Preserve attributes on HTML paragraphs

MANUAL.txt

@@ -5374,7 +5374,7 @@ from being interpreted as Markdown.
 
 ### Extension: `native_divs` ###
 
-Use native pandoc `Div` blocks for content inside `<div>` tags.
+Use native pandoc `Div` blocks for content inside `<div>` tags. This extension also influences how HTML `<p>` tags with attributes are processed, by wrapping them in `Div` blocks to better preserve their attributes during conversion.
 For the most part this should give the same output as
 `markdown_in_html_blocks`, but it makes it easier to write pandoc
 filters to manipulate groups of blocks.

src/Text/Pandoc/Readers/HTML.hs

@@ -623,13 +623,38 @@ pPlain = do
      then return mempty
      else return $ B.plain contents
 
-pPara :: PandocMonad m => TagParser m Blocks
-pPara = do
+-- Helper function for pPara when significant attributes are present
+pParaWithWrapper :: PandocMonad m => Attr -> TagParser m Blocks
+pParaWithWrapper (ident, classes, kvs) = do
+  guardEnabled Ext_native_divs -- Ensure native_divs is enabled for this behavior
+  contents <- trimInlines <$> pInTags "p" inline
+  (do guardDisabled Ext_empty_paragraphs
+      guard (null contents)
+      return mempty) <|> do
+    let wrapperAttr = ("wrapper", "1")
+    let finalKVs = wrapperAttr : kvs
+    let finalAttrs = (ident, classes, finalKVs)
+    return $ B.divWith finalAttrs (B.para contents)
+
+-- Helper function for pPara when no significant attributes are present
+pParaSimple :: PandocMonad m => TagParser m Blocks
+pParaSimple = do
   contents <- trimInlines <$> pInTags "p" inline
   (do guardDisabled Ext_empty_paragraphs
       guard (null contents)
-      return mempty)
-    <|> return (B.para contents)
+      return mempty) <|>
+    return (B.para contents)
+
+pPara :: PandocMonad m => TagParser m Blocks
+pPara = do
+  TagOpen _ attr' <- lookAhead $ pSatisfy (matchTagOpen "p" [])
+  let attr@(ident, classes, kvs) = toAttr attr'
+  -- "Significant" attributes are any id, class, or key-value pair.
+  let hasSignificantAttributes = not (T.null ident) || not (null classes) || not (null kvs)
+
+  if hasSignificantAttributes
+    then pParaWithWrapper attr
+    else pParaSimple
 
 pFigure :: PandocMonad m => TagParser m Blocks
 pFigure = do

src/Text/Pandoc/Writers/HTML.hs

@@ -750,6 +750,13 @@ blockToHtmlInner opts (Para lst) = do
 blockToHtmlInner opts (LineBlock lns) = do
   htmlLines <- inlineListToHtml opts $ intercalate [LineBreak] lns
   return $ H.div ! A.class_ "line-block" $ htmlLines
+blockToHtmlInner opts (Div (ident, classes, kvs) [Para pans]) | Just "1" <- lookup "wrapper" kvs = do
+  -- This is a paragraph that was wrapped in a Div by the reader
+  -- Unwrap it back to a <p> tag, transferring attributes from the Div
+  let pKVs = filter (\(k,_) -> k /= "wrapper") kvs
+  let pAttr = (ident, classes, pKVs)
+  inner <- inlineListToHtml opts pans
+  addAttrs opts pAttr (H.p inner)
 blockToHtmlInner opts (Div (ident, "section":dclasses, dkvs)
                    (Header level
                      hattr@(hident,hclasses,hkvs) ils : xs)) = do

