From 052684712b8a36854c622468f5d4ddd367199c1c Mon Sep 17 00:00:00 2001
From: John MacFarlane <jgm@berkeley.edu>
Date: Mon, 25 Mar 2019 08:43:59 -0700
Subject: HTML reader: read `data-foo` attribute into `foo`.

The HTML writer adds the `data-` prefix for HTML5
for nonstandard attributes.  But the attributes are
represented in the AST without the `data-` prefix,
so we should strip this when reading HTML.

Closes #5392.
---
 src/Text/Pandoc/Readers/HTML.hs | 3 ++-
 test/Tests/Readers/HTML.hs      | 2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/src/Text/Pandoc/Readers/HTML.hs b/src/Text/Pandoc/Readers/HTML.hs
index 45af20ca8..dfc31f7af 100644
--- a/src/Text/Pandoc/Readers/HTML.hs
+++ b/src/Text/Pandoc/Readers/HTML.hs
@@ -807,7 +807,8 @@ mathMLToTeXMath s = writeTeX <$> readMathML s
 
 toStringAttr :: [(Text, Text)] -> [(String, String)]
 toStringAttr = map go
-  where go (x,y) = (T.unpack x, T.unpack y)
+  where go (x,y) = (T.unpack (fromMaybe x $ T.stripPrefix "data-" x),
+                    T.unpack y)
 
 pScriptMath :: PandocMonad m => TagParser m Inlines
 pScriptMath = try $ do
diff --git a/test/Tests/Readers/HTML.hs b/test/Tests/Readers/HTML.hs
index 69078d955..be5178e07 100644
--- a/test/Tests/Readers/HTML.hs
+++ b/test/Tests/Readers/HTML.hs
@@ -83,7 +83,7 @@ tests = [ testGroup "base tag"
           , test htmlNativeDivs "<main role=X> becomes <div role=X>" $ "<main role=foobar>hello</main>" =?>
             doc (divWith ("", [], [("role", "foobar")]) (plain (text "hello")))
           , test htmlNativeDivs "<main> has attributes preserved" $ "<main id=foo class=bar data-baz=qux>hello</main>" =?>
-            doc (divWith ("foo", ["bar"], [("role", "main"), ("data-baz", "qux")]) (plain (text "hello")))
+            doc (divWith ("foo", ["bar"], [("role", "main"), ("baz", "qux")]) (plain (text "hello")))
           , test htmlNativeDivs "<main> closes <p>" $ "<p>hello<main>main content</main>" =?>
             doc (para (text "hello") <> divWith ("", [], [("role", "main")]) (plain (text "main content")))
           , test htmlNativeDivs "<main> followed by text" $ "<main>main content</main>non-main content" =?>
-- 
cgit v1.2.3