0352f78 · typeable/pandoc

git.iko.soy/typeable/pandoc

Improve emailAddress in Text.Pandoc.Parsing.

Previously the parser would accept characters in domains that are illegal in domains, and this sometimes caused it to gobble bits of the following text. Closes #7398. Note that this change, by itself, caused some txt2tag reader tests to fail. txt2tags allows bare email addresses with a following form query. So, in addition to the change to emailAddress, we modify the txt2tags parser so it can still handle these cases.

commit0352f7845bfa2053797850c3639414978285b63e authorJohn MacFarlane <jgm@berkeley.edu> dateMon 21 Jun 2021, 22:35 (5 years ago) parent2ef2049

2 files changed+24−5

compare with master

unifiedsplit

src/Text/Pandoc/Parsing.hs+3−4 src/Text/Pandoc/Readers/Txt2Tags.hs+21−1

src/Text/Pandoc/Parsing.hs+3−4

@@ -693,13 +693,12 @@ emailAddress = try $ toResult <$> mailbox <*> (char '@' *> domain)

693693 mailbox = intercalate "." <$> (emailWord `sepBy1'` dot)

694694 domain = intercalate "." <$> (subdomain `sepBy1'` dot)

695695 dot = char '.'

696− subdomain = many1 $ alphaNum <|> innerPunct

696+ subdomain = many1 $ alphaNum <|> innerPunct (=='-')

697697 -- this excludes some valid email addresses, since an

698698 -- email could contain e.g. '__', but gives better results

699699 -- for our purposes, when combined with markdown parsing:

700− innerPunct = try (satisfy (\c -> isEmailPunct c || c == '@')

701− <* notFollowedBy space

702− <* notFollowedBy (satisfy isPunctuation))

700+ innerPunct f = try (satisfy f

701+ <* notFollowedBy (satisfy (not . isAlphaNum)))

703702 -- technically an email address could begin with a symbol,

704703 -- but allowing this creates too many problems.

705704 -- See e.g. https://github.com/jgm/pandoc/issues/2940

src/Text/Pandoc/Readers/Txt2Tags.hs+21−1

@@ -478,9 +478,29 @@ macro = try $ do

478478-- raw URLs in text are automatically linked

479479url :: T2T Inlines

480480url = try $ do

481− (rawUrl, escapedUrl) <- try uri <|> emailAddress

481+ (rawUrl, escapedUrl) <- try uri <|> emailAddress'

482482 return $ B.link rawUrl "" (B.str escapedUrl)

483483

484+emailAddress' :: T2T (Text, Text)

485+emailAddress' = do

486+ (base, mailURI) <- emailAddress

487+ query <- option "" emailQuery

488+ return (base <> query, mailURI <> query)

489+

490+emailQuery :: T2T Text

491+emailQuery = do

492+ char '?'

493+ parts <- kv `sepBy1` (char '&')

494+ return $ "?" <> T.intercalate "&" parts

495+

496+kv :: T2T Text

497+kv = do

498+ k <- T.pack <$> many1 alphaNum

499+ char '='

500+ let vchar = alphaNum <|> try (oneOf "%._/~:,=$@&+-;*" <* lookAhead alphaNum)

501+ v <- T.pack <$> many1 vchar

502+ return (k <> "=" <> v)

503+

484504uri :: T2T (Text, Text)

485505uri = try $ do

486506 address <- t2tURI