1e8a25a · typeable/pandoc

git.iko.soy/typeable/pandoc

Percent-encode more special characters in URLs.

HTML, LaTeX writers adjusted. The special characters are '<','>','|','"','{','}','[',']','^', '`'. Closes #1640, #2377.

commit1e8a25ad69db8fa372ac4ddbb5ac05ffb00ed052 authorJohn MacFarlane <jgm@berkeley.edu> dateSun 11 Oct 2015, 17:06 (11 years ago) parent04307a1

3 files changed+14−9

compare with master

unifiedsplit

src/Text/Pandoc/Shared.hs+5−2 src/Text/Pandoc/Writers/LaTeX.hs+5−5 tests/Tests/Readers/Markdown.hs+4−2

src/Text/Pandoc/Shared.hs+5−2

@@ -288,9 +288,12 @@ toRomanNumeral x =

288288 _ | x >= 1 -> "I" ++ toRomanNumeral (x - 1)

289289 _ -> ""

290290

291−-- | Escape whitespace in URI.

291+-- | Escape whitespace and some punctuation characters in URI.

292292escapeURI :: String -> String

293−escapeURI = escapeURIString (not . isSpace)

293+escapeURI = escapeURIString (not . needsEscaping)

294+ where needsEscaping c = isSpace c || c `elem`

295+ ['<','>','|','"','{','}','[',']','^', '`']

296+

294297

295298-- | Convert tabs to spaces and filter out DOS line endings.

296299-- Tabs will be preserved if tab stop is set to 0.

src/Text/Pandoc/Writers/LaTeX.hs+5−5

@@ -242,7 +242,7 @@ stringToLaTeX ctx (x:xs) = do

242242 '^' -> "\\^{}" ++ rest

243243 '\\'| isUrl -> '/' : rest -- NB. / works as path sep even on Windows

244244 | otherwise -> "\\textbackslash{}" ++ rest

245− '|' -> "\\textbar{}" ++ rest

245+ '|' | not isUrl -> "\\textbar{}" ++ rest

246246 '<' -> "\\textless{}" ++ rest

247247 '>' -> "\\textgreater{}" ++ rest

248248 '[' -> "{[}" ++ rest -- to avoid interpretation as

@@ -848,17 +848,17 @@ inlineToLaTeX (Link txt (src, _)) =

848848 case txt of

849849 [Str x] | escapeURI x == src -> -- autolink

850850 do modify $ \s -> s{ stUrl = True }

851− src' <- stringToLaTeX URLString src

851+ src' <- stringToLaTeX URLString (escapeURI src)

852852 return $ text $ "\\url{" ++ src' ++ "}"

853853 [Str x] | Just rest <- stripPrefix "mailto:" src,

854854 escapeURI x == rest -> -- email autolink

855855 do modify $ \s -> s{ stUrl = True }

856− src' <- stringToLaTeX URLString src

856+ src' <- stringToLaTeX URLString (escapeURI src)

857857 contents <- inlineListToLaTeX txt

858858 return $ "\\href" <> braces (text src') <>

859859 braces ("\\nolinkurl" <> braces contents)

860860 _ -> do contents <- inlineListToLaTeX txt

861− src' <- stringToLaTeX URLString src

861+ src' <- stringToLaTeX URLString (escapeURI src)

862862 return $ text ("\\href{" ++ src' ++ "}{") <>

863863 contents <> char '}'

864864inlineToLaTeX (Image _ (source, _)) = do

@@ -866,7 +866,7 @@ inlineToLaTeX (Image _ (source, _)) = do

866866 let source' = if isURI source

867867 then source

868868 else unEscapeString source

869− source'' <- stringToLaTeX URLString source'

869+ source'' <- stringToLaTeX URLString (escapeURI source')

870870 inHeading <- gets stInHeading

871871 return $

872872 (if inHeading then "\\protect\\includegraphics" else "\\includegraphics")

tests/Tests/Readers/Markdown.hs+4−2

@@ -66,9 +66,11 @@ bareLinkTests =

6666 , ("http://en.wikipedia.org/wiki/Sprite_(computer_graphics)",

6767 autolink "http://en.wikipedia.org/wiki/Sprite_(computer_graphics)")

6868 , ("http://en.wikipedia.org/wiki/Sprite_[computer_graphics]",

69− autolink "http://en.wikipedia.org/wiki/Sprite_[computer_graphics]")

69+ link "http://en.wikipedia.org/wiki/Sprite_%5Bcomputer_graphics%5D" ""

70+ (str "http://en.wikipedia.org/wiki/Sprite_[computer_graphics]"))

7071 , ("http://en.wikipedia.org/wiki/Sprite_{computer_graphics}",

71− autolink "http://en.wikipedia.org/wiki/Sprite_{computer_graphics}")

72+ link "http://en.wikipedia.org/wiki/Sprite_%7Bcomputer_graphics%7D" ""

73+ (str "http://en.wikipedia.org/wiki/Sprite_{computer_graphics}"))

7274 , ("http://example.com/Notification_Center-GitHub-20101108-140050.jpg",

7375 autolink "http://example.com/Notification_Center-GitHub-20101108-140050.jpg")

7476 , ("https://github.com/github/hubot/blob/master/scripts/cream.js#L20-20",