Texinfo writer: Fixed internal cross-references.

Now we insert anchors after each header, and use @ref
instead of @uref for links.

Commas are now escaped as @comma{} only when needed; previously
all commas were escaped.  (This change is needed, in part, because @ref
commands must be followed by a real comma or period.)

Also insert a blank line in from of @verbatim environments.

2 files changed, 93 insertions, 33 deletions

diff --git a/src/Text/Pandoc/Writers/Texinfo.hs b/src/Text/Pandoc/Writers/Texinfo.hs
index 40e76c615..5ced0d9ec 100644
--- a/src/Text/Pandoc/Writers/Texinfo.hs
+++ b/src/Text/Pandoc/Writers/Texinfo.hs
@@ -45,6 +45,8 @@ data WriterState =
   WriterState { stStrikeout   :: Bool  -- document contains strikeout
               , stSuperscript :: Bool -- document contains superscript
               , stSubscript   :: Bool -- document contains subscript
+              , stEscapeComma :: Bool -- in a context where we need @comma
+              , stIdentifiers :: [String] -- header ids used already
               }
 
 {- TODO:
@@ -56,7 +58,8 @@ data WriterState =
 writeTexinfo :: WriterOptions -> Pandoc -> String
 writeTexinfo options document =
   evalState (pandocToTexinfo options $ wrapTop document) $
-  WriterState { stStrikeout = False, stSuperscript = False, stSubscript = False }
+  WriterState { stStrikeout = False, stSuperscript = False,
+                stEscapeComma = False, stSubscript = False, stIdentifiers = [] }
 
 -- | Add a "Top" node around the document, needed by Texinfo.
 wrapTop :: Pandoc -> Pandoc
@@ -95,7 +98,6 @@ stringToTexinfo = escapeStringUsing texinfoEscapes
   where texinfoEscapes = [ ('{', "@{")
                          , ('}', "@}")
                          , ('@', "@@")
-                         , (',', "@comma{}") -- only needed in argument lists
                          , ('\160', "@ ")
                          , ('\x2014', "---")
                          , ('\x2013', "--")
@@ -103,6 +105,14 @@ stringToTexinfo = escapeStringUsing texinfoEscapes
                          , ('\x2019', "'")
                          ]
 
+escapeCommas :: State WriterState Doc -> State WriterState Doc
+escapeCommas parser = do
+  oldEscapeComma <- gets stEscapeComma
+  modify $ \st -> st{ stEscapeComma = True }
+  res <- parser
+  modify $ \st -> st{ stEscapeComma = oldEscapeComma }
+  return res
+
 -- | Puts contents into Texinfo command.
 inCmd :: String -> Doc -> Doc
 inCmd cmd contents = char '@' <> text cmd <> braces contents
@@ -134,7 +144,8 @@ blockToTexinfo (BlockQuote lst) = do
            text "@end quotation"
 
 blockToTexinfo (CodeBlock _ str) = do
-  return $ text "@verbatim" $$
+  return $ blankline $$
+           text "@verbatim" $$
            flush (text str) $$
            text "@end verbatim" <> blankline
 
@@ -194,9 +205,13 @@ blockToTexinfo (Header 0 lst) = do
 blockToTexinfo (Header level lst) = do
   node <- inlineListForNode lst
   txt <- inlineListToTexinfo lst
+  idsUsed <- gets stIdentifiers
+  let id' = uniqueIdent lst idsUsed
+  modify $ \st -> st{ stIdentifiers = id' : idsUsed }
   return $ if (level > 0) && (level <= 4)
-              then blankline <> text "@node " <> node <> cr <>
-                   text (seccmd level) <> txt
+              then blankline <> text "@node " <> node $$
+                   text (seccmd level) <> txt $$
+                   text "@anchor" <> braces (text $ '#':id')
               else txt
   where
     seccmd 1 = "@chapter "
@@ -404,17 +419,21 @@ inlineToTexinfo (RawInline _ _) = return empty
 inlineToTexinfo (LineBreak) = return $ text "@*"
 inlineToTexinfo Space = return $ char ' '
 
+inlineToTexinfo (Link txt (src@('#':_), _)) = do
+  contents <- escapeCommas $ inlineListToTexinfo txt
+  return $ text "@ref" <>
+           braces (text (stringToTexinfo src) <> text "," <> contents)
 inlineToTexinfo (Link txt (src, _)) = do
   case txt of
         [Code _ x] | x == src ->  -- autolink
              do return $ text $ "@url{" ++ x ++ "}"
-        _ -> do contents <- inlineListToTexinfo txt
+        _ -> do contents <- escapeCommas $ inlineListToTexinfo txt
                 let src1 = stringToTexinfo src
                 return $ text ("@uref{" ++ src1 ++ ",") <> contents <>
                          char '}'
 
 inlineToTexinfo (Image alternate (source, _)) = do
-  content <- inlineListToTexinfo alternate
+  content <- escapeCommas $ inlineListToTexinfo alternate
   return $ text ("@image{" ++ base ++ ",,,") <> content <> text "," <>
            text (ext ++ "}")
   where
diff --git a/tests/writer.texinfo b/tests/writer.texinfo
index aed237f2a..d88f404b7 100644
--- a/tests/writer.texinfo
+++ b/tests/writer.texinfo
@@ -30,7 +30,7 @@ _@{\text\@}
 @title Pandoc Test Suite
 @author John MacFarlane
 @author Anonymous
-July 17@comma{} 2006
+July 17, 2006
 @end titlepage
 
 @node Top
@@ -64,28 +64,33 @@ This is a set of tests for pandoc. Most of them are adapted from John Gruber's m
 
 @node Headers
 @chapter Headers
+@anchor{#headers}
 @menu
 * Level 2 with an embedded link::
 @end menu
 
 @node Level 2 with an embedded link
 @section Level 2 with an @uref{/url,embedded link}
+@anchor{#level-2-with-an-embedded-link}
 @menu
 * Level 3 with emphasis::
 @end menu
 
 @node Level 3 with emphasis
 @subsection Level 3 with @emph{emphasis}
+@anchor{#level-3-with-emphasis}
 @menu
 * Level 4::
 @end menu
 
 @node Level 4
 @subsubsection Level 4
+@anchor{#level-4}
 Level 5
 
 @node Level 1
 @chapter Level 1
+@anchor{#level-1}
 @menu
 * Level 2 with emphasis::
 * Level 2::
@@ -93,16 +98,19 @@ Level 5
 
 @node Level 2 with emphasis
 @section Level 2 with @emph{emphasis}
+@anchor{#level-2-with-emphasis}
 @menu
 * Level 3::
 @end menu
 
 @node Level 3
 @subsection Level 3
+@anchor{#level-3}
 with no blank line
 
 @node Level 2
 @section Level 2
+@anchor{#level-2}
 with no blank line
 
 @iftex
@@ -114,6 +122,7 @@ with no blank line
 
 @node Paragraphs
 @chapter Paragraphs
+@anchor{#paragraphs}
 Here's a regular paragraph.
 
 In Markdown 1.0.0 and earlier. Version 8. This line turns into a list item. Because a hard-wrapped line in the middle of a paragraph looked like a list item.
@@ -131,6 +140,7 @@ There should be a hard line break@*here.
 
 @node Block Quotes
 @chapter Block Quotes
+@anchor{#block-quotes}
 E-mail style:
 
 @quotation
@@ -138,6 +148,7 @@ This is a block quote. It is pretty short.
 @end quotation
 @quotation
 Code in a block quote:
+
 @verbatim
 sub status {
     print "working";
@@ -175,7 +186,9 @@ And a following paragraph.
 
