136bf90 · typeable/pandoc

git.iko.soy/typeable/pandoc

Markdown reader: distinguish autolinks in the AST.

With this change, autolinks are parsed as Links with the `uri` class. (The same is true for bare links, if the `autolink_bare_uris` extension is enabled.) Email autolinks are parsed as Links with the `email` class. This allows the distinction to be represented in the URI. Formerly the `uri` class was added to autolinks by the HTML writer, but it had to guess what was an autolink and could not distinguish `[http://example.com](http://example.com)` from `<http://example.com>`. It also incorrectly recognized `[pandoc](pandoc)` as an autolink. Now the HTML writer simply passes through the `uri` attribute if it is present, but does not add anything. The Textile writer has been modified so that the `uri` class is not explicitly added for autolinks, even if it is present. Closes #4913.

commit136bf901aa088eaf4e5c996c71e0a36c171f1587 authorJohn MacFarlane <jgm@berkeley.edu> dateWed 19 Sep 2018, 14:49 (8 years ago) parentdba5c8d

14 files changed+73−38

compare with master

unifiedsplit

src/Text/Pandoc/Readers/Markdown.hs+7−5 src/Text/Pandoc/Writers/HTML.hs+2−5 src/Text/Pandoc/Writers/Textile.hs+3−3 test/Tests/Readers/Markdown.hs+5−3 test/command/3716.md+1−1 test/command/4913.md+34 test/markdown-reader-more.native+4−4 test/testsuite.native+4−4 test/writer.docbook4+3−3 test/writer.docbook5+3−3 test/writer.html4+1−1 test/writer.html5+1−1 test/writer.native+4−4 test/writer.textile+1−1

src/Text/Pandoc/Readers/Markdown.hs+7−5

@@ -1,6 +1,7 @@

11{-# LANGUAGE NoImplicitPrelude #-}

22{-# LANGUAGE RelaxedPolyRec #-}

33{-# LANGUAGE ScopedTypeVariables #-}

4+{-# LANGUAGE TupleSections #-}

56{-

@@ -1879,23 +1880,24 @@ bareURL :: PandocMonad m => MarkdownParser m (F Inlines)

18791880bareURL = try $ do

18801881 guardEnabled Ext_autolink_bare_uris

18811882 getState >>= guard . stateAllowLinks

1882− (orig, src) <- uri <|> emailAddress

1883+ (cls, (orig, src)) <- (("uri",) <$> uri) <|> (("email",) <$> emailAddress)

18831884 notFollowedBy $ try $ spaces >> htmlTag (~== TagClose "a")

1884− return $ return $ B.link src "" (B.str orig)

1885+ return $ return $ B.linkWith ("",[cls],[]) src "" (B.str orig)

18851886

18861887autoLink :: PandocMonad m => MarkdownParser m (F Inlines)

18871888autoLink = try $ do

18881889 getState >>= guard . stateAllowLinks

18891890 char '<'

1890− (orig, src) <- uri <|> emailAddress

1891+ (cls, (orig, src)) <- (("uri",) <$> uri) <|> (("email",) <$> emailAddress)

18911892 -- in rare cases, something may remain after the uri parser

18921893 -- is finished, because the uri parser tries to avoid parsing

18931894 -- final punctuation. for example: in `<http://hi---there>`,

18941895 -- the URI parser will stop before the dashes.

18951896 extra <- fromEntities <$> manyTill nonspaceChar (char '>')

1896− attr <- option nullAttr $ try $

1897+ attr <- option ("", [cls], []) $ try $

18971898 guardEnabled Ext_link_attributes >> attributes

1898− return $ return $ B.linkWith attr (src ++ escapeURI extra) "" (B.str $ orig ++ extra)

1899+ return $ return $ B.linkWith attr (src ++ escapeURI extra) ""

1900+ (B.str $ orig ++ extra)

18991901

19001902image :: PandocMonad m => MarkdownParser m (F Inlines)

19011903image = try $ do

src/Text/Pandoc/Writers/HTML.hs+2−5

@@ -56,7 +56,7 @@ import Data.String (fromString)

5656import Data.Text (Text)

5757import qualified Data.Text.Lazy as TL

5858import Network.HTTP (urlEncode)

