Markdown reader: better handling of paragraph in div.

Previously text that ended a div would be parsed as Plain
unless there was a blank line before the closing div tag.

Test case:

    <div class="first">
    This is a paragraph.

    This is another paragraph.
    </div>

Closes #1591.

src/Text/Pandoc/Readers/Markdown.hs

@@ -927,6 +927,12 @@ para = try $ do
               <|> (guardDisabled Ext_blank_before_header >> () <$ lookAhead header)
               <|> (guardEnabled Ext_lists_without_preceding_blankline >>
                        () <$ lookAhead listStart)
+              <|> do guardEnabled Ext_native_divs
+                     inHtmlBlock <- stateInHtmlBlock <$> getState
+                     case inHtmlBlock of
+                          Just "div" -> () <$
+                                       lookAhead (htmlTag (~== TagClose "div"))
+                          _          -> mzero
             return $ do
               result' <- result
               case B.toList result' of
@@ -1611,6 +1617,7 @@ endline = try $ do
   guardEnabled Ext_blank_before_header <|> notFollowedBy (char '#') -- atx header
   guardDisabled Ext_backtick_code_blocks <|>
      notFollowedBy (() <$ (lookAhead (char '`') >> codeBlockFenced))
+  notFollowedByHtmlCloser
   (eof >> return mempty)
     <|> (guardEnabled Ext_hard_line_breaks >> return (return B.linebreak))
     <|> (guardEnabled Ext_ignore_line_breaks >> return mempty)

tests/Tests/Readers/Markdown.hs

@@ -266,10 +266,10 @@ tests = [ testGroup "inline code"
         , testGroup "lists"
           [ "issue #1154" =:
               " -  <div>\n    first div breaks\n    </div>\n\n    <button>if this button exists</button>\n\n    <div>\n    with this div too.\n    </div>\n"
-              =?> bulletList [divWith nullAttr (plain $ text "first div breaks") <>
+              =?> bulletList [divWith nullAttr (para $ text "first div breaks") <>
                               rawBlock "html" "<button>" <>
                               plain (text "if this button exists") <>
                               rawBlock "html" "</button>" <>
-                              divWith nullAttr (plain $ text "with this div too.")]
+                              divWith nullAttr (para $ text "with this div too.")]
           ]
         ]

tests/testsuite.native

@@ -230,7 +230,7 @@ Pandoc (Meta {unMeta = fromList [("author",MetaList [MetaInlines [Str "John",Spa
 ,Para [Str "Simple",Space,Str "block",Space,Str "on",Space,Str "one",Space,Str "line:"]
 ,Div ("",[],[]) [Plain [Str "foo"]]
 ,Para [Str "And",Space,Str "nested",Space,Str "without",Space,Str "indentation:"]
-,Div ("",[],[]) [Div ("",[],[]) [Div ("",[],[]) [Plain [Str "foo"]]],Div ("",[],[]) [Plain [Str "bar"]]]
+,Div ("",[],[]) [Div ("",[],[]) [Div ("",[],[]) [Para [Str "foo"]]],Div ("",[],[]) [Plain [Str "bar"]]]
 ,Para [Str "Interpreted",Space,Str "markdown",Space,Str "in",Space,Str "a",Space,Str "table:"]
 ,RawBlock (Format "html") "<table>"
 ,RawBlock (Format "html") "<tr>"
@@ -244,7 +244,7 @@ Pandoc (Meta {unMeta = fromList [("author",MetaList [MetaInlines [Str "John",Spa
 ,RawBlock (Format "html") "</table>"
 ,RawBlock (Format "html") "<script type=\"text/javascript\">document.write('This *should not* be interpreted as markdown');</script>"
 ,Para [Str "Here\8217s",Space,Str "a",Space,Str "simple",Space,Str "block:"]
-,Div ("",[],[]) [Plain [Str "foo"]]
+,Div ("",[],[]) [Para [Str "foo"]]
 ,Para [Str "This",Space,Str "should",Space,Str "be",Space,Str "a",Space,Str "code",Space,Str "block,",Space,Str "though:"]
 ,CodeBlock ("",[],[]) "<div>\n    foo\n</div>"
 ,Para [Str "As",Space,Str "should",Space,Str "this:"]

tests/writer.asciidoc

@@ -378,6 +378,7 @@ foo
 And nested without indentation:
 
 foo
+
 bar
 Interpreted markdown in a table:
 
@@ -386,6 +387,7 @@ And this is *strong*
 Here’s a simple block:
 
 foo
+
 This should be a code block, though:
 
 -------

tests/writer.context

@@ -548,6 +548,7 @@ foo
 And nested without indentation:
 
 foo
+
 bar
 Interpreted markdown in a table:
 
@@ -556,6 +557,7 @@ And this is {\bf strong}
 Here's a simple block:
 
 foo
+
 This should be a code block, though:
 
 \starttyping

tests/writer.dokuwiki

@@ -298,6 +298,7 @@ And nested without indentation:
 foo
 
 
+
 bar
 
 
@@ -324,6 +325,7 @@ Here’s a simple block:
 
 foo
 
+
 This should be a code block, though:
 
