) import qualified Data.Text.Lazy.Builder as B import Prelude hiding (take, takeWhile) +-- | A tag name (e.g. @body@) type TagName   = Text++-- | An attribute name (e.g. @href@) type AttrName  = Text++-- | The value of an attribute type AttrValue = Text  -- | An HTML token data Token-  -- | An opening tag. Attribute ordering is arbitrary+  -- | An opening tag. Attribute ordering is arbitrary.   = TagOpen !TagName [Attr]   -- | A closing tag.   | TagClose !TagName   -- | The content between tags.-  | ContentChar !Char   | ContentText !Text+  -- | A single character of content+  | ContentChar !Char   -- | Contents of a comment.   | Comment !Builder   -- | Doctype@@ -195,6 +206,7 @@ commentStart = do           (char '-' >> commentStartDash)       <|> (char '>' >> return (Comment mempty))+      <|> comment mempty  -- | /§8.2.4.47/: Comment start dash state commentStartDash :: Parser Token@@ -238,9 +250,9 @@ bogusComment :: Parser Token bogusComment = fail "Bogus comment" --- | Produce a lazy list of tokens.-tagStream :: Text -> [Token]-tagStream = unfoldr f+-- | Parse a lazy list of tokens from strict 'Text'.+parseTokens :: Text -> [Token]+parseTokens = unfoldr f   where     f :: Text -> Maybe (Token, Text)     f t@@ -248,4 +260,20 @@       | otherwise =         case parse token t of             Done rest tok -> Just (tok, rest)+            Partial cont  ->+                case cont mempty of+                  Done rest tok -> Just (tok, rest)+                  _             -> Nothing             _             -> Nothing++-- | Parse a lazy list of tokens from lazy 'TL.Text'.+parseTokensLazy :: TL.Text -> [Token]+parseTokensLazy = unfoldr f+  where+    f :: TL.Text -> Maybe (Token, TL.Text)+    f t+      | TL.null t = Nothing+      | otherwise =+        case AL.parse token t of+            AL.Done rest tok -> Just (tok, rest)+            _                -> Nothing

@@ -1,5 +1,5 @@ name:                html-parse-version:             0.1.0.0+version:             0.2.0.0 synopsis:            A high-performance HTML tokenizer description:     This package provides a fast and reasonably robust HTML5 tokenizer built