diff options
Diffstat (limited to 'src/Text/Pandoc/Readers/HTML.hs')
| -rw-r--r-- | src/Text/Pandoc/Readers/HTML.hs | 21 |
1 files changed, 14 insertions, 7 deletions
diff --git a/src/Text/Pandoc/Readers/HTML.hs b/src/Text/Pandoc/Readers/HTML.hs index 02ff07e73..52358e51e 100644 --- a/src/Text/Pandoc/Readers/HTML.hs +++ b/src/Text/Pandoc/Readers/HTML.hs @@ -1,4 +1,5 @@ -{-# LANGUAGE FlexibleContexts, FlexibleInstances, MultiParamTypeClasses #-} +{-# LANGUAGE FlexibleContexts, FlexibleInstances, MultiParamTypeClasses, +ViewPatterns#-} {- Copyright (C) 2006-2014 John MacFarlane <jgm@berkeley.edu> @@ -43,7 +44,7 @@ import Text.Pandoc.Definition import qualified Text.Pandoc.Builder as B import Text.Pandoc.Builder (Blocks, Inlines, trimInlines, HasMeta(..)) import Text.Pandoc.Shared ( extractSpaces, renderTags' - , escapeURI, safeRead ) + , escapeURI, safeRead, mapLeft ) import Text.Pandoc.Options (ReaderOptions(readerParseRaw, readerTrace) , Extension (Ext_epub_html_exts, Ext_native_divs, Ext_native_spans)) @@ -62,15 +63,18 @@ import Text.TeXMath (readMathML, writeTeX) import Data.Default (Default (..), def) import Control.Monad.Reader (Reader,ask, asks, local, runReader) +import Text.Pandoc.Error + +import Text.Parsec.Error + -- | Convert HTML-formatted string to 'Pandoc' document. readHtml :: ReaderOptions -- ^ Reader options -> String -- ^ String to parse (assumes @'\n'@ line endings) - -> Pandoc + -> Either PandocError Pandoc readHtml opts inp = - case flip runReader def $ runParserT parseDoc (HTMLState def{ stateOptions = opts } []) "source" tags of - Left err' -> error $ "\nError at " ++ show err' - Right result -> result + mapLeft (ParseFailure . getError) . flip runReader def $ + runParserT parseDoc (HTMLState def{ stateOptions = opts } []) "source" tags where tags = stripPrefixes . canonicalizeTags $ parseTagsOptions parseOptions{ optTagPosition = True } inp parseDoc = do @@ -78,6 +82,9 @@ readHtml opts inp = meta <- stateMeta . parserState <$> getState bs' <- replaceNotes (B.toList blocks) return $ Pandoc meta bs' + getError (errorMessages -> ms) = case ms of + [] -> "" + (m:_) -> messageString m replaceNotes :: [Block] -> TagParser [Block] replaceNotes = walkM replaceNotes' @@ -880,7 +887,7 @@ htmlTag :: Monad m => (Tag String -> Bool) -> ParserT [Char] st m (Tag String, String) htmlTag f = try $ do - lookAhead $ char '<' >> (oneOf "/!?" <|> letter) + lookAhead $ char '<' >> ((oneOf "/!?" >> nonspaceChar) <|> letter) (next : _) <- getInput >>= return . canonicalizeTags . parseTags guard $ f next -- advance the parser |