From 971e4c43647cf28d1de0dbc109a6ec6f269fa563 Mon Sep 17 00:00:00 2001
From: John MacFarlane <fiddlosopher@gmail.com>
Date: Sat, 5 Apr 2014 23:05:11 -0700
Subject: [PATCH] HTML reader: Updated `closes` with rules from HTML5 spec.
---
src/Text/Pandoc/Readers/HTML.hs | 17 ++++++++++++-----
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/src/Text/Pandoc/Readers/HTML.hs b/src/Text/Pandoc/Readers/HTML.hs
index 38619d040..2101b2fc2 100644
--- a/src/Text/Pandoc/Readers/HTML.hs
+++ b/src/Text/Pandoc/Readers/HTML.hs
@@ -469,7 +469,7 @@ pCloses :: String -> TagParser ()
pCloses tagtype = try $ do
t <- lookAhead $ pSatisfy $ \tag -> isTagClose tag || isTagOpen tag
case t of
- (TagClose t') | t' == tagtype -> pAnyTag >> return ()
+ (TagClose t') | t' == tagtype -> pAnyTag >> return ()
(TagOpen t' _) | t' `closes` tagtype -> return ()
(TagClose "ul") | tagtype == "li" -> return ()
(TagClose "ol") | tagtype == "li" -> return ()
@@ -627,7 +627,7 @@ isCommentTag :: Tag String -> Bool
isCommentTag = tagComment (const True)
-- taken from HXT and extended
-
+-- See http://www.w3.org/TR/html5/syntax.html sec 8.1.2.4 optional tags
closes :: String -> String -> Bool
_ `closes` "body" = False
_ `closes` "html" = False
@@ -635,11 +635,18 @@ _ `closes` "html" = False
"li" `closes` "li" = True
"th" `closes` t | t `elem` ["th","td"] = True
"tr" `closes` t | t `elem` ["th","td","tr"] = True
+"dd" `closes` t | t `elem` ["dt", "dd"] = True
"dt" `closes` t | t `elem` ["dt","dd"] = True
-"hr" `closes` "p" = True
-"p" `closes` "p" = True
+"rt" `closes` t | t `elem` ["rb", "rt", "rtc"] = True
+"optgroup" `closes` "optgroup" = True
+"optgroup" `closes` "option" = True
+"option" `closes` "option" = True
+-- http://www.w3.org/TR/html-markup/p.html
+x `closes` "p" | x `elem` ["address", "article", "aside", "blockquote",
+ "dir", "div", "dl", "fieldset", "footer", "form", "h1", "h2", "h3", "h4",
+ "h5", "h6", "header", "hr", "menu", "nav", "ol", "p", "pre", "section",
+ "table", "ul"] = True
"meta" `closes` "meta" = True
-"colgroup" `closes` "colgroup" = True
"form" `closes` "form" = True
"label" `closes` "label" = True
"map" `closes` "map" = True