2f8f8f0 · typeable/pandoc

git.iko.soy/typeable/pandoc

Issue warning for duplicate header identifiers.

As noted in the previous commit, an autogenerated identifier may still coincide with an explicit identifier that is given for a header later in the document, or with an identifier on a div, span, link, or image. This commit adds a warning in this case, so users can supply an explicit identifier. * Added `DuplicateIdentifier` to LogMessage. * Modified HTML, Org, MediaWiki readers so their custom state type is an instance of HasLogMessages. This is necessary for `registerHeader` to issue warnings. See #1745.

commit2f8f8f0da64388fa01d83fccf3cf1f2899c64269 authorJohn MacFarlane <jgm@berkeley.edu> dateSun 12 Mar 2017, 22:03 (10 years ago) parentc8b9062

6 files changed+46−8

compare with master

unifiedsplit

src/Text/Pandoc/Logging.hs+10−1 src/Text/Pandoc/Parsing.hs+8−2 src/Text/Pandoc/Readers/HTML.hs+11−5 src/Text/Pandoc/Readers/MediaWiki.hs+7 src/Text/Pandoc/Readers/Org.hs+2 src/Text/Pandoc/Readers/Org/ParserState.hs+8

src/Text/Pandoc/Logging.hs+10−1

@@ -62,6 +62,7 @@ data LogMessage =

6262 | CouldNotParseYamlMetadata String SourcePos

6363 | DuplicateLinkReference String SourcePos

6464 | DuplicateNoteReference String SourcePos

65+ | DuplicateIdentifier String SourcePos

6566 | ReferenceNotFound String SourcePos

6667 | CircularReference String SourcePos

6768 | ParsingUnescaped String SourcePos

@@ -106,6 +107,11 @@ instance ToJSON LogMessage where

106107 "source" .= Text.pack (sourceName pos),

107108 "line" .= toJSON (sourceLine pos),

108109 "column" .= toJSON (sourceColumn pos)]

110+ DuplicateIdentifier s pos ->

111+ ["contents" .= Text.pack s,

112+ "source" .= Text.pack (sourceName pos),

113+ "line" .= toJSON (sourceLine pos),

114+ "column" .= toJSON (sourceColumn pos)]

109115 ReferenceNotFound s pos ->

110116 ["contents" .= Text.pack s,

111117 "source" .= Text.pack (sourceName pos),

@@ -184,6 +190,8 @@ showLogMessage msg =

184190 "Duplicate link reference '" ++ s ++ "' at " ++ showPos pos

185191 DuplicateNoteReference s pos ->

186192 "Duplicate note reference '" ++ s ++ "' at " ++ showPos pos

193+ DuplicateIdentifier s pos ->

194+ "Duplicate identifier '" ++ s ++ "' at " ++ showPos pos

187195 ReferenceNotFound s pos ->

188196 "Reference not found for '" ++ s ++ "' at " ++ showPos pos

189197 CircularReference s pos ->

@@ -233,6 +241,7 @@ messageVerbosity msg =

233241 CouldNotParseYamlMetadata{} -> WARNING

234242 DuplicateLinkReference{} -> WARNING

235243 DuplicateNoteReference{} -> WARNING

244+ DuplicateIdentifier{} -> WARNING

236245 ReferenceNotFound{} -> WARNING

237246 CircularReference{} -> WARNING

238247 CouldNotLoadIncludeFile{} -> WARNING

@@ -249,4 +258,4 @@ messageVerbosity msg =

249258 CouldNotParseCSS{} -> WARNING

250259 Fetching{} -> INFO

251260 NoTitleElement{} -> WARNING

252− NoLangSpecified -> INFO

261+ NoLangSpecified -> INFO

src/Text/Pandoc/Parsing.hs+8−2

@@ -1112,8 +1112,11 @@ type SubstTable = M.Map Key Inlines

11121112-- with its associated identifier. If the identifier is null

11131113-- and the auto_identifers extension is set, generate a new

11141114-- unique identifier, and update the list of identifiers

1115−-- in state.