test/Tests/Readers/HTML.hs

@@ -137,6 +137,44 @@ tests = [ testGroup "base tag"
             =?>
             codeBlockWith ("c", [], []) "print('hi mom!')"
           ]
+        , testGroup "paragraph attributes"
+          [ test htmlNativeDivs "paragraph with id and class" $
+            "<p id=\"mypara\" class=\"important\">This is a paragraph.</p>" =?>
+            doc (divWith ("mypara", ["important"], [("wrapper", "1")]) (para (text "This is a paragraph.")))
+          , test htmlNativeDivs "paragraph with id only" $
+            "<p id=\"mypara\">This is a paragraph.</p>" =?>
+            doc (divWith ("mypara", [], [("wrapper", "1")]) (para (text "This is a paragraph.")))
+          , test htmlNativeDivs "paragraph with class only" $
+            "<p class=\"important\">This is a paragraph.</p>" =?>
+            doc (divWith ("", ["important"], [("wrapper", "1")]) (para (text "This is a paragraph.")))
+          , test htmlNativeDivs "paragraph with multiple classes" $
+            "<p class=\"important urgent\">This is a paragraph.</p>" =?>
+            doc (divWith ("", ["important", "urgent"], [("wrapper", "1")]) (para (text "This is a paragraph.")))
+          , test htmlNativeDivs "paragraph with key-value attributes" $
+            "<p data-foo=\"bar\">This is a paragraph.</p>" =?>
+            doc (divWith ("", [], [("wrapper", "1"), ("foo", "bar")]) (para (text "This is a paragraph.")))
+          , test htmlNativeDivs "paragraph without attributes" $
+            "<p>This is a normal paragraph.</p>" =?>
+            doc (para (text "This is a normal paragraph."))
+          , test htmlNativeDivs "paragraph with align only (center)" $
+            "<p align=\"center\">Aligned paragraph.</p>" =?>
+            doc (divWith ("", [], [("wrapper", "1"), ("align", "center")]) (para (text "Aligned paragraph.")))
+          , test htmlNativeDivs "paragraph with align only (right)" $
+            "<p align=\"right\">Aligned paragraph.</p>" =?>
+            doc (divWith ("", [], [("wrapper", "1"), ("align", "right")]) (para (text "Aligned paragraph.")))
+          , test htmlNativeDivs "paragraph with align and id" $
+            "<p id=\"foo\" align=\"left\">Aligned paragraph with id.</p>" =?>
+            doc (divWith ("foo", [], [("wrapper", "1"), ("align", "left")]) (para (text "Aligned paragraph with id.")))
+          , test htmlNativeDivs "paragraph with align and class" $
+            "<p class=\"bar\" align=\"justify\">Aligned paragraph with class.</p>" =?>
+            doc (divWith ("", ["bar"], [("wrapper", "1"), ("align", "justify")]) (para (text "Aligned paragraph with class.")))
+          , test htmlNativeDivs "paragraph with invalid align" $
+            "<p align=\"invalid\">Invalid align.</p>" =?>
+            doc (divWith ("", [], [("wrapper", "1"), ("align", "invalid")]) (para (text "Invalid align.")))
+          , test htmlNativeDivs "paragraph with invalid align and id" $
+            "<p id=\"baz\" align=\"invalid\">Invalid align with id.</p>" =?>
+            doc (divWith ("baz", [], [("wrapper", "1"), ("align", "invalid")]) (para (text "Invalid align with id.")))
+          ]
         , askOption $ \(QuickCheckTests numtests) ->
             testProperty "Round trip" $
               withMaxSuccess (if QuickCheckTests numtests == defaultValue

test/Tests/Writers/HTML.hs

@@ -21,6 +21,9 @@ htmlQTags = unpack
   . purely (writeHtml4String def{ writerWrapText = WrapNone, writerHtmlQTags = True })
   . toPandoc
 
+html5 :: (ToPandoc a) => a -> String
+html5 = unpack . purely (writeHtml5String def{ writerWrapText = WrapNone }) . toPandoc
+
 {-
   "my test" =: X =?> Y
 
