From dd970030ba5be6aac9bf831ea08df39b1d4f874c Mon Sep 17 00:00:00 2001
From: John MacFarlane <jgm@berkeley.edu>
Date: Mon, 17 Aug 2026 09:27:38 -0700
Subject: [PATCH] HTML reader: handle pre without code.

We preserve whitespace as nonbreaking spaces.

Closes #11810.
---
 src/Text/Pandoc/Readers/HTML.hs       | 31 +++++++++++++++++++++------
 src/Text/Pandoc/Readers/HTML/Types.hs |  1 +
 test/command/11810.md                 | 15 +++++++++++++
 3 files changed, 41 insertions(+), 6 deletions(-)
 create mode 100644 test/command/11810.md

diff --git a/src/Text/Pandoc/Readers/HTML.hs b/src/Text/Pandoc/Readers/HTML.hs
index 1baeda555..9e57f95da 100644
--- a/src/Text/Pandoc/Readers/HTML.hs
+++ b/src/Text/Pandoc/Readers/HTML.hs
@@ -92,7 +92,7 @@ readHtml opts inp = do
   result <- flip runReaderT def $
        runParserT parseDoc
        (HTMLState def{ stateOptions = opts }
-         [] Nothing Set.empty [] M.empty opts False)
+         [] Nothing Set.empty [] M.empty opts False False)
        "source" tags
   case result of
     Right doc -> return doc
@@ -222,7 +222,7 @@ block = ((do
         "h5" -> pHeader
         "h6" -> pHeader
         "blockquote" -> pBlockQuote
-        "pre" -> pCodeBlock
+        "pre" -> pCodeBlock <|> pPreBlock
         "ul" -> pBulletList
         "ol" -> pOrderedList
         "dl" -> pDefinitionList
@@ -646,6 +646,15 @@ pFigure = do
                         (B.simpleCaption (mconcat captions))
                         (mconcat rest)
 
+pPreBlock :: PandocMonad m => TagParser m Blocks
+pPreBlock = try $ do
+  pSatisfy (matchTagOpen "pre" [])
+  oldInPre <- inPre <$> getState
+  updateState $ \st -> st{ inPre = True }
+  contents <- mconcat <$> manyTill block (pCloses "pre" <|> eof)
+  updateState $ \st -> st{ inPre = oldInPre }
+  return contents
+
 pCodeBlock :: PandocMonad m => TagParser m Blocks
 pCodeBlock = try $ do
   TagOpen _ attr' <- pSatisfy (matchTagOpen "pre" [])
@@ -1062,10 +1071,20 @@ pBad = do
   return $ B.str $ T.singleton c'
 
 pSpace :: PandocMonad m => InlinesParser m Inlines
-pSpace = many1 (satisfy isSpace) >>= \xs ->
-            if '\n' `elem` xs
-               then return B.softbreak
-               else return B.space
+pSpace = do
+  inpre <- inPre <$> getState
+  xs <- many1 (satisfy isSpace)
+  if inpre
+     then return $ makePreInlines xs
+     else if '\n' `elem` xs
+             then return B.softbreak
+             else return B.space
+ where
+  makePreInlines cs =
+    let chunks = splitWhen (=='\n') cs
+        tostr = B.str . T.pack . map (\c -> if c == ' ' then '\160' else c)
+    in  mconcat $ filter (/= B.str "")
+                $ L.intersperse B.linebreak (map tostr chunks)
 
 getTagName :: Tag Text -> Maybe Text
 getTagName (TagOpen t _) = Just t
diff --git a/src/Text/Pandoc/Readers/HTML/Types.hs b/src/Text/Pandoc/Readers/HTML/Types.hs
index 28798de41..9398767e5 100644
--- a/src/Text/Pandoc/Readers/HTML/Types.hs
+++ b/src/Text/Pandoc/Readers/HTML/Types.hs
@@ -53,6 +53,7 @@ data HTMLState = HTMLState
   , macros      :: Map Text Macro
   , readerOpts  :: ReaderOptions
   , inFootnotes :: Bool
+  , inPre       :: Bool
   }
 
 -- | Local HTML parser state
diff --git a/test/command/11810.md b/test/command/11810.md
new file mode 100644
index 000000000..36bf5246e
--- /dev/null
+++ b/test/command/11810.md
@@ -0,0 +1,15 @@
+```
+% pandoc -f html -t native
+<pre>alpha
+    beta
+        gamma</pre>
+^D
+[ Plain
+    [ Str "alpha"
+    , LineBreak
+    , Str "\160\160\160\160beta"
+    , LineBreak
+    , Str "\160\160\160\160\160\160\160\160gamma"
+    ]
+]
+```