Implemented implicit reference-style links to section headers in markdown.

For example, if you have a header '# Supported architectures', you can
link to it with '[Supported architectures]'.  If there are multiple
headers with this label, the link will point to the first of them.
Implicit references are always overridden by explicitly specified references.
Addresses Issue #20.

+ Moved isPunctuation, uniqueIdentifiers, and inlineListToIdentifier from
  Text.Pandoc.Writers.HTML to Text.Pandoc.Shared.

+ Added stHeaders to ParserState.   This holds a list of header texts
  used in the document, and is used to construct implicit header references.

+ In Text.Pandoc.Readers.Markdown, added call to headerReference
  parser in initial parsing pass, constructing a list of section header
  labels. This is then passed to uniqueIdentifiers to produce
  identifiers, and a list of implicit references is constructed. This is
  added to the end of the explicitly specified references, so it will be
  overridden by explicitly specified references. All of this processing
  is skipped if --strict was specified.

+ Modified documentation in README.



git-svn-id: https://pandoc.googlecode.com/svn/trunk@1086 788f1e2b-df1e-0410-8736-df70ead52e1b

1 files changed, 0 insertions, 46 deletions

diff --git a/src/Text/Pandoc/Writers/HTML.hs b/src/Text/Pandoc/Writers/HTML.hs
index 7ec95d8ef..56ca5ca48 100644
--- a/src/Text/Pandoc/Writers/HTML.hs
+++ b/src/Text/Pandoc/Writers/HTML.hs
@@ -193,15 +193,6 @@ obfuscateChar char =
 obfuscateString :: String -> String
 obfuscateString = concatMap obfuscateChar . decodeCharacterReferences
 
--- | True if character is a punctuation character (unicode).
-isPunctuation :: Char -> Bool
-isPunctuation c =
-  let c' = ord c
-  in  if c `elem` "!\"'()*,-./:;<>?[\\]`{|}~" || c' >= 0x2000 && c' <= 0x206F ||
-         c' >= 0xE000 && c' <= 0xE0FF
-         then True
-         else False
-
 -- | Add CSS for document header.
 addToCSS :: String -> State WriterState ()
 addToCSS item = do
@@ -209,43 +200,6 @@ addToCSS item = do
   let current = stCSS st
   put $ st {stCSS = S.insert item current}
 
--- | Convert Pandoc inline list to plain text identifier.
-inlineListToIdentifier :: [Inline] -> String
-inlineListToIdentifier [] = ""
-inlineListToIdentifier (x:xs) = 
-  xAsText ++ inlineListToIdentifier xs
-  where xAsText = case x of
-          Str s          -> filter (\c -> c == '-' || not (isPunctuation c)) $
-                            concat $ intersperse "-" $ words $ map toLower s
-          Emph lst       -> inlineListToIdentifier lst
-          Strikeout lst  -> inlineListToIdentifier lst
-          Superscript lst -> inlineListToIdentifier lst
-          Subscript lst  -> inlineListToIdentifier lst
-          Strong lst     -> inlineListToIdentifier lst
-          Quoted _ lst   -> inlineListToIdentifier lst
-          Code s         -> s
-          Space          -> "-"
-          EmDash         -> "-"
-          EnDash         -> "-"
-          Apostrophe     -> ""
-          Ellipses       -> ""
-          LineBreak      -> "-"
-          TeX _          -> ""
-          HtmlInline _   -> ""
-          Link lst _     -> inlineListToIdentifier lst
-          Image lst _    -> inlineListToIdentifier lst
-          Note _         -> ""
-
--- | Return unique identifiers for list of inline lists.
-uniqueIdentifiers :: [[Inline]] -> [String]
-uniqueIdentifiers ls =
-  let addIdentifier (nonuniqueIds, uniqueIds) l =
-        let new = inlineListToIdentifier l
-            matches = length $ filter (== new) nonuniqueIds
-            new' = new ++ if matches > 0 then ("-" ++ show matches) else ""
-        in  (new:nonuniqueIds, new':uniqueIds)
-  in  reverse $ snd $ foldl addIdentifier ([],[]) ls
-
 -- | Convert Pandoc block element to HTML.
 blockToHtml :: WriterOptions -> Block -> State WriterState Html
 blockToHtml opts Null = return $ noHtml