Merge pull request #1885 from mb21/html-reader-tables

fixes HTML Reader: tables

3 files changed, 397 insertions, 12 deletions

diff --git a/src/Text/Pandoc/Readers/HTML.hs b/src/Text/Pandoc/Readers/HTML.hs
index 2a23f2a62..02ff07e73 100644
--- a/src/Text/Pandoc/Readers/HTML.hs
+++ b/src/Text/Pandoc/Readers/HTML.hs
@@ -374,12 +374,20 @@ pTable = try $ do
   caption <- option mempty $ pInTags "caption" inline <* skipMany pBlank
   -- TODO actually read these and take width information from them
   widths' <- pColgroup <|> many pCol
-  head' <- option [] $ pOptInTag "thead" $ pInTags "tr" (pCell "th")
-  skipMany pBlank
-  rows <- pOptInTag "tbody"
-          $ many1 $ try $ skipMany pBlank >> pInTags "tr" (pCell "td")
-  skipMany pBlank
+  let pTh = option [] $ pInTags "tr" (pCell "th")
+      pTr = try $ skipMany pBlank >> pInTags "tr" (pCell "td" <|> pCell "th")
+      pTBody = do pOptInTag "tbody" $ many1 pTr
+  head'' <- pOptInTag "thead" pTh
+  head'  <- pOptInTag "tbody" $ do
+              if null head''
+                 then pTh
+                 else return head''
+  rowsLs <- many pTBody
+  rows'  <- pOptInTag "tfoot" $ many pTr
   TagClose _ <- pSatisfy (~== TagClose "table")
+  let rows = (concat rowsLs) ++ rows'
+  -- fail on empty table
+  guard $ not $ null head' && null rows
   let isSinglePlain x = case B.toList x of
                              [Plain _] -> True
                              _         -> False
@@ -624,14 +632,17 @@ pInTags tagtype parser = try $ do
   pSatisfy (~== TagOpen tagtype [])
   mconcat <$> manyTill parser (pCloses tagtype <|> eof)
 
-pOptInTag :: String -> TagParser a
-          -> TagParser a
-pOptInTag tagtype parser = try $ do
-  open <- option False (pSatisfy (~== TagOpen tagtype []) >> return True)
+-- parses p, preceeded by an optional opening tag
+-- and followed by an optional closing tags
+pOptInTag :: String -> TagParser a -> TagParser a
+pOptInTag tagtype p = try $ do
+  skipMany pBlank
+  optional $ pSatisfy (~== TagOpen tagtype [])
+  skipMany pBlank
+  x <- p
   skipMany pBlank
-  x <- parser
+  optional $ pSatisfy (~== TagClose tagtype)
   skipMany pBlank
-  when open $ pCloses tagtype
   return x
 
