7a602d2 · typeable/pandoc

git.iko.soy/typeable/pandoc

Limit nesting of strong/emph.

This avoids exponential lookahead in parasitic cases, like a**a*a**a*a**a*a**a*a**a*a**a*a**a*a**. Added stateMaxNestingLevel to ParserState. We set this to 6, so you can still have Emph inside Emph, just not indefinitely.

commit7a602d222f6ba005a901d6b3322f9392602d6b9d authorJohn MacFarlane <fiddlosopher@gmail.com> dateTue 7 Feb 2012, 21:50 (15 years ago) parent12aa83f

2 files changed+16−2

compare with master

unifiedsplit

src/Text/Pandoc/Parsing.hs+2 src/Text/Pandoc/Readers/Markdown.hs+14−2

src/Text/Pandoc/Parsing.hs+2

@@ -629,6 +629,7 @@ data ParserState = ParserState

629629 { stateParseRaw :: Bool, -- ^ Parse raw HTML and LaTeX?

630630 stateParserContext :: ParserContext, -- ^ Inside list?

631631 stateQuoteContext :: QuoteContext, -- ^ Inside quoted environment?

632+ stateMaxNestingLevel :: Int, -- ^ Max # of nested Strong/Emph

632633 stateLastStrPos :: Maybe SourcePos, -- ^ Position after last str parsed

633634 stateKeys :: KeyTable, -- ^ List of reference keys

634635 stateCitations :: [String], -- ^ List of available citations

@@ -660,6 +661,7 @@ defaultParserState =

660661 ParserState { stateParseRaw = False,

661662 stateParserContext = NullState,

662663 stateQuoteContext = NoQuote,

664+ stateMaxNestingLevel = 6,

663665 stateLastStrPos = Nothing,

664666 stateKeys = M.empty,

665667 stateCitations = [],

src/Text/Pandoc/Readers/Markdown.hs+14−2

@@ -1041,8 +1041,20 @@ inlinesBetween start end =

10411041 where inner = innerSpace <|> (notFollowedBy' whitespace >> inline)

10421042 innerSpace = try $ whitespace >>~ notFollowedBy' end

10431043

1044+-- This is used to prevent exponential blowups for things like:

1045+-- a**a*a**a*a**a*a**a*a**a*a**a*a**

1046+nested :: GenParser Char ParserState a

1047+ -> GenParser Char ParserState a

1048+nested p = do

1049+ nestlevel <- stateMaxNestingLevel `fmap` getState

1050+ guard $ nestlevel > 0

1051+ updateState $ \st -> st{ stateMaxNestingLevel = stateMaxNestingLevel st - 1 }

1052+ res <- p

1053+ updateState $ \st -> st{ stateMaxNestingLevel = nestlevel }

1054+ return res

1055+

10441056emph :: GenParser Char ParserState Inline

1045−emph = Emph `liftM`

1057+emph = Emph `fmap` nested

10461058 (inlinesBetween starStart starEnd <|> inlinesBetween ulStart ulEnd)

10471059 where starStart = char '*' >> lookAhead nonspaceChar

10481060 starEnd = notFollowedBy' strong >> char '*'

@@ -1050,7 +1062,7 @@ emph = Emph `liftM`

10501062 ulEnd = notFollowedBy' strong >> char '_'

10511063

10521064strong :: GenParser Char ParserState Inline

1053−strong = Strong `liftM`

1065+strong = Strong `liftM` nested

10541066 (inlinesBetween starStart starEnd <|> inlinesBetween ulStart ulEnd)

10551067 where starStart = string "**" >> lookAhead nonspaceChar

10561068 starEnd = try $ string "**"