bf72a48 · typeable/pandoc

git.iko.soy/typeable/pandoc

on master help sign in

Tighten up parsing of raw email addresses.

bf72a48John MacFarlane10 years ago

src/Text/Pandoc/Parsing.hs+13−4

compare with master

unifiedsplit

@@ -178,7 +178,7 @@ import qualified Text.Pandoc.UTF8 as UTF8 (putStrLn)

178import Text.Parsec hiding (token)

179import Text.Parsec.Pos (newPos)

180import Data.Char ( toLower, toUpper, ord, chr, isAscii, isAlphaNum,

181− isHexDigit, isSpace )

181+ isHexDigit, isSpace, isPunctuation )

182import Data.List ( intercalate, transpose, isSuffixOf )

183import Text.Pandoc.Shared

184import qualified Data.Map as M

@@ -405,9 +405,18 @@ emailAddress = try $ toResult <$> mailbox <*> (char '@' *> domain)

405 domain = intercalate "." <$> (subdomain `sepby1` dot)

406 dot = char '.'

407 subdomain = many1 $ alphaNum <|> innerPunct

408+ -- this excludes some valid email addresses, since an

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

410+ -- for our purposes, when combined with markdown parsing:

408− innerPunct = try (satisfy (\c -> isEmailPunct c || c == '@') <*

411+ innerPunct = try (satisfy (\c -> isEmailPunct c || c == '@')

409− notFollowedBy space)

412+ <* notFollowedBy space

413+ <* notFollowedBy (satisfy isPunctuation))

414+ -- technically an email address could begin with a symbol,

415+ -- but allowing this creates too many problems.

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

417+ emailWord = do x <- satisfy isAlphaNum

410− emailWord = many1 $ satisfy isEmailChar

418+ xs <- many (satisfy isEmailChar)

419+ return (x:xs)

411 isEmailChar c = isAlphaNum c || isEmailPunct c

420 isEmailChar c = isAlphaNum c || isEmailPunct c

412 isEmailPunct c = c `elem` "!\"#$%&'*+-/=?^_{|}~;"

421 isEmailPunct c = c `elem` "!\"#$%&'*+-/=?^_{|}~;"

413 -- note: sepBy1 from parsec consumes input when sep

422 -- note: sepBy1 from parsec consumes input when sep

tests/Tests/Readers/Markdown.hs+5

@@ -185,6 +185,11 @@ tests = [ testGroup "inline code"

185 "<\n\na>" =?>

186 para (text "<") <> para (text "a>")

187 ]

188+ , testGroup "raw email addresses"

189+ [ test markdownGH "issue 2940" $

190+ "**@user**" =?>

191+ para (strong (text "@user"))

192+ ]

188 , testGroup "emoji"

193 , testGroup "emoji"

189 [ test markdownGH "emoji symbols" $

194 [ test markdownGH "emoji symbols" $

190 ":smile: and :+1:" =?> para (text "😄 and 👍")

195 ":smile: and :+1:" =?> para (text "😄 and 👍")