 pCloses :: String -> TagParser ()
diff --git a/tests/html-reader.html b/tests/html-reader.html
index e9ba2a68b..749925b2a 100644
--- a/tests/html-reader.html
+++ b/tests/html-reader.html
@@ -433,6 +433,7 @@ An e-mail address: nobody [at] nowhere.net<blockquote>
 <p>text<em>   Leading spaces</em></p>
 <p><em>Trailing spaces    </em>text</p>
 <h1>Tables</h1>
+<h2>Tables with Headers</h2>
 <table>
     <tr>
         <th>X</th>
@@ -450,5 +451,251 @@ An e-mail address: nobody [at] nowhere.net<blockquote>
         <td>6</td>
     </tr>
 </table>
+<hr />
+<table>
+    <thead>
+    <tr>
+        <th>X</th>
+        <th>Y</th>
+        <th>Z</th>
+    </tr>
+    </thead>
+    <tbody>
+    <tr>
+        <td>1</td>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tbody>
+</table>
+<hr />
+<table>
+    <thead>
+    <tr>
+        <th>X</th>
+        <th>Y</th>
+        <th>Z</th>
+    </tr>
+    </thead>
+    <tbody>
+    <tr>
+        <th>1</th>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    <tr>
+        <th>4</th>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tbody>
+</table>
+<hr />
+<table>
+    <thead>
+    <tr>
+        <th>X</th>
+        <th>Y</th>
+        <th>Z</th>
+    </tr>
+    </thead>
+    <tbody>
+    <tr>
+        <th>1</th>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    </tbody>
+    <tfoot>
+    <tr>
+        <th>4</th>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tfoot>
+</table>
+<hr />
+<table>
+    <tr>
+        <th>X</th>
+        <th>Y</th>
+        <th>Z</th>
+    </tr>
+    <tr>
+        <th>1</th>
+        <th>2</th>
+        <th>3</th>
+    </tr>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+</table>
+<hr />
+<table>
+    <tbody>
+    <tr>
+        <th>X</th>
+        <th>Y</th>
+        <th>Z</th>
+    </tr>
+    <tr>
+        <td>1</td>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tbody>
+</table>
+<hr />
+<table>
+    <thead>
+    </thead>
+    <tbody>
+    <tr>
+        <th>X</th>
+        <th>Y</th>
+        <th>Z</th>
+    </tr>
+    <tr>
+        <td>1</td>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tbody>
+</table>
+<hr />
+<table>
+    <thead>
+    <tr>
+        <th>X</th>
+        <th>Y</th>
+        <th>Z</th>
+    </tr>
+    </thead>
+    <tbody>
+    <tr>
+        <td>1</td>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    </tbody>
+    <tbody>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tbody>
+</table>
+<hr />
+<table>
+    <thead>
+    <tr>
+        <th>X</th>
+        <th>Y</th>
+        <th>Z</th>
+    </tr>
+    </thead>
+    <tbody>
+    <tr>
+        <td>1</td>
+        <td><p>2</p></td>
+        <td>3</td>
+    </tr>
+    </tbody>
+    <tbody>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tbody>
+</table>
+<h2>Tables without Headers</h2>
+<table>
+    <tbody>
+    <tr>
+        <td>1</td>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tbody>
+</table>
+<hr />
+<table>
+    <tr>
+        <td>1</td>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+</table>
+<hr />
+<table>
+    <thead>
+    </thead>
+    <tbody>
+    <tr>
+        <td>1</td>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tbody>
+</table>
+<hr />
+<table>
+    <tbody>
+    <tr>
+        <td>1</td>
+        <td>2</td>
+        <td>3</td>
+    </tr>
+    </tbody>
+    <tfoot>
+    <tr>
+        <td>4</td>
+        <td>5</td>
+        <td>6</td>
+    </tr>
+    </tfoot>
+</table>
+<h2>Empty Tables</h2>
+<p>This section should be empty.</p>
+<table>
+    <tbody>
+    </tbody>
+</table>
+<table>
+</table>
 </body>
 </html>
diff --git a/tests/html-reader.native b/tests/html-reader.native
index aef6e40fc..b2d660fda 100644
--- a/tests/html-reader.native
+++ b/tests/html-reader.native
@@ -311,6 +311,7 @@ Pandoc (Meta {unMeta = fromList [("generator",MetaInlines [Str "pandoc"]),("titl
 ,Para [Str "text",Space,Emph [Str "Leading",Space,Str "spaces"]]
 ,Para [Emph [Str "Trailing",Space,Str "spaces"],Space,Str "text"]
 ,Header 1 ("",[],[]) [Str "Tables"]
+,Header 2 ("",[],[]) [Str "Tables",Space,Str "with",Space,Str "Headers"]
 ,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
  [[Plain [Str "X"]]
  ,[Plain [Str "Y"]]
@@ -320,4 +321,130 @@ Pandoc (Meta {unMeta = fromList [("generator",MetaInlines [Str "pandoc"]),("titl
   ,[Plain [Str "3"]]]
  ,[[Plain [Str "4"]]
   ,[Plain [Str "5"]]
-  ,[Plain [Str "6"]]]]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ [[Plain [Str "X"]]
+ ,[Plain [Str "Y"]]
+ ,[Plain [Str "Z"]]]
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ [[Plain [Str "X"]]
+ ,[Plain [Str "Y"]]
+ ,[Plain [Str "Z"]]]
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ [[Plain [Str "X"]]
+ ,[Plain [Str "Y"]]
+ ,[Plain [Str "Z"]]]
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ [[Plain [Str "X"]]
+ ,[Plain [Str "Y"]]
+ ,[Plain [Str "Z"]]]
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ [[Plain [Str "X"]]
+ ,[Plain [Str "Y"]]
+ ,[Plain [Str "Z"]]]
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ [[Plain [Str "X"]]
+ ,[Plain [Str "Y"]]
+ ,[Plain [Str "Z"]]]
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ [[Plain [Str "X"]]
+ ,[Plain [Str "Y"]]
+ ,[Plain [Str "Z"]]]
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.3333333333333333,0.3333333333333333,0.3333333333333333]
+ [[Plain [Str "X"]]
+ ,[Plain [Str "Y"]]
+ ,[Plain [Str "Z"]]]
+ [[[Plain [Str "1"]]
+  ,[Para [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,Header 2 ("",[],[]) [Str "Tables",Space,Str "without",Space,Str "Headers"]
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ []
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ []
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ []
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,HorizontalRule
+,Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]
+ []
+ [[[Plain [Str "1"]]
+  ,[Plain [Str "2"]]
+  ,[Plain [Str "3"]]]
+ ,[[Plain [Str "4"]]
+  ,[Plain [Str "5"]]
+  ,[Plain [Str "6"]]]]
+,Header 2 ("",[],[]) [Str "Empty",Space,Str "Tables"]
+,Para [Str "This",Space,Str "section",Space,Str "should",Space,Str "be",Space,Str "empty."]]