 @node Code Blocks
 @chapter Code Blocks
+@anchor{#code-blocks}
 Code:
+
 @verbatim
 ---- (should be four hyphens)
 
@@ -187,6 +200,7 @@ this code block is indented by one tab
 @end verbatim
 
 And:
+
 @verbatim
     this code block is indented by two tabs
 
@@ -202,6 +216,7 @@ These should not be escaped:  \$ \\ \> \[ \{
 
 @node Lists
 @chapter Lists
+@anchor{#lists}
 @menu
 * Unordered::
 * Ordered::
@@ -212,6 +227,7 @@ These should not be escaped:  \$ \\ \> \[ \{
 
 @node Unordered
 @section Unordered
+@anchor{#unordered}
 Asterisks tight:
 
 @itemize
@@ -289,6 +305,7 @@ Minus 3
 
 @node Ordered
 @section Ordered
+@anchor{#ordered}
 Tight:
 
 @enumerate 
@@ -343,7 +360,7 @@ Multiple paragraphs:
 
 @enumerate 
 @item
-Item 1@comma{} graf one.
+Item 1, graf one.
 
 Item 1. graf two. The quick brown fox jumped over the lazy dog's back.
 
@@ -357,6 +374,7 @@ Item 3.
 
 @node Nested
 @section Nested
+@anchor{#nested}
 @itemize
 @item
 Tab
@@ -417,6 +435,7 @@ Third
 
 @node Tabs and spaces
 @section Tabs and spaces
+@anchor{#tabs-and-spaces}
 @itemize
 @item
 this is a list item indented with tabs
@@ -437,6 +456,7 @@ this is an example list item indented with spaces
 
 @node Fancy list markers
 @section Fancy list markers
+@anchor{#fancy-list-markers}
 @enumerate 2
 @item
 begins with 2
@@ -447,7 +467,7 @@ with a continuation
 
 @enumerate 4
 @item
-sublist with roman numerals@comma{} starting with 4
+sublist with roman numerals, starting with 4
 @item
 more items
 @enumerate A
@@ -512,6 +532,7 @@ B. Williams
 
 @node Definition Lists
 @chapter Definition Lists
+@anchor{#definition-lists}
 Tight using spaces:
 
 @table @asis
@@ -564,11 +585,12 @@ Multiple blocks with italics:
 
 red fruit
 
-contains seeds@comma{} crisp@comma{} pleasant to taste
+contains seeds, crisp, pleasant to taste
 
 @item @emph{orange}
 
 orange fruit
+
 @verbatim
 { orange code block }
 @end verbatim
@@ -578,7 +600,7 @@ orange block quote
 @end quotation
 @end table
 
-Multiple definitions@comma{} tight:
+Multiple definitions, tight:
 
 @table @asis
 @item apple
@@ -591,7 +613,7 @@ orange fruit
 bank
 @end table
 
-Multiple definitions@comma{} loose:
+Multiple definitions, loose:
 
 @table @asis
 @item apple
@@ -608,7 +630,7 @@ bank
 
 @end table
 
-Blank line after term@comma{} indented marker@comma{} alternate markers:
+Blank line after term, indented marker, alternate markers:
 
 @table @asis
 @item apple
@@ -632,6 +654,7 @@ sublist
 
 @node HTML Blocks
 @chapter HTML Blocks
+@anchor{#html-blocks}
 Simple block on one line:
 
 foo
@@ -646,7 +669,8 @@ And this is @strong{strong}
 Here's a simple block:
 
 foo
-This should be a code block@comma{} though:
+This should be a code block, though:
+
 @verbatim
 <div>
     foo
@@ -654,11 +678,12 @@ This should be a code block@comma{} though:
 @end verbatim
 
 As should this:
+
 @verbatim
 <div>foo</div>
 @end verbatim
 
-Now@comma{} nested:
+Now, nested:
 
 foo
 This should just be an HTML comment:
@@ -666,13 +691,15 @@ This should just be an HTML comment:
 Multiline:
 
 Code block:
+
 @verbatim
 <!-- Comment -->
 @end verbatim
 
-Just plain comment@comma{} with trailing spaces on the line:
+Just plain comment, with trailing spaces on the line:
 
 Code:
+
 @verbatim
 <hr />
 @end verbatim
@@ -688,9 +715,10 @@ Hr's:
 
 @node Inline Markup
 @chapter Inline Markup
-This is @emph{emphasized}@comma{} and so @emph{is this}.
+@anchor{#inline-markup}
+This is @emph{emphasized}, and so @emph{is this}.
 
