00d20cc · typeable/pandoc

git.iko.soy/typeable/pandoc

Org reader: allow changing emphasis syntax

The characters allowed before and after emphasis can be configured via `#+pandoc-emphasis-pre` and `#+pandoc-emphasis-post`, respectively. This allows to change which strings are recognized as emphasized text on a per-document or even per-paragraph basis. The allowed characters must be given as (Haskell) string. #+pandoc-emphasis-pre: "-\t ('\"{" #+pandoc-emphasis-post: "-\t\n .,:!?;'\")}[" If the argument cannot be read as a string, the default value is restored. Closes: #4378

commit00d20ccd09a8542fda631ab16c7f569098f2918d authorAlbert Krewinkel <albert@zeitkraut.de> dateWed 21 Feb 2018, 08:53 (9 years ago) parent84db7e4

4 files changed+58−12

compare with master

unifiedsplit

src/Text/Pandoc/Readers/Org/Inlines.hs+6−10 src/Text/Pandoc/Readers/Org/Meta.hs+22 src/Text/Pandoc/Readers/Org/ParserState.hs+11−1 test/Tests/Readers/Org/Meta.hs+19−1

src/Text/Pandoc/Readers/Org/Inlines.hs+6−10

@@ -603,6 +603,8 @@ updatePositions :: PandocMonad m

603603 => Char

604604 -> OrgParser m Char

605605updatePositions c = do

606+ st <- getState

607+ let emphasisPreChars = orgStateEmphasisPreChars st

606608 when (c `elem` emphasisPreChars) updateLastPreCharPos

607609 when (c `elem` emphasisForbiddenBorderChars) updateLastForbiddenCharPos

608610 return c

@@ -681,8 +683,10 @@ emphasisEnd c = try $ do

681683 updateLastStrPos

682684 popInlineCharStack

683685 return c

684− where acceptablePostChars =

685− surroundingEmphasisChar >>= \x -> oneOf (x ++ emphasisPostChars)

686+ where

687+ acceptablePostChars = do

688+ emphasisPostChars <- orgStateEmphasisPostChars <$> getState

689+ surroundingEmphasisChar >>= \x -> oneOf (x ++ emphasisPostChars)

686690

687691mathStart :: PandocMonad m => Char -> OrgParser m Char

688692mathStart c = try $

@@ -734,14 +738,6 @@ many1TillNOrLessNewlines n p end = try $

734738-- here (see, e.g., the Emacs Lisp variable `org-emphasis-regexp-components`

735739-- for details).

736740

737−-- | Chars allowed to occur before emphasis (spaces and newlines are ok, too)

738−emphasisPreChars :: [Char]

739−emphasisPreChars = "-\t ('\"{"

740−

741−-- | Chars allowed at after emphasis

742−emphasisPostChars :: [Char]

743−emphasisPostChars = "-\t\n .,:!?;'\")}["

744−

745741-- | Chars not allowed at the (inner) border of emphasis

746742emphasisForbiddenBorderChars :: [Char]

747743emphasisForbiddenBorderChars = "\t\n\r "

src/Text/Pandoc/Readers/Org/Meta.hs+22

@@ -43,6 +43,7 @@ import Text.Pandoc.Builder (Blocks, Inlines)

4343import qualified Text.Pandoc.Builder as B

4444import Text.Pandoc.Class (PandocMonad)

4545import Text.Pandoc.Definition

46+import Text.Pandoc.Shared (safeRead)

4647

4748import Control.Monad (mzero, void, when)

4849import Data.Char (toLower)

@@ -154,6 +155,8 @@ optionLine = try $ do

154155 "seq_todo" -> todoSequence >>= updateState . registerTodoSequence

155156 "typ_todo" -> todoSequence >>= updateState . registerTodoSequence

156157 "macro" -> macroDefinition >>= updateState . registerMacro

158+ "pandoc-emphasis-pre" -> emphChars >>= updateState . setEmphasisPreChar

159+ "pandoc-emphasis-post" -> emphChars >>= updateState . setEmphasisPostChar

157160 _ -> mzero

158161

159162addLinkFormat :: Monad m => String

@@ -184,6 +187,25 @@ parseFormat = try $ replacePlain <|> replaceUrl <|> justAppend

184187 rest = manyTill anyChar (eof <|> () <$ oneOf "\n\r")

185188 tillSpecifier c = manyTill (noneOf "\n\r") (try $ string ('%':c:""))

