ee0fc19 · typeable/pandoc

git.iko.soy/typeable/pandoc

Parsing: Further improvements to uri parser.

Don't treat punctuation before percent-encoding as final punctuation. Don't treat '+' as final punctuation.

commitee0fc19bc54208c5d8828eab872e3bbe303c47bf authorJohn MacFarlane <fiddlosopher@gmail.com> dateThu 28 Mar 2013, 11:33 (14 years ago) parente81b87c

1 file changed+4−2

unifiedsplit

src/Text/Pandoc/Parsing.hs+4−2

@@ -425,7 +425,8 @@ uri = try $ do

425425 -- http://en.wikipedia.org/wiki/State_of_emergency_(disambiguation)

426426 -- as a URL, while NOT picking up the closing paren in

427427 -- (http://wikipedia.org). So we include balanced parens in the URL.

428− let isWordChar c = isAlphaNum c || c == '_' || c == '/' || not (isAscii c)

428+ let isWordChar c = isAlphaNum c || c == '_' || c == '/' || c == '+' ||

429+ not (isAscii c)

429430 let wordChar = satisfy isWordChar

430431 let percentEscaped = try $ char '%' >> skipMany1 (satisfy isHexDigit)

431432 let entity = () <$ characterReference

@@ -434,7 +435,8 @@ uri = try $ do

434435 let uriChunk = skipMany1 wordChar

435436 <|> percentEscaped

436437 <|> entity

437− <|> (try $ punct >> lookAhead (satisfy isWordChar) >> return ())

438+ <|> (try $ punct >>

439+ lookAhead (void (satisfy isWordChar) <|> percentEscaped))

438440 str <- snd `fmap` withRaw (skipMany1 ( () <$

439441 (enclosed (char '(') (char ')') uriChunk

440442 <|> enclosed (char '{') (char '}') uriChunk