59−import Network.URI (URI (..), parseURIReference, unEscapeString)

59+import Network.URI (URI (..), parseURIReference)

6060import Numeric (showHex)

6161import Text.Blaze.Internal (customLeaf, customParent, MarkupM(Empty))

6262#if MIN_VERSION_blaze_markup(0,6,3)

@@ -1084,10 +1084,7 @@ inlineToHtml opts inline = do

10841084 in '#' : prefix ++ xs

10851085 _ -> s

10861086 let link = H.a ! A.href (toValue s') $ linkText

1087− let attr = if txt == [Str (unEscapeString s)]

1088− then (ident, "uri" : classes, kvs)

1089− else (ident, classes, kvs)

1090− link' <- addAttrs opts attr link

1087+ link' <- addAttrs opts (ident, classes, kvs) link

10911088 return $ if null tit

10921089 then link'

10931090 else link' ! A.title (toValue tit)

src/Text/Pandoc/Writers/Textile.hs+3−3

@@ -463,15 +463,15 @@ inlineToTextile _ SoftBreak = return " "

463463inlineToTextile _ Space = return " "

464464

465465inlineToTextile opts (Link (_, cls, _) txt (src, _)) = do

466− let classes = if null cls

467− then ""

468− else "(" ++ unwords cls ++ ")"

469466 label <- case txt of

470467 [Code _ s]

471468 | s == src -> return "$"

472469 [Str s]

473470 | s == src -> return "$"

474471 _ -> inlineListToTextile opts txt

472+ let classes = if null cls || cls == ["uri"] && label == "$"

473+ then ""

474+ else "(" ++ unwords cls ++ ")"

475475 return $ "\"" ++ classes ++ label ++ "\":" ++ src

476476

477477inlineToTextile opts (Image attr@(_, cls, _) alt (source, tit)) = do

test/Tests/Readers/Markdown.hs+5−3

@@ -39,7 +39,7 @@ testBareLink (inp, ils) =

3939 (unpack inp) (inp, doc $ para ils)

4040

4141autolink :: String -> Inlines

42−autolink = autolinkWith nullAttr

42+autolink = autolinkWith ("",["uri"],[])

4343

4444autolinkWith :: Attr -> String -> Inlines

4545autolinkWith attr s = linkWith attr s "" (str s)

@@ -72,10 +72,12 @@ bareLinkTests =

7272 , ("http://en.wikipedia.org/wiki/Sprite_(computer_graphics)",

7373 autolink "http://en.wikipedia.org/wiki/Sprite_(computer_graphics)")

7474 , ("http://en.wikipedia.org/wiki/Sprite_[computer_graphics]",

75− link "http://en.wikipedia.org/wiki/Sprite_%5Bcomputer_graphics%5D" ""

75+ linkWith ("",["uri"],[])

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

7677 (str "http://en.wikipedia.org/wiki/Sprite_[computer_graphics]"))

7778 , ("http://en.wikipedia.org/wiki/Sprite_{computer_graphics}",

78− link "http://en.wikipedia.org/wiki/Sprite_%7Bcomputer_graphics%7D" ""

79+ linkWith ("",["uri"],[])

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

7981 (str "http://en.wikipedia.org/wiki/Sprite_{computer_graphics}"))

8082 , ("http://example.com/Notification_Center-GitHub-20101108-140050.jpg",

8183 autolink "http://example.com/Notification_Center-GitHub-20101108-140050.jpg")

test/command/3716.md+1−1

@@ -2,5 +2,5 @@

22% pandoc

33<http://example.com>{.foo}

44^D

5−<a href="http://example.com" class="uri foo">http://example.com</a>

5+<a href="http://example.com" class="foo">http://example.com</a>

66```

test/command/4913.mdnew file+34

@@ -0,0 +1,34 @@

1+```

2+% pandoc -f markdown -t html

3+[https://pandoc.org](https://pandoc.org)

4+^D

5+<a href="https://pandoc.org">https://pandoc.org</a>

6+```

8+```

9+% pandoc -f markdown -t markdown

10+[https://pandoc.org](https://pandoc.org)

