From 0640c4c9859aa5a3ede082c190fcd5883c24ac83 Mon Sep 17 00:00:00 2001
From: nibras shami <74427567+vreoo@users.noreply.github.com>
Date: Sun, 17 May 2026 12:02:34 +0300
Subject: [PATCH] EPUB writer: support multiple EPUB versions for raw content
 (#11628).

This change ensures that raw content marked `epub2` will appear in (only) EPUBv2 output
and content marked `epub3` will appear in (only) EPUBv3 output.
---
 src/Text/Pandoc/Writers/EPUB.hs | 39 +++++++++++++++++-------------
 src/Text/Pandoc/Writers/HTML.hs | 10 ++++++++
 test/command/8880.md            | 42 +++++++++++++++++++++++++++++++++
 3 files changed, 74 insertions(+), 17 deletions(-)
 create mode 100644 test/command/8880.md

diff --git a/src/Text/Pandoc/Writers/EPUB.hs b/src/Text/Pandoc/Writers/EPUB.hs
index 030270571..39d3f6274 100644
--- a/src/Text/Pandoc/Writers/EPUB.hs
+++ b/src/Text/Pandoc/Writers/EPUB.hs
@@ -466,8 +466,8 @@ pandocToEPUB :: PandocMonad m
 pandocToEPUB version opts doc = do
   let doc' = ensureValidXmlIdentifiers doc
   -- handle pictures
-  Pandoc meta blocks <- walkM (transformInline opts) doc' >>=
-                        walkM transformBlock
+  Pandoc meta blocks <- walkM (transformInline version opts) doc' >>=
+                        walkM (transformBlock version)
   picEntries <- mapMaybe (snd . snd) <$> gets stMediaPaths
 
   epubSubdir <- gets stEpubSubdir
@@ -1203,42 +1203,47 @@ getMediaNextNewName ext = do
   modify $ \st -> st { stMediaNextId = nextId + 1 }
   return $ "file" ++ show nextId ++ ext
 
-isHtmlFormat :: Format -> Bool
-isHtmlFormat (Format "html") = True
-isHtmlFormat (Format "html4") = True
-isHtmlFormat (Format "html5") = True
-isHtmlFormat _ = False
+isHtmlFormat :: EPUBVersion -> Format -> Bool
+isHtmlFormat _ (Format "html") = True
+isHtmlFormat _ (Format "html4") = True
+isHtmlFormat _ (Format "html5") = True
+isHtmlFormat _ (Format "epub") = True
+isHtmlFormat EPUB2 (Format "epub2") = True
+isHtmlFormat EPUB3 (Format "epub3") = True
+isHtmlFormat _ _ = False
 
 transformBlock  :: PandocMonad m
-                => Block
+                => EPUBVersion
+                -> Block
                 -> E m Block
-transformBlock (RawBlock fmt raw)
-  | isHtmlFormat fmt = do
+transformBlock version (RawBlock fmt raw)
+  | isHtmlFormat version fmt = do
   let tags = parseTags raw
   tags' <- mapM transformTag tags
   return $ RawBlock fmt (renderTags' tags')
-transformBlock b = return b
+transformBlock _ b = return b
 
 transformInline  :: PandocMonad m
-                 => WriterOptions
+                 => EPUBVersion
+                 -> WriterOptions
                  -> Inline
                  -> E m Inline
-transformInline _opts (Image attr@(_,_,kvs) lab (src,tit))
+transformInline _ _opts (Image attr@(_,_,kvs) lab (src,tit))
   | isNothing (lookup "external" kvs) = do
     newsrc <- modifyMediaRef $ T.unpack src
     return $ Image attr lab ("../" <> newsrc, tit)
-transformInline opts x@(Math t m)
+transformInline _ opts x@(Math t m)
   | WebTeX url <- writerHTMLMathMethod opts = do
     newsrc <- modifyMediaRef (T.unpack (url <> urlEncode m))
     let mathclass = if t == DisplayMath then "display" else "inline"
     return $ Span ("",["math",mathclass],[])
                 [Image nullAttr [x] ("../" <> newsrc, "")]
-transformInline _opts (RawInline fmt raw)
-  | isHtmlFormat fmt = do
+transformInline version _opts (RawInline fmt raw)
+  | isHtmlFormat version fmt = do
   let tags = parseTags raw
   tags' <- mapM transformTag tags
   return $ RawInline fmt (renderTags' tags')
-transformInline _ x = return x
+transformInline _ _ x = return x
 
