HTML reader: better handle KaTeX-generated math.

KaTeX emits the mathml followed by a span with an HTML fallback.
Previously pandoc was converting both.  We now ignore the HTML
fallback span, marked with class `katex-html`.

Closes #9971.

1 files changed, 20 insertions, 18 deletions

diff --git a/src/Text/Pandoc/Readers/HTML.hs b/src/Text/Pandoc/Readers/HTML.hs
index 5ee990db4..6fcf574c4 100644
--- a/src/Text/Pandoc/Readers/HTML.hs
+++ b/src/Text/Pandoc/Readers/HTML.hs
@@ -512,7 +512,7 @@ pIframe = try $ do
                | "image/" `T.isPrefixOf` mt -> do
                     return $ B.divWith ("",["iframe"],[]) $
                       B.plain $ B.image url "" mempty
-             _ -> return $ B.divWith ("",["iframe"],[("src", url)]) $ mempty)
+             _ -> return $ B.divWith ("",["iframe"],[("src", url)]) mempty)
        (\e -> do
          logMessage $ CouldNotFetchResource url (renderError e)
          ignore $ renderTags' [tag, TagClose "iframe"])
@@ -852,23 +852,25 @@ pSpan :: PandocMonad m => TagParser m Inlines
 pSpan = do
   (TagOpen _ attr') <- lookAhead (pSatisfy $ tagOpen (=="span") (const True))
   exts <- getOption readerExtensions
-  if extensionEnabled Ext_native_spans exts
-     then do
-       contents <- pInTags "span" inline
-       let attr = toAttr attr'
-       let classes = maybe [] T.words $ lookup "class" attr'
-       let styleAttr   = fromMaybe "" $ lookup "style" attr'
-       let fontVariant = fromMaybe "" $
-                          pickStyleAttrProps ["font-variant"] styleAttr
-       let isSmallCaps = fontVariant == "small-caps" ||
-                           "smallcaps" `elem` classes
-       let tag = if isSmallCaps then B.smallcaps else B.spanWith attr
-       return $ tag contents
-     else if extensionEnabled Ext_raw_html exts
-             then do
-               tag <- pSatisfy $ tagOpen (=="span") (const True)
-               return $ B.rawInline "html" $ renderTags' [tag]
-             else pInTags "span" inline -- just contents
+  let attr = toAttr attr'
+  case attr of
+     (_,["katex-html"],_) -> mempty <$ pInTags "span" inline
+       -- skip HTML generated by KaTeX, since we get
+       -- the math by parsing mathml (#9971)
+     _ | extensionEnabled Ext_native_spans exts -> do
+           contents <- pInTags "span" inline
+           let classes = maybe [] T.words $ lookup "class" attr'
+           let styleAttr   = fromMaybe "" $ lookup "style" attr'
+           let fontVariant = fromMaybe "" $
+                              pickStyleAttrProps ["font-variant"] styleAttr
+           let isSmallCaps = fontVariant == "small-caps" ||
+                               "smallcaps" `elem` classes
+           let tag = if isSmallCaps then B.smallcaps else B.spanWith attr
+           return $ tag contents
+       | extensionEnabled Ext_raw_html exts -> do
+            tag <- pSatisfy $ tagOpen (=="span") (const True)
+            return $ B.rawInline "html" $ renderTags' [tag]
+       | otherwise  -> pInTags "span" inline -- just contents
 
 pRawHtmlInline :: PandocMonad m => TagParser m Inlines
 pRawHtmlInline = do