11+^D

12+<https://pandoc.org>

13+```

14+

15+```

16+% pandoc -f markdown -t html

17+<https://pandoc.org>

18+^D

19+<a href="https://pandoc.org" class="uri">https://pandoc.org</a>

20+```

21+

22+```

23+% pandoc -f markdown -t html

24+<https://pandoc.org>{.foo}

25+^D

26+<a href="https://pandoc.org" class="foo">https://pandoc.org</a>

27+```

28+

29+```

30+% pandoc -f markdown -t html

31+<me@example.com>

32+^D

33+<a href="mailto:me@example.com" class="email">me@example.com</a>

34+```

test/markdown-reader-more.native+4−4

@@ -45,9 +45,9 @@ Pandoc (Meta {unMeta = fromList [("author",MetaList [MetaInlines [Str "Author",S

4545,Para [Str "`hi"]

4646,Para [Str "there`"]

4747,Header 2 ("multilingual-urls",[],[]) [Str "Multilingual",Space,Str "URLs"]

48−,Para [Link ("",[],[]) [Str "http://\27979.com?\27979=\27979"] ("http://\27979.com?\27979=\27979","")]

48+,Para [Link ("",["uri"],[]) [Str "http://\27979.com?\27979=\27979"] ("http://\27979.com?\27979=\27979","")]

4949,Para [Link ("",[],[]) [Str "foo"] ("/bar/\27979?x=\27979","title")]

50−,Para [Link ("",[],[]) [Str "\27979@foo.\27979.baz"] ("mailto:\27979@foo.\27979.baz","")]

50+,Para [Link ("",["email"],[]) [Str "\27979@foo.\27979.baz"] ("mailto:\27979@foo.\27979.baz","")]

5151,Header 2 ("numbered-examples",[],[]) [Str "Numbered",Space,Str "examples"]

5252,OrderedList (1,Example,TwoParens)

5353 [[Plain [Str "First",Space,Str "example."]]

@@ -176,8 +176,8 @@ Pandoc (Meta {unMeta = fromList [("author",MetaList [MetaInlines [Str "Author",S

176176 ,[]]]

177177,Header 2 ("entities-in-links-and-titles",[],[]) [Str "Entities",Space,Str "in",Space,Str "links",Space,Str "and",Space,Str "titles"]

178178,Para [Link ("",[],[]) [Str "link"] ("/\252rl","\246\246!")]

179−,Para [Link ("",[],[]) [Str "http://g\246\246gle.com"] ("http://g\246\246gle.com","")]

180−,Para [Link ("",[],[]) [Str "me@ex\228mple.com"] ("mailto:me@ex\228mple.com","")]

179+,Para [Link ("",["uri"],[]) [Str "http://g\246\246gle.com"] ("http://g\246\246gle.com","")]

180+,Para [Link ("",["email"],[]) [Str "me@ex\228mple.com"] ("mailto:me@ex\228mple.com","")]

181181,Para [Link ("",[],[]) [Str "foobar"] ("/\252rl","\246\246!")]

182182,Header 2 ("parentheses-in-urls",[],[]) [Str "Parentheses",Space,Str "in",Space,Str "URLs"]

183183,Para [Link ("",[],[]) [Str "link"] ("/hi(there)","")]

test/testsuite.native+4−4

@@ -384,14 +384,14 @@ Pandoc (Meta {unMeta = fromList [("author",MetaList [MetaInlines [Str "John",Spa

384384,Para [Str "Here\8217s",Space,Str "an",Space,Link ("",[],[]) [Str "inline",Space,Str "link"] ("/script?foo=1&bar=2",""),Str "."]

385385,Para [Str "Here\8217s",Space,Str "an",Space,Link ("",[],[]) [Str "inline",Space,Str "link",Space,Str "in",Space,Str "pointy",Space,Str "braces"] ("/script?foo=1&bar=2",""),Str "."]

386386,Header 2 ("autolinks",[],[]) [Str "Autolinks"]

387−,Para [Str "With",Space,Str "an",Space,Str "ampersand:",Space,Link ("",[],[]) [Str "http://example.com/?foo=1&bar=2"] ("http://example.com/?foo=1&bar=2","")]

387+,Para [Str "With",Space,Str "an",Space,Str "ampersand:",Space,Link ("",["uri"],[]) [Str "http://example.com/?foo=1&bar=2"] ("http://example.com/?foo=1&bar=2","")]

388388,BulletList

389389 [[Plain [Str "In",Space,Str "a",Space,Str "list?"]]