 (!) :: (t -> Element) -> [(Text, Text)] -> t -> Element
 (!) f attrs n = add_attrs (map (\(k,v) -> Attr (unqual k) v) attrs) (f n)
diff --git a/src/Text/Pandoc/Writers/HTML.hs b/src/Text/Pandoc/Writers/HTML.hs
index 93473a327..0f1d378a9 100644
--- a/src/Text/Pandoc/Writers/HTML.hs
+++ b/src/Text/Pandoc/Writers/HTML.hs
@@ -1795,10 +1795,20 @@ intrinsicEventsHTML4 =
 isRawHtml :: PandocMonad m => Format -> StateT WriterState m Bool
 isRawHtml f = do
   html5 <- gets stHtml5
+  epubVersion <- gets stEPUBVersion
   return $ f == Format "html" ||
            ((html5 && f == Format "html5") || f == Format "html4") ||
+           isEpubFormat epubVersion f ||
            isSlideVariant f
 
+-- | Check to see if Format matches with an EPUB variant
+isEpubFormat :: Maybe EPUBVersion -> Format -> Bool
+isEpubFormat Nothing _ = False
+isEpubFormat (Just EPUB2) f =
+  f == Format "epub" || f == Format "epub2"
+isEpubFormat (Just EPUB3) f =
+  f == Format "epub" || f == Format "epub3"
+
 -- | Check to see if Format matches with an HTML slide variant
 isSlideVariant :: Format -> Bool
 isSlideVariant f = f `elem` [Format "s5", Format "slidy", Format "slideous",
diff --git a/test/command/8880.md b/test/command/8880.md
new file mode 100644
index 000000000..3aa9417b2
--- /dev/null
+++ b/test/command/8880.md
@@ -0,0 +1,42 @@
+Raw EPUB attributes are rendered in EPUB output.
+
+```
+% pandoc -f native -t epub --metadata title=Raw -o - | pandoc -f epub -t html
+[ RawBlock (Format "epub") "<p>ok</p>" ]
+^D
+<p><span id="title_page.xhtml"></span></p>
+<p><span id="ch001.xhtml"></span></p>
+<section id="ch001.xhtml_raw" class="unnumbered level1">
+<h1 class="unnumbered">Raw</h1>
+<p>ok</p>
+</section>
+```
+
+Raw EPUB2 attributes are omitted from EPUB3 output.
+
+```
+% pandoc -f markdown -t epub3 --metadata title=Raw -o - | pandoc -f epub -t html
+~~~ {=epub2}
+<p>ok</p>
+~~~
+^D
+<p><span id="title_page.xhtml"></span></p>
+<p><span id="ch001.xhtml"></span></p>
+<section id="ch001.xhtml_raw" class="unnumbered level1">
+<h1 class="unnumbered">Raw</h1>
+</section>
+```
+
+```
+% pandoc -f markdown -t epub3 --metadata title=Raw -o - | pandoc -f epub -t html
+~~~ {=epub3}
+<p>ok</p>
+~~~
+^D
+<p><span id="title_page.xhtml"></span></p>
+<p><span id="ch001.xhtml"></span></p>
+<section id="ch001.xhtml_raw" class="unnumbered level1">
+<h1 class="unnumbered">Raw</h1>
+<p>ok</p>
+</section>
+```