-This is @strong{strong}@comma{} and so @strong{is this}.
+This is @strong{strong}, and so @strong{is this}.
 
 An @emph{@uref{/url,emphasized link}}.
 
@@ -702,15 +730,15 @@ So is @strong{@emph{this}} word.
 
 So is @strong{@emph{this}} word.
 
-This is code: @code{>}@comma{} @code{$}@comma{} @code{\}@comma{} @code{\$}@comma{} @code{<html>}.
+This is code: @code{>}, @code{$}, @code{\}, @code{\$}, @code{<html>}.
 
 @textstrikeout{This is @emph{strikeout}.}
 
 Superscripts: a@textsuperscript{bc}d a@textsuperscript{@emph{hello}} a@textsuperscript{hello@ there}.
 
-Subscripts: H@textsubscript{2}O@comma{} H@textsubscript{23}O@comma{} H@textsubscript{many@ of@ them}O.
+Subscripts: H@textsubscript{2}O, H@textsubscript{23}O, H@textsubscript{many@ of@ them}O.
 
-These should not be superscripts or subscripts@comma{} because of the unescaped spaces: a^b c^d@comma{} a~b c~d.
+These should not be superscripts or subscripts, because of the unescaped spaces: a^b c^d, a~b c~d.
 
 @iftex
 @bigskip@hrule@bigskip
@@ -720,20 +748,21 @@ These should not be superscripts or subscripts@comma{} because of the unescaped
 @end ifnottex
 
 @node Smart quotes ellipses dashes
-@chapter Smart quotes@comma{} ellipses@comma{} dashes
-``Hello@comma{}'' said the spider. ```Shelob' is my name.''
+@chapter Smart quotes, ellipses, dashes
+@anchor{#smart-quotes-ellipses-dashes}
+``Hello,'' said the spider. ```Shelob' is my name.''
 
-`A'@comma{} `B'@comma{} and `C' are letters.
+`A', `B', and `C' are letters.
 
-`Oak@comma{}' `elm@comma{}' and `beech' are names of trees. So is `pine.'
+`Oak,' `elm,' and `beech' are names of trees. So is `pine.'
 
-`He said@comma{} ``I want to go.''' Were you alive in the 70's?
+`He said, ``I want to go.''' Were you alive in the 70's?
 
 Here is some quoted `@code{code}' and a ``@uref{http://example.com/?foo=1&bar=2,quoted link}''.
 
 Some dashes: one---two --- three---four --- five.
 
-Dashes between numbers: 5--7@comma{} 255--66@comma{} 1987--1999.
+Dashes between numbers: 5--7, 255--66, 1987--1999.
 
 Ellipses@dots{}and@dots{}and@dots{}.
 
@@ -746,6 +775,7 @@ Ellipses@dots{}and@dots{}and@dots{}.
 
 @node LaTeX
 @chapter LaTeX
+@anchor{#latex}
 @itemize
 @item
 @tex
@@ -771,9 +801,9 @@ These shouldn't be math:
 
 @itemize
 @item
-To get the famous equation@comma{} write @code{$e = mc^2$}.
+To get the famous equation, write @code{$e = mc^2$}.
 @item
-$22@comma{}000 is a @emph{lot} of money. So is $34@comma{}000. (It worked if ``lot'' is emphasized.)
+$22,000 is a @emph{lot} of money. So is $34,000. (It worked if ``lot'' is emphasized.)
 @item
 Shoes ($20) and socks ($5).
 @item
@@ -798,6 +828,7 @@ Cat    & 1      \\ \hline
 
 @node Special Characters
 @chapter Special Characters
+@anchor{#special-characters}
 Here is some unicode:
 
 @itemize
@@ -864,6 +895,7 @@ Minus: -
 
 @node Links
 @chapter Links
+@anchor{#links}
 @menu
 * Explicit::
 * Reference::
@@ -873,6 +905,7 @@ Minus: -
 
 @node Explicit
 @section Explicit
+@anchor{#explicit}
 Just a @uref{/url/,URL}.
 