390− ,[Plain [Link ("",[],[]) [Str "http://example.com/"] ("http://example.com/","")]]

390+ ,[Plain [Link ("",["uri"],[]) [Str "http://example.com/"] ("http://example.com/","")]]

391391 ,[Plain [Str "It",Space,Str "should."]]]

392−,Para [Str "An",Space,Str "e-mail",Space,Str "address:",Space,Link ("",[],[]) [Str "nobody@nowhere.net"] ("mailto:nobody@nowhere.net","")]

392+,Para [Str "An",Space,Str "e-mail",Space,Str "address:",Space,Link ("",["email"],[]) [Str "nobody@nowhere.net"] ("mailto:nobody@nowhere.net","")]

393393,BlockQuote

394− [Para [Str "Blockquoted:",Space,Link ("",[],[]) [Str "http://example.com/"] ("http://example.com/","")]]

394+ [Para [Str "Blockquoted:",Space,Link ("",["uri"],[]) [Str "http://example.com/"] ("http://example.com/","")]]

395395,Para [Str "Auto-links",Space,Str "should",Space,Str "not",Space,Str "occur",Space,Str "here:",Space,Code ("",[],[]) "<http://example.com/>"]

396396,CodeBlock ("",[],[]) "or here: <http://example.com/>"

397397,HorizontalRule

test/writer.docbook4+3−3

@@ -1298,7 +1298,7 @@ These should not be escaped: \$ \\ \> \[ \{

12981298 <title>Autolinks</title>

12991299 <para>

13001300 With an ampersand:

1301− <ulink url="http://example.com/?foo=1&bar=2">http://example.com/?foo=1&bar=2</ulink>

1301+ <ulink url="http://example.com/?foo=1&bar=2" role="uri">http://example.com/?foo=1&bar=2</ulink>

13021302 </para>

13031303 <itemizedlist spacing="compact">

13041304 <listitem>

@@ -1308,7 +1308,7 @@ These should not be escaped: \$ \\ \> \[ \{

13081308 </listitem>

13091309 <listitem>

13101310 <para>

1311− <ulink url="http://example.com/">http://example.com/</ulink>

1311+ <ulink url="http://example.com/" role="uri">http://example.com/</ulink>

13121312 </para>

13131313 </listitem>

13141314 <listitem>

@@ -1323,7 +1323,7 @@ These should not be escaped: \$ \\ \> \[ \{

13231323 <blockquote>

13241324 <para>

13251325 Blockquoted:

1326− <ulink url="http://example.com/">http://example.com/</ulink>

1326+ <ulink url="http://example.com/" role="uri">http://example.com/</ulink>

13271327 </para>

13281328 </blockquote>

13291329 <para>

test/writer.docbook5+3−3

@@ -1273,7 +1273,7 @@ These should not be escaped: \$ \\ \> \[ \{

12731273 <title>Autolinks</title>

12741274 <para>

12751275 With an ampersand:

1276− <link xlink:href="http://example.com/?foo=1&bar=2">http://example.com/?foo=1&bar=2</link>

1276+ <link xlink:href="http://example.com/?foo=1&bar=2" role="uri">http://example.com/?foo=1&bar=2</link>

12771277 </para>

12781278 <itemizedlist spacing="compact">

12791279 <listitem>

@@ -1283,7 +1283,7 @@ These should not be escaped: \$ \\ \> \[ \{

12831283 </listitem>

12841284 <listitem>

12851285 <para>

1286− <link xlink:href="http://example.com/">http://example.com/</link>

1286+ <link xlink:href="http://example.com/" role="uri">http://example.com/</link>

12871287 </para>

12881288 </listitem>

12891289 <listitem>

@@ -1298,7 +1298,7 @@ These should not be escaped: \$ \\ \> \[ \{

12981298 <blockquote>

12991299 <para>

13001300 Blockquoted:

1301− <link xlink:href="http://example.com/">http://example.com/</link>

1301+ <link xlink:href="http://example.com/" role="uri">http://example.com/</link>