 <code><div>

tests/writer.html

@@ -328,7 +328,7 @@ <h1 id="html-blocks">HTML Blocks</h1>
 <div>
 <div>
 <div>
-foo
+<p>foo</p>
 </div>
 </div>
 <div>
@@ -349,7 +349,7 @@ <h1 id="html-blocks">HTML Blocks</h1>
 <script type="text/javascript">document.write('This *should not* be interpreted as markdown');</script>
 <p>Here’s a simple block:</p>
 <div>
-foo
+<p>foo</p>
 </div>
 <p>This should be a code block, though:</p>
 <pre><code>&lt;div&gt;

tests/writer.icml

@@ -1374,7 +1374,7 @@ These should not be escaped:  \$ \\ \&gt; \[ \{</Content>
     <Content>And nested without indentation:</Content>
   </CharacterStyleRange><Br />
 </ParagraphStyleRange>
-<ParagraphStyleRange AppliedParagraphStyle="">
+<ParagraphStyleRange AppliedParagraphStyle="ParagraphStyle/Paragraph">
   <CharacterStyleRange AppliedCharacterStyle="$ID/NormalCharacterStyle">
     <Content>foo</Content>
   </CharacterStyleRange><Br />
@@ -1455,7 +1455,7 @@ These should not be escaped:  \$ \\ \&gt; \[ \{</Content>
     <Content>Here’s a simple block:</Content>
   </CharacterStyleRange><Br />
 </ParagraphStyleRange>
-<ParagraphStyleRange AppliedParagraphStyle="">
+<ParagraphStyleRange AppliedParagraphStyle="ParagraphStyle/Paragraph">
   <CharacterStyleRange AppliedCharacterStyle="$ID/NormalCharacterStyle">
     <Content>foo</Content>
   </CharacterStyleRange><Br />

tests/writer.man

@@ -450,6 +450,7 @@ Simple block on one line:
 foo
 .PP
 And nested without indentation:
+.PP
 foo
 bar
 .PP
@@ -458,6 +459,7 @@ This is \f[I]emphasized\f[]
 And this is \f[B]strong\f[]
 .PP
 Here's a simple block:
+.PP
 foo
 .PP
 This should be a code block, though:

tests/writer.mediawiki

@@ -325,6 +325,7 @@ And nested without indentation:
 
 foo
 
+
 </div>
 
 </div>
@@ -354,6 +355,7 @@ Here’s a simple block:
 
 foo
 
+
 </div>
 This should be a code block, though:
 

tests/writer.native

@@ -230,7 +230,7 @@ Pandoc (Meta {unMeta = fromList [("author",MetaList [MetaInlines [Str "John",Spa
 ,Para [Str "Simple",Space,Str "block",Space,Str "on",Space,Str "one",Space,Str "line:"]
 ,Div ("",[],[]) [Plain [Str "foo"]]
 ,Para [Str "And",Space,Str "nested",Space,Str "without",Space,Str "indentation:"]
-,Div ("",[],[]) [Div ("",[],[]) [Div ("",[],[]) [Plain [Str "foo"]]],Div ("",[],[]) [Plain [Str "bar"]]]
+,Div ("",[],[]) [Div ("",[],[]) [Div ("",[],[]) [Para [Str "foo"]]],Div ("",[],[]) [Plain [Str "bar"]]]
 ,Para [Str "Interpreted",Space,Str "markdown",Space,Str "in",Space,Str "a",Space,Str "table:"]
 ,RawBlock (Format "html") "<table>"
 ,RawBlock (Format "html") "<tr>"
@@ -244,7 +244,7 @@ Pandoc (Meta {unMeta = fromList [("author",MetaList [MetaInlines [Str "John",Spa
 ,RawBlock (Format "html") "</table>"
 ,RawBlock (Format "html") "<script type=\"text/javascript\">document.write('This *should not* be interpreted as markdown');</script>"
 ,Para [Str "Here\8217s",Space,Str "a",Space,Str "simple",Space,Str "block:"]
-,Div ("",[],[]) [Plain [Str "foo"]]
+,Div ("",[],[]) [Para [Str "foo"]]
 ,Para [Str "This",Space,Str "should",Space,Str "be",Space,Str "a",Space,Str "code",Space,Str "block,",Space,Str "though:"]
 ,CodeBlock ("",[],[]) "<div>\n    foo\n</div>"
 ,Para [Str "As",Space,Str "should",Space,Str "this:"]

tests/writer.rtf

@@ -208,13 +208,13 @@ These should not be escaped:  \\$ \\\\ \\> \\[ \\\{\par}
 {\pard \ql \f0 \sa180 \li0 \fi0 Simple block on one line:\par}
 {\pard \ql \f0 \sa0 \li0 \fi0 foo\par}
 {\pard \ql \f0 \sa180 \li0 \fi0 And nested without indentation:\par}
-{\pard \ql \f0 \sa0 \li0 \fi0 foo\par}
+{\pard \ql \f0 \sa180 \li0 \fi0 foo\par}
 {\pard \ql \f0 \sa0 \li0 \fi0 bar\par}
 {\pard \ql \f0 \sa180 \li0 \fi0 Interpreted markdown in a table:\par}
 {\pard \ql \f0 \sa0 \li0 \fi0 This is {\i emphasized}\par}
 {\pard \ql \f0 \sa0 \li0 \fi0 And this is {\b strong}\par}
 {\pard \ql \f0 \sa180 \li0 \fi0 Here\u8217's a simple block:\par}
-{\pard \ql \f0 \sa0 \li0 \fi0 foo\par}
+{\pard \ql \f0 \sa180 \li0 \fi0 foo\par}
 {\pard \ql \f0 \sa180 \li0 \fi0 This should be a code block, though:\par}
 {\pard \ql \f0 \sa180 \li0 \fi0 \f1 <div>\line
     foo\line

tests/writer.textile

@@ -367,6 +367,7 @@ And nested without indentation:
 
 foo
 
+
 </div>
 
 
@@ -400,6 +401,7 @@ Here's a simple block:
 
 foo
 
+
 </div>
 
 This should be a code block, though: