0558ea9 · typeable/pandoc

git.iko.soy/typeable/pandoc

Org reader: modify handling of example blocks. (#5717)

* Org reader: allow the `-i` switch to ignore leading spaces. * Org reader: handle awkwardly-aligned code blocks within lists. Code blocks in Org lists must have their #+BEGIN_ aligned in a reasonable way, but their other components can be positioned otherwise.

commit0558ea9836f5350a00f3df19d55ca6bddda1947b authorBrian Leung <29217594+leungbk@users.noreply.github.com> dateMon 9 Sep 2019, 07:34 (7 years ago) parent8f5ab97

3 files changed+103−14

compare with master

unifiedsplit

src/Text/Pandoc/Readers/Org/Blocks.hs+41−14 src/Text/Pandoc/Readers/Org/ParserState.hs+2 test/command/4186.md+60

src/Text/Pandoc/Readers/Org/Blocks.hs+41−14

@@ -186,7 +186,7 @@ orgBlock = try $ do

186186 "html" -> rawBlockLines (return . B.rawBlock (lowercase blkType))

187187 "latex" -> rawBlockLines (return . B.rawBlock (lowercase blkType))

188188 "ascii" -> rawBlockLines (return . B.rawBlock (lowercase blkType))

189− "example" -> rawBlockLines (return . exampleCode)

189+ "example" -> exampleBlock blockAttrs

190190 "quote" -> parseBlockLines (fmap B.blockQuote)

191191 "verse" -> verseBlock

192192 "src" -> codeBlock blockAttrs

@@ -200,6 +200,16 @@ orgBlock = try $ do

200200 lowercase :: String -> String

201201 lowercase = map toLower

202202

203+exampleBlock :: PandocMonad m => BlockAttributes -> String -> OrgParser m (F Blocks)

204+exampleBlock blockAttrs _label = do

205+ skipSpaces

206+ (classes, kv) <- switchesAsAttributes

207+ newline

208+ content <- rawBlockContent "example"

209+ let id' = fromMaybe mempty $ blockAttrName blockAttrs

210+ let codeBlck = B.codeBlockWith (id', "example":classes, kv) content

211+ return . return $ codeBlck

212+

203213rawBlockLines :: Monad m => (String -> F Blocks) -> String -> OrgParser m (F Blocks)

204214rawBlockLines f blockType = ignHeaders *> (f <$> rawBlockContent blockType)

205215

@@ -216,11 +226,13 @@ rawBlockContent :: Monad m => String -> OrgParser m String

216226rawBlockContent blockType = try $ do

217227 blkLines <- manyTill rawLine blockEnder

218228 tabLen <- getOption readerTabStop

219− return

220− . unlines

221− . stripIndent

222− . map (tabsToSpaces tabLen . commaEscaped)

223− $ blkLines

229+ trimP <- orgStateTrimLeadBlkIndent <$> getState

230+ let stripIndent strs = if trimP then map (drop (shortestIndent strs)) strs else strs

231+ (unlines

232+ . stripIndent

233+ . map (tabsToSpaces tabLen . commaEscaped)

234+ $ blkLines)

235+ <$ updateState (\s -> s { orgStateTrimLeadBlkIndent = True })

224236 where

225237 rawLine :: Monad m => OrgParser m String

226238 rawLine = try $ ("" <$ blankline) <|> anyLine

@@ -228,9 +240,6 @@ rawBlockContent blockType = try $ do

228240 blockEnder :: Monad m => OrgParser m ()

229241 blockEnder = try $ skipSpaces <* stringAnyCase ("#+end_" <> blockType)

230242

231− stripIndent :: [String] -> [String]

232− stripIndent strs = map (drop (shortestIndent strs)) strs

233−

234243 shortestIndent :: [String] -> Int

235244 shortestIndent = foldr (min . length . takeWhile isSpace) maxBound

236245 . filter (not . null)

@@ -357,12 +366,19 @@ switchPolarity = (SwitchMinus <$ char '-') <|> (SwitchPlus <$ char '+')

357366

358367-- | Parses a source block switch option.

359368switch :: Monad m => OrgParser m (Char, Maybe String, SwitchPolarity)

360−switch = try $ lineNumberSwitch <|> labelSwitch <|> simpleSwitch

369+switch = try $ lineNumberSwitch <|> labelSwitch

370+ <|> whitespaceSwitch <|> simpleSwitch

361371 where

362372 simpleSwitch = (\pol c -> (c, Nothing, pol)) <$> switchPolarity <*> letter

363373 labelSwitch = genericSwitch 'l' $

364374 char '"' *> many1Till nonspaceChar (char '"')

365375

376+whitespaceSwitch :: Monad m => OrgParser m (Char, Maybe String, SwitchPolarity)

377+whitespaceSwitch = do

378+ string "-i"

379+ updateState $ \s -> s { orgStateTrimLeadBlkIndent = False }

380+ return ('i', Nothing, SwitchMinus)

381+

366382-- | Generic source block switch-option parser.

367383genericSwitch :: Monad m

368384 => Char

@@ -821,11 +837,22 @@ listItem parseIndentedMarker = try . withContext ListItemState $ do

821837

822838-- continuation of a list item - indented and separated by blankline or endline.

823839-- Note: nested lists are parsed as continuations.

824−listContinuation :: Monad m => Int

825− -> OrgParser m String

840+listContinuation :: PandocMonad m => Int -> OrgParser m String

826841listContinuation markerLength = try $ do

827842 notFollowedBy' blankline

828− mappend <$> (concat <$> many1 listLine)

843+ mappend <$> (concat <$> many1 (listContinuation' markerLength))

829844 <*> many blankline

830845 where

831− listLine = try $ indentWith markerLength *> anyLineNewline

846+ listContinuation' indentation =

847+ blockLines indentation <|> listLine indentation

848+ listLine indentation = try $ indentWith indentation *> anyLineNewline

849+ -- The block attributes and start must be appropriately indented,

850+ -- but the contents, and end do not.

851+ blockLines indentation =

852+ try $ lookAhead (indentWith indentation

853+ >> blockAttributes

854+ >>= (\blockAttrs ->

855+ case attrFromBlockAttributes blockAttrs of

856+ ("", [], []) -> count 1 anyChar

857+ _ -> indentWith indentation))

858+ >> (snd <$> withRaw orgBlock)

src/Text/Pandoc/Readers/Org/ParserState.hs+2

@@ -117,6 +117,7 @@ data OrgParserState = OrgParserState

117117 , orgStateSelectTags :: Set.Set Tag

118118 , orgStateSelectTagsChanged :: Bool

119119 , orgStateTodoSequences :: [TodoSequence]

120+ , orgStateTrimLeadBlkIndent :: Bool

120121 , orgLogMessages :: [LogMessage]

121122 , orgMacros :: M.Map Text Macro

122123 }

@@ -184,6 +185,7 @@ defaultOrgParserState = OrgParserState

184185 , orgStateParserContext = NullState

185186 , orgStateSelectTags = Set.singleton $ Tag "export"

186187 , orgStateSelectTagsChanged = False

188+ , orgStateTrimLeadBlkIndent = True

187189 , orgStateTodoSequences = []

188190 , orgLogMessages = []

189191 , orgMacros = M.empty

test/command/4186.mdnew file+60

@@ -0,0 +1,60 @@

1+```

2+% pandoc -f org -t native

3+#+BEGIN_EXAMPLE -i

4+ This should retain the four leading spaces

5+#+END_EXAMPLE

6+^D

7+[CodeBlock ("",["example"],[]) " This should retain the four leading spaces\n"]

8+```

10+```

11+% pandoc -f org -t html

12+- depth 1

13+ #+NAME: bob

14+ #+BEGIN_EXAMPLE -i

15+ Vertical alignment is four spaces beyond the appearance of the word "depth".

16+ #+END_EXAMPLE

17+ - depth 2

18+ #+begin_example

19+ Vertically aligned with the second appearance of the word "depth".

20+ #+end_example

21+ #+begin_example -i

22+ Vertical alignment is four spaces beyond the second

23+ appearance of the word "depth".

24+ The "begin" portion is a component of

25+ this deeper list element, so that guarantees

26+ that the entire block must be a component of the

27+ inner list element.

28+ #+end_example

29+ Still inside the inner list element

30+ #+NAME: carrie

31+ #+BEGIN_EXAMPLE

32+ This belongs to the outer list element, and is aligned accordingly, since the NAME attribute is not indented deeply enough. It is not enough for the BEGIN alone to be aligned deeply if the block is meant to have a NAME.

33+ #+END_EXAMPLE

34+ Still in the shallower list element since the preceding example

35+ block forced the deeper list element to terminate.

36+Outside all lists.

37+^D

38+<ul>

39+<li>depth 1

40+<pre id="bob" class="example"><code> Vertical alignment is four spaces beyond the appearance of the word "depth".

41+</code></pre>

42+<ul>

43+<li>depth 2

44+<pre class="example"><code>Vertically aligned with the second appearance of the word "depth".

45+</code></pre>

46+<pre class="example"><code> Vertical alignment is four spaces beyond the second

47+ appearance of the word "depth".

48+ The "begin" portion is a component of

49+ this deeper list element, so that guarantees

50+ that the entire block must be a component of the

51+ inner list element.

52+</code></pre>

53+Still inside the inner list element</li>

54+</ul>

55+<pre id="carrie" class="example"><code>This belongs to the outer list element, and is aligned accordingly, since the NAME attribute is not indented deeply enough. It is not enough for the BEGIN alone to be aligned deeply if the block is meant to have a NAME.

56+</code></pre>

57+Still in the shallower list element since the preceding example block forced the deeper list element to terminate.</li>

58+</ul>

59+Outside all lists.

60+```