13021302 </para>

13031303 </blockquote>

13041304 <para>

test/writer.html4+1−1

@@ -508,7 +508,7 @@ Blah

508508<li><a href="http://example.com/" class="uri">http://example.com/</a></li>

509509<li>It should.</li>

510510</ul>

511−An e-mail address: <a href="mailto:nobody@nowhere.net">nobody@nowhere.net</a>

511+An e-mail address: <a href="mailto:nobody@nowhere.net" class="email">nobody@nowhere.net</a>

512512<blockquote>

513513Blockquoted: <a href="http://example.com/" class="uri">http://example.com/</a>

514514</blockquote>

test/writer.html5+1−1

@@ -511,7 +511,7 @@ Blah

511511<li><a href="http://example.com/" class="uri">http://example.com/</a></li>

512512<li>It should.</li>

513513</ul>

514−An e-mail address: <a href="mailto:nobody@nowhere.net">nobody@nowhere.net</a>

514+An e-mail address: <a href="mailto:nobody@nowhere.net" class="email">nobody@nowhere.net</a>

515515<blockquote>

516516Blockquoted: <a href="http://example.com/" class="uri">http://example.com/</a>

517517</blockquote>

test/writer.native+4−4

@@ -384,14 +384,14 @@ Pandoc (Meta {unMeta = fromList [("author",MetaList [MetaInlines [Str "John",Spa

384384,Para [Str "Here\8217s",Space,Str "an",Space,Link ("",[],[]) [Str "inline",Space,Str "link"] ("/script?foo=1&bar=2",""),Str "."]

385385,Para [Str "Here\8217s",Space,Str "an",Space,Link ("",[],[]) [Str "inline",Space,Str "link",Space,Str "in",Space,Str "pointy",Space,Str "braces"] ("/script?foo=1&bar=2",""),Str "."]

386386,Header 2 ("autolinks",[],[]) [Str "Autolinks"]

387−,Para [Str "With",Space,Str "an",Space,Str "ampersand:",Space,Link ("",[],[]) [Str "http://example.com/?foo=1&bar=2"] ("http://example.com/?foo=1&bar=2","")]

387+,Para [Str "With",Space,Str "an",Space,Str "ampersand:",Space,Link ("",["uri"],[]) [Str "http://example.com/?foo=1&bar=2"] ("http://example.com/?foo=1&bar=2","")]

388388,BulletList

389389 [[Plain [Str "In",Space,Str "a",Space,Str "list?"]]

390− ,[Plain [Link ("",[],[]) [Str "http://example.com/"] ("http://example.com/","")]]

390+ ,[Plain [Link ("",["uri"],[]) [Str "http://example.com/"] ("http://example.com/","")]]

391391 ,[Plain [Str "It",Space,Str "should."]]]

392−,Para [Str "An",Space,Str "e-mail",Space,Str "address:",Space,Link ("",[],[]) [Str "nobody@nowhere.net"] ("mailto:nobody@nowhere.net","")]

392+,Para [Str "An",Space,Str "e-mail",Space,Str "address:",Space,Link ("",["email"],[]) [Str "nobody@nowhere.net"] ("mailto:nobody@nowhere.net","")]

393393,BlockQuote

394− [Para [Str "Blockquoted:",Space,Link ("",[],[]) [Str "http://example.com/"] ("http://example.com/","")]]

394+ [Para [Str "Blockquoted:",Space,Link ("",["uri"],[]) [Str "http://example.com/"] ("http://example.com/","")]]

395395,Para [Str "Auto-links",Space,Str "should",Space,Str "not",Space,Str "occur",Space,Str "here:",Space,Code ("",[],[]) "<http://example.com/>"]

396396,CodeBlock ("",[],[]) "or here: <http://example.com/>"

397397,HorizontalRule

test/writer.textile+1−1

@@ -660,7 +660,7 @@ With an ampersand: "$":http://example.com/?foo=1&bar=2

660660* "$":http://example.com/

661661* It should.

662662

663−An e-mail address: "nobody@nowhere.net":mailto:nobody@nowhere.net

663+An e-mail address: "(email)nobody@nowhere.net":mailto:nobody@nowhere.net

664664

665665bq. Blockquoted: "$":http://example.com/

666666