1116−registerHeader :: (Stream s m a, HasReaderOptions st, HasHeaderMap st, HasIdentifierList st)

1115+-- in state. Issue a warning if an explicit identifier

1116+-- is encountered that duplicates an earlier identifier

1117+-- (explict or automatically generated).

1118+registerHeader :: (Stream s m a, HasReaderOptions st,

1119+ HasHeaderMap st, HasLogMessages st, HasIdentifierList st)

11171120 => Attr -> Inlines -> ParserT s st m Attr

11181121registerHeader (ident,classes,kvs) header' = do

11191122 ids <- extractIdentifierList <$> getState

@@ -1131,6 +1134,9 @@ registerHeader (ident,classes,kvs) header' = do

11311134 return (id'',classes,kvs)

11321135 else do

11331136 unless (null ident) $ do

1137+ when (ident `Set.member` ids) $ do

1138+ pos <- getPosition

1139+ logMessage $ DuplicateIdentifier ident pos

11341140 updateState $ updateIdentifierList $ Set.insert ident

11351141 updateState $ updateHeaderMap $ insert' header' ident

11361142 return (ident,classes,kvs)

src/Text/Pandoc/Readers/HTML.hs+11−5

@@ -83,14 +83,15 @@ readHtml opts inp = do

8383 blocks <- (fixPlains False) . mconcat <$> manyTill block eof

8484 meta <- stateMeta . parserState <$> getState

8585 bs' <- replaceNotes (B.toList blocks)

86+ reportLogMessages

8687 return $ Pandoc meta bs'

8788 getError (errorMessages -> ms) = case ms of

8889 [] -> ""

8990 (m:_) -> messageString m

9091 result <- flip runReaderT def $

91− runParserT parseDoc

92− (HTMLState def{ stateOptions = opts } [] Nothing Set.empty M.empty)

93− "source" tags

92+ runParserT parseDoc

93+ (HTMLState def{ stateOptions = opts } [] Nothing Set.empty M.empty [])

94+ "source" tags

9495 case result of

9596 Right doc -> return doc

9697 Left err -> throwError $ PandocParseError $ getError err

@@ -110,7 +111,8 @@ data HTMLState =

110111 noteTable :: [(String, Blocks)],

111112 baseHref :: Maybe URI,

112113 identifiers :: Set.Set String,

113− headerMap :: M.Map Inlines String

114+ headerMap :: M.Map Inlines String,

115+ logMessages :: [LogMessage]

114116 }

115117

116118data HTMLLocal = HTMLLocal { quoteContext :: QuoteContext

@@ -376,7 +378,7 @@ ignore raw = do

376378 -- raw can be null for tags like <!DOCTYPE>; see paRawTag

377379 -- in this case we don't want a warning:

378380 unless (null raw) $

379− report $ SkippedContent raw pos

381+ logMessage $ SkippedContent raw pos

380382 return mempty

381383

382384pHtmlBlock :: PandocMonad m => String -> TagParser m String

@@ -1092,6 +1094,10 @@ instance HasHeaderMap HTMLState where

10921094 extractHeaderMap = headerMap

10931095 updateHeaderMap f s = s{ headerMap = f (headerMap s) }

10941096

1097+instance HasLogMessages HTMLState where

1098+ addLogMessage m s = s{ logMessages = m : logMessages s }

1099+ getLogMessages = reverse . logMessages

1100+

10951101-- This signature should be more general

10961102-- MonadReader HTMLLocal m => HasQuoteContext st m

10971103instance PandocMonad m => HasQuoteContext HTMLState (ReaderT HTMLLocal m) where

src/Text/Pandoc/Readers/MediaWiki.hs+7

@@ -73,6 +73,7 @@ readMediaWiki opts s = do

7373 , mwCategoryLinks = []

7474 , mwHeaderMap = M.empty

7575 , mwIdentifierList = Set.empty

76+ , mwLogMessages = []

