ed3d466 · typeable/pandoc

git.iko.soy/typeable/pandoc

Really fix #3989.

The previous fix only worked in certain cases. Other cases with `>` in an HTML attribute broke.

commited3d46638425825de30aaa3d1152b9343292c315 authorJohn MacFarlane <jgm@berkeley.edu> dateWed 1 Nov 2017, 09:27 (9 years ago) parentf1ebdb8

2 files changed+13−6

compare with master

unifiedsplit

src/Text/Pandoc/Readers/HTML.hs+12−5 test/command/3989.md+1−1

src/Text/Pandoc/Readers/HTML.hs+12−5

@@ -1133,6 +1133,7 @@ htmlTag :: (HasReaderOptions st, Monad m)

11331133 -> ParserT [Char] st m (Tag String, String)

11341134htmlTag f = try $ do

11351135 lookAhead (char '<')

1136+ startpos <- getPosition

11361137 inp <- getInput

11371138 let ts = canonicalizeTags $ parseTagsOptions

11381139 parseOptions{ optTagWarning = False

@@ -1153,11 +1154,17 @@ htmlTag f = try $ do

11531154 [] -> False

11541155 (c:cs) -> isLetter c && all isNameChar cs

11551156

1156− let endAngle = try $ do char '>'

1157− pos <- getPosition

1158− guard $ (sourceLine pos == ln &&

1159− sourceColumn pos >= col) ||

1160− sourceLine pos > ln

1157+ let endpos = if ln == 1

1158+ then setSourceColumn startpos

1159+ (sourceColumn startpos + (col - 1))

1160+ else setSourceColumn (setSourceLine startpos

1161+ (sourceLine startpos + (ln - 1)))

1162+ col

1163+ let endAngle = try $

1164+ do char '>'

1165+ pos <- getPosition

1166+ guard $ pos >= endpos

1167+

11611168 let handleTag tagname = do

11621169 -- basic sanity check, since the parser is very forgiving

11631170 -- and finds tags in stuff like x<y)

test/command/3989.md+1−1

@@ -1,7 +1,7 @@

11```

22pandoc -f markdown -t native

33 2nd line of text">foo

4− 2nd line of text">foo

4+ 2nd line of text">foo

55^D

66[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"]]]

77```