 @uref{/url/,URL and title}.
@@ -893,6 +926,7 @@ Just a @uref{/url/,URL}.
 
 @node Reference
 @section Reference
+@anchor{#reference}
 Foo @uref{/url/,bar}.
 
 Foo @uref{/url/,bar}.
@@ -910,6 +944,7 @@ Indented @uref{/url,twice}.
 Indented @uref{/url,thrice}.
 
 This should [not][] be a link.
+
 @verbatim
 [not]: /url
 @end verbatim
@@ -920,6 +955,7 @@ Foo @uref{/url/,biz}.
 
 @node With ampersands
 @section With ampersands
+@anchor{#with-ampersands}
 Here's a @uref{http://example.com/?foo=1&bar=2,link with an ampersand in the URL}.
 
 Here's a link with an amersand in the link text: @uref{http://att.com/,AT&T}.
@@ -930,6 +966,7 @@ Here's an @uref{/script?foo=1&bar=2,inline link in pointy braces}.
 
 @node Autolinks
 @section Autolinks
+@anchor{#autolinks}
 With an ampersand: @url{http://example.com/?foo=1&bar=2}
 
 @itemize
@@ -947,6 +984,7 @@ An e-mail address: @uref{mailto:nobody@@nowhere.net,@code{nobody@@nowhere.net}}
 Blockquoted: @url{http://example.com/}
 @end quotation
 Auto-links should not occur here: @code{<http://example.com/>}
+
 @verbatim
 or here: <http://example.com/>
 @end verbatim
@@ -960,6 +998,7 @@ or here: <http://example.com/>
 
 @node Images
 @chapter Images
+@anchor{#images}
 From ``Voyage dans la Lune'' by Georges Melies (1902):
 
 @float
@@ -978,14 +1017,16 @@ Here is a movie @image{movie,,,movie,jpg} icon.
 
 @node Footnotes
 @chapter Footnotes
-Here is a footnote reference@comma{}@footnote{Here is the footnote. It can go anywhere after the footnote reference. It need not be placed at the end of the document.} and another.@footnote{Here's the long note. This one contains multiple blocks.
+@anchor{#footnotes}
+Here is a footnote reference,@footnote{Here is the footnote. It can go anywhere after the footnote reference. It need not be placed at the end of the document.} and another.@footnote{Here's the long note. This one contains multiple blocks.
 
 Subsequent blocks are indented to show that they belong to the footnote (as with list items).
+
 @verbatim
   { <code> }
 @end verbatim
 
-If you want@comma{} you can indent every line@comma{} but you can also be lazy and just indent the first line of each block.} This should @emph{not} be a footnote reference@comma{} because it contains a space.[^my note] Here is an inline note.@footnote{This is @emph{easier} to type. Inline notes may contain @uref{http://google.com,links} and @code{]} verbatim characters@comma{} as well as [bracketed text].}
+If you want, you can indent every line, but you can also be lazy and just indent the first line of each block.} This should @emph{not} be a footnote reference, because it contains a space.[^my note] Here is an inline note.@footnote{This is @emph{easier} to type. Inline notes may contain @uref{http://google.com,links} and @code{]} verbatim characters, as well as [bracketed text].}
 
 @quotation
 Notes can go in quotes.@footnote{In quote.}
@@ -995,6 +1036,6 @@ Notes can go in quotes.@footnote{In quote.}
 And in list items.@footnote{In list.}
 @end enumerate
 
-This paragraph should not be part of the note@comma{} as it is not indented.
+This paragraph should not be part of the note, as it is not indented.
 
 @bye