From 5314df51f368853f7ee4bf8f75eedf43afc0400e Mon Sep 17 00:00:00 2001
From: John MacFarlane <jgm@berkeley.edu>
Date: Wed, 4 Dec 2013 10:00:40 -0800
Subject: [PATCH] Stop parsing "list lines" when we hit a block tag.
This fixes exponential slowdown in certain input, e.g.
a series of lists followed by `</div>`.
---
src/Text/Pandoc/Readers/Markdown.hs | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/Text/Pandoc/Readers/Markdown.hs b/src/Text/Pandoc/Readers/Markdown.hs
index 33d1a9620..4cb75d86c 100644
--- a/src/Text/Pandoc/Readers/Markdown.hs
+++ b/src/Text/Pandoc/Readers/Markdown.hs
@@ -732,7 +732,8 @@ listLine = try $ do
notFollowedBy' (do indentSpaces
many (spaceChar)
listStart)
- chunks <- manyTill (liftM snd (htmlTag isCommentTag) <|> count 1 anyChar) newline
+ chunks <- manyTill (liftM snd (htmlTag isCommentTag) <|> count 1 (satisfy (/='<'))
+ <|> (notFollowedBy' (htmlTag isBlockTag) >> count 1 anyChar)) newline
return $ concat chunks
-- parse raw text for one list item, excluding start marker and continuations
@@ -759,6 +760,7 @@ listContinuationLine :: MarkdownParser String
listContinuationLine = try $ do
notFollowedBy blankline
notFollowedBy' listStart
+ notFollowedBy' $ try $ skipMany spaceChar >> htmlTag (~== TagClose "div")
optional indentSpaces
result <- anyLine
return $ result ++ "\n"