7677 }

7778 (s ++ "\n")

7879 case parsed of

@@ -85,6 +86,7 @@ data MWState = MWState { mwOptions :: ReaderOptions

8586 , mwCategoryLinks :: [Inlines]

8687 , mwHeaderMap :: M.Map Inlines String

8788 , mwIdentifierList :: Set.Set String

89+ , mwLogMessages :: [LogMessage]

8890 }

8991

9092type MWParser m = ParserT [Char] MWState m

@@ -100,6 +102,10 @@ instance HasIdentifierList MWState where

100102 extractIdentifierList = mwIdentifierList

101103 updateIdentifierList f st = st{ mwIdentifierList = f $ mwIdentifierList st }

102104

105+instance HasLogMessages MWState where

106+ addLogMessage m s = s{ mwLogMessages = m : mwLogMessages s }

107+ getLogMessages = reverse . mwLogMessages

108+

103109--

104110-- auxiliary functions

105111--

@@ -187,6 +193,7 @@ parseMediaWiki = do

187193 let categories = if null categoryLinks

188194 then mempty

189195 else B.para $ mconcat $ intersperse B.space categoryLinks

196+ reportLogMessages

190197 return $ B.doc $ bs <> categories

191198

192199--

src/Text/Pandoc/Readers/Org.hs+2

@@ -35,6 +35,7 @@ import Text.Pandoc.Class (PandocMonad)

3535import Text.Pandoc.Definition

3636import Text.Pandoc.Error

3737import Text.Pandoc.Options

38+import Text.Pandoc.Parsing (reportLogMessages)

3839

3940import Control.Monad.Except (throwError)

4041import Control.Monad.Reader (runReaderT)

@@ -59,4 +60,5 @@ parseOrg :: PandocMonad m => OrgParser m Pandoc

5960parseOrg = do

6061 blocks' <- blockList

6162 meta' <- meta

63+ reportLogMessages

6264 return $ Pandoc meta' blocks'

src/Text/Pandoc/Readers/Org/ParserState.hs+8

@@ -60,7 +60,9 @@ import qualified Data.Set as Set

6060import Text.Pandoc.Builder (Blocks, Inlines, trimInlines)

6161import Text.Pandoc.Definition (Meta (..), nullMeta)

6262import Text.Pandoc.Options (ReaderOptions (..))

63+import Text.Pandoc.Logging

6364import Text.Pandoc.Parsing (HasHeaderMap (..), HasIdentifierList (..),

65+ HasLogMessages (..),

6466 HasLastStrPosition (..), HasQuoteContext (..),

6567 HasReaderOptions (..), ParserContext (..),

6668 QuoteContext (..), SourcePos)

@@ -104,6 +106,7 @@ data OrgParserState = OrgParserState

104106 , orgStateOptions :: ReaderOptions

105107 , orgStateParserContext :: ParserContext

106108 , orgStateTodoSequences :: [TodoSequence]

109+ , orgLogMessages :: [LogMessage]

107110 }

108111

109112data OrgParserLocal = OrgParserLocal { orgLocalQuoteContext :: QuoteContext }

@@ -130,6 +133,10 @@ instance HasHeaderMap OrgParserState where

130133 extractHeaderMap = orgStateHeaderMap

131134 updateHeaderMap f s = s{ orgStateHeaderMap = f (orgStateHeaderMap s) }

132135

136+instance HasLogMessages OrgParserState where

137+ addLogMessage msg st = st{ orgLogMessages = msg : orgLogMessages st }

138+ getLogMessages st = reverse $ orgLogMessages st

139+

133140instance Default OrgParserState where

134141 def = defaultOrgParserState

135142

@@ -150,6 +157,7 @@ defaultOrgParserState = OrgParserState

150157 , orgStateOptions = def

151158 , orgStateParserContext = NullState

152159 , orgStateTodoSequences = []

160+ , orgLogMessages = []

153161 }

154162

155163optionsToParserState :: ReaderOptions -> OrgParserState