Commonmark reader: parse HTML as plain text if -raw_html.

changelog

@@ -7,6 +7,14 @@ pandoc (2.1.2)
     interfered with list parsing, so now we require the fences to be flush with
     the margin of the containing block.
 
+  * Commonmark reader:
+
+    + `raw_html` is now on by default.  It can be disabled explicitly
+      using `-f commonmark-raw_html`.
+    + When `raw_html` is disabled, HTML is parsed as regular text,
+      rather than disappearing altogether.  This matches the behavior
+      of `markdown`.
+
   * Org reader (Albert Krewinkel):
 
     + Move citation tests to separate module.

src/Text/Pandoc/Readers/CommonMark.hs

@@ -109,7 +109,7 @@ addBlock opts (Node _ BLOCK_QUOTE nodes) =
   (BlockQuote (addBlocks opts nodes) :)
 addBlock opts (Node _ (HTML_BLOCK t) _)
   | isEnabled Ext_raw_html opts = (RawBlock (Format "html") (unpack t) :)
-  | otherwise                 = id
+  | otherwise                   = (Para [Str (unpack t)] :)
 -- Note:  the cmark parser will never generate CUSTOM_BLOCK,
 -- so we don't need to handle it:
 addBlock _ (Node _ (CUSTOM_BLOCK _onEnter _onExit) _nodes) =
@@ -215,7 +215,7 @@ addInline opts (Node _ SOFTBREAK _)
   | otherwise                           = (SoftBreak :)
 addInline opts (Node _ (HTML_INLINE t) _)
   | isEnabled Ext_raw_html opts = (RawInline (Format "html") (unpack t) :)
-  | otherwise                 = id
+  | otherwise                   = (Str (unpack t) :)
 -- Note:  the cmark parser will never generate CUSTOM_BLOCK,
 -- so we don't need to handle it:
 addInline _ (Node _ (CUSTOM_INLINE _onEnter _onExit) _nodes) =