186189

190+setEmphasisPreChar :: Maybe [Char] -> OrgParserState -> OrgParserState

191+setEmphasisPreChar csMb st =

192+ let preChars = case csMb of

193+ Nothing -> orgStateEmphasisPreChars defaultOrgParserState

194+ Just cs -> cs

195+ in st { orgStateEmphasisPreChars = preChars }

196+

197+setEmphasisPostChar :: Maybe [Char] -> OrgParserState -> OrgParserState

198+setEmphasisPostChar csMb st =

199+ let postChars = case csMb of

200+ Nothing -> orgStateEmphasisPostChars defaultOrgParserState

201+ Just cs -> cs

202+ in st { orgStateEmphasisPostChars = postChars }

203+

204+emphChars :: Monad m => OrgParser m (Maybe [Char])

205+emphChars = do

206+ skipSpaces

207+ safeRead <$> anyLine

208+

187209inlinesTillNewline :: PandocMonad m => OrgParser m (F Inlines)

188210inlinesTillNewline = do

189211 updateLastPreCharPos

src/Text/Pandoc/Readers/Org/ParserState.hs+11−1

@@ -29,6 +29,7 @@ Define the Org-mode parser state.

2929-}

3030module Text.Pandoc.Readers.Org.ParserState

3131 ( OrgParserState (..)

32+ , defaultOrgParserState

3233 , OrgParserLocal (..)

3334 , OrgNoteRecord

3435 , HasReaderOptions (..)

@@ -104,6 +105,11 @@ type TodoSequence = [TodoMarker]

104105data OrgParserState = OrgParserState

105106 { orgStateAnchorIds :: [String]

106107 , orgStateEmphasisCharStack :: [Char]

108+ , orgStateEmphasisPreChars :: [Char] -- ^ Chars allowed to occur before

109+ -- emphasis; spaces and newlines are

110+ -- always ok in addition to what is

111+ -- specified here.

112+ , orgStateEmphasisPostChars :: [Char] -- ^ Chars allowed at after emphasis

107113 , orgStateEmphasisNewlines :: Maybe Int

108114 , orgStateExportSettings :: ExportSettings

109115 , orgStateHeaderMap :: M.Map Inlines String

@@ -124,7 +130,9 @@ data OrgParserState = OrgParserState

124130 , orgMacros :: M.Map Text Macro

125131 }

126132

127−data OrgParserLocal = OrgParserLocal { orgLocalQuoteContext :: QuoteContext }

133+data OrgParserLocal = OrgParserLocal

134+ { orgLocalQuoteContext :: QuoteContext

135+ }

128136

129137instance Default OrgParserLocal where

130138 def = OrgParserLocal NoQuote

@@ -168,6 +176,8 @@ instance Default OrgParserState where

168176defaultOrgParserState :: OrgParserState

169177defaultOrgParserState = OrgParserState

170178 { orgStateAnchorIds = []

179+ , orgStateEmphasisPreChars = "-\t ('\"{"

180+ , orgStateEmphasisPostChars = "-\t\n .,:!?;'\")}["

171181 , orgStateEmphasisCharStack = []

172182 , orgStateEmphasisNewlines = Nothing

173183 , orgStateExportSettings = def

test/Tests/Readers/Org/Meta.hs+19−1

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

11{-# LANGUAGE OverloadedStrings #-}

22module Tests.Readers.Org.Meta (tests) where

4−import Test.Tasty (TestTree)

4+import Test.Tasty (TestTree, testGroup)

55import Tests.Helpers ((=?>))

66import Tests.Readers.Org.Shared ((=:), spcSep)

77import Text.Pandoc

@@ -170,4 +170,22 @@ tests =

170170 , "[[expl:foo][bar]]"

171171 ] =?>

172172 para (link "http://example.com/foo" "" "bar")

173+

174+ , testGroup "emphasis config"

175+ [ "Changing pre and post chars for emphasis" =:

176+ T.unlines [ "#+pandoc-emphasis-pre: \"[)\""

177+ , "#+pandoc-emphasis-post: \"]\\n\""

178+ , "([/emph/])*foo*"

179+ ] =?>

180+ para ("([" <> emph "emph" <> "])" <> strong "foo")

181+

182+ , "setting an invalid value restores the default" =:

183+ T.unlines [ "#+pandoc-emphasis-pre: \"[\""

184+ , "#+pandoc-emphasis-post: \"]\""

185+ , "#+pandoc-emphasis-pre:"

186+ , "#+pandoc-emphasis-post:"

187+ , "[/noemph/]"

188+ ] =?>

189+ para ("[/noemph/]")

190+ ]

173191 ]