From ed3d46638425825de30aaa3d1152b9343292c315 Mon Sep 17 00:00:00 2001 From: John MacFarlane Date: Wed, 1 Nov 2017 09:27:51 -0700 Subject: Really fix #3989. The previous fix only worked in certain cases. Other cases with `>` in an HTML attribute broke. --- src/Text/Pandoc/Readers/HTML.hs | 17 ++++++++++++----- test/command/3989.md | 2 +- 2 files changed, 13 insertions(+), 6 deletions(-) diff --git a/src/Text/Pandoc/Readers/HTML.hs b/src/Text/Pandoc/Readers/HTML.hs index e2be1c5bd..2b667c63c 100644 --- a/src/Text/Pandoc/Readers/HTML.hs +++ b/src/Text/Pandoc/Readers/HTML.hs @@ -1133,6 +1133,7 @@ htmlTag :: (HasReaderOptions st, Monad m) -> ParserT [Char] st m (Tag String, String) htmlTag f = try $ do lookAhead (char '<') + startpos <- getPosition inp <- getInput let ts = canonicalizeTags $ parseTagsOptions parseOptions{ optTagWarning = False @@ -1153,11 +1154,17 @@ htmlTag f = try $ do [] -> False (c:cs) -> isLetter c && all isNameChar cs - let endAngle = try $ do char '>' - pos <- getPosition - guard $ (sourceLine pos == ln && - sourceColumn pos >= col) || - sourceLine pos > ln + let endpos = if ln == 1 + then setSourceColumn startpos + (sourceColumn startpos + (col - 1)) + else setSourceColumn (setSourceLine startpos + (sourceLine startpos + (ln - 1))) + col + let endAngle = try $ + do char '>' + pos <- getPosition + guard $ pos >= endpos + let handleTag tagname = do -- basic sanity check, since the parser is very forgiving -- and finds tags in stuff like xfoo -foo + foo ^D [Para [Span ("",[],[("title","1st line of text
2nd line of text")]) [Str "foo"],SoftBreak,Span ("",[],[("title","1st line of text
2nd line of text")]) [Str "foo"]]] ``` -- cgit v1.2.3