@@ -208,6 +211,32 @@ tests =
           , "</div>"
           ]
       ]
+  , testGroup "paragraph attributes"
+    [ "paragraph with id and class" =:
+      divWith ("mypara", ["important"], [("wrapper", "1")]) (para (text "This is a paragraph."))
+      =?> "<p id=\"mypara\" class=\"important\">This is a paragraph.</p>"
+    , "paragraph with id only" =:
+      divWith ("mypara", [], [("wrapper", "1")]) (para (text "This is a paragraph."))
+      =?> "<p id=\"mypara\">This is a paragraph.</p>"
+    , "paragraph with class only" =:
+      divWith ("", ["important"], [("wrapper", "1")]) (para (text "This is a paragraph."))
+      =?> "<p class=\"important\">This is a paragraph.</p>"
+    , "paragraph with multiple classes" =:
+      divWith ("", ["important", "urgent"], [("wrapper", "1")]) (para (text "This is a paragraph."))
+      =?> "<p class=\"important urgent\">This is a paragraph.</p>"
+    , test html5 "paragraph with key-value attributes"
+        (divWith ("", [], [("wrapper", "1"), ("foo", "bar")]) (para (text "This is a paragraph."))
+        , "<p data-foo=\"bar\">This is a paragraph.</p>" :: String)
+    , "paragraph without wrapper attribute" =:
+      divWith ("mydiv", ["someclass"], []) (para (text "This is a div, not a p."))
+      =?> "<div id=\"mydiv\" class=\"someclass\">\n<p>This is a div, not a p.</p>\n</div>"
+    , "paragraph with wrapper and other attributes" =:
+      divWith ("mypara", ["important"], [("wrapper", "1"), ("data-value", "123")]) (para (text "This is a paragraph."))
+      =?> "<p id=\"mypara\" class=\"important\" data-value=\"123\">This is a paragraph.</p>"
+    , "paragraph with wrapper and align" =:
+      divWith ("mypara", [], [("wrapper", "1"), ("align", "center")]) (para (text "Aligned paragraph."))
+      =?> "<p id=\"mypara\" align=\"center\">Aligned paragraph.</p>"
+    ]
   ]
   where
     tQ :: (ToString a, ToPandoc a)

test/command/10768.md

@@ -0,0 +1,68 @@
+```
+% pandoc -f html -t native
+<p id="test" class="foo bar" data-custom="value">This is a paragraph with attributes.</p>
+^D
+[ Div
+    ( "test"
+    , [ "foo" , "bar" ]
+    , [ ( "wrapper" , "1" ) , ( "custom" , "value" ) ]
+    )
+    [ Para
+        [ Str "This"
+        , Space
+        , Str "is"
+        , Space
+        , Str "a"
+        , Space
+        , Str "paragraph"
+        , Space
+        , Str "with"
+        , Space
+        , Str "attributes."
+        ]
+    ]
+]
+```
+
+```
+% pandoc -f native -t html
+[ Div
+    ( "test"
+    , [ "foo" , "bar" ]
+    , [ ( "wrapper" , "1" ) , ( "custom" , "value" ) ]
+    )
+    [ Para
+        [ Str "This"
+        , Space
+        , Str "is"
+        , Space
+        , Str "a"
+        , Space
+        , Str "paragraph"
+        , Space
+        , Str "with"
+        , Space
+        , Str "attributes."
+        ]
+    ]
+]
+^D
+<p id="test" class="foo bar" data-custom="value">This is a paragraph
+with attributes.</p>
+```
+
+```
+% pandoc -f html -t html
+<p id="test" class="foo bar" data-custom="value">This is a paragraph with attributes.</p>
+^D
+<p id="test" class="foo bar" data-custom="value">This is a paragraph
+with attributes.</p>
+```
+
+```
+% pandoc -f html -t html5
+<p id="test" class="foo bar" data-custom="value">This is a paragraph with attributes.</p>
+^D
+<p id="test" class="foo bar" data-custom="value">This is a paragraph
+with attributes.</p>
+```