From 46d8b42da5d42b981bc56c49416cec15abaf2d09 Mon Sep 17 00:00:00 2001
From: John MacFarlane <jgm@berkeley.edu>
Date: Sun, 2 Oct 2016 21:51:34 +0200
Subject: [PATCH] AsciiDoc writer: avoid unnecessary use of "unconstrained"
emphasis.
In AsciiDoc, you must use a special form of emphasis (double `__`)
for intraword emphasis. Pandoc was previously using this more
than necessary.
Closes #3068.
---
src/Text/Pandoc/Writers/AsciiDoc.hs | 24 +++++++++++++++++-------
tests/writer.asciidoc | 8 ++++----
2 files changed, 21 insertions(+), 11 deletions(-)
diff --git a/src/Text/Pandoc/Writers/AsciiDoc.hs b/src/Text/Pandoc/Writers/AsciiDoc.hs
index 4ac6aa093..0dfbd705e 100644
--- a/src/Text/Pandoc/Writers/AsciiDoc.hs
+++ b/src/Text/Pandoc/Writers/AsciiDoc.hs
@@ -51,6 +51,7 @@ import Control.Monad.State
import qualified Data.Map as M
import Data.Aeson (Value(String), fromJSON, toJSON, Result(..))
import qualified Data.Text as T
+import Data.Char (isSpace, isPunctuation)
data WriterState = WriterState { defListMarker :: String
, orderedListLevel :: Int
@@ -321,6 +322,8 @@ blockListToAsciiDoc :: WriterOptions -- ^ Options
-> State WriterState Doc
blockListToAsciiDoc opts blocks = cat `fmap` mapM (blockToAsciiDoc opts) blocks
+data SpacyLocation = End | Start
+
-- | Convert list of Pandoc inline elements to asciidoc.
inlineListToAsciiDoc :: WriterOptions -> [Inline] -> State WriterState Doc
inlineListToAsciiDoc opts lst = do
@@ -331,14 +334,14 @@ inlineListToAsciiDoc opts lst = do
return result
where go [] = return empty
go (y:x:xs)
- | not (isSpacy y) = do
- y' <- if isSpacy x
+ | not (isSpacy End y) = do
+ y' <- if isSpacy Start x
then inlineToAsciiDoc opts y
else withIntraword $ inlineToAsciiDoc opts y
x' <- withIntraword $ inlineToAsciiDoc opts x
xs' <- go xs
return (y' <> x' <> xs')
- | not (isSpacy x) = do
+ | not (isSpacy Start x) = do
y' <- withIntraword $ inlineToAsciiDoc opts y
xs' <- go (x:xs)
return (y' <> xs')
@@ -346,10 +349,17 @@ inlineListToAsciiDoc opts lst = do
x' <- inlineToAsciiDoc opts x
xs' <- go xs
return (x' <> xs')
- isSpacy Space = True
- isSpacy LineBreak = True
- isSpacy SoftBreak = True
- isSpacy _ = False
+ isSpacy :: SpacyLocation -> Inline -> Bool
+ isSpacy _ Space = True
+ isSpacy _ LineBreak = True
+ isSpacy _ SoftBreak = True
+ -- Note that \W characters count as spacy in AsciiDoc
+ -- for purposes of determining interword:
+ isSpacy End (Str xs) = case reverse xs of
+ c:_ -> isPunctuation c || isSpace c
+ _ -> False
+ isSpacy Start (Str (c:_)) = isPunctuation c || isSpace c
+ isSpacy _ _ = False
setIntraword :: Bool -> State WriterState ()
setIntraword b = modify $ \st -> st{ intraword = b }
diff --git a/tests/writer.asciidoc b/tests/writer.asciidoc
index 8d0973675..2bf62e36f 100644
--- a/tests/writer.asciidoc
+++ b/tests/writer.asciidoc
@@ -435,11 +435,11 @@ Hr’s:
Inline Markup
-------------
-This is __emphasized__, and so __is this__.
+This is _emphasized_, and so _is this_.
-This is **strong**, and so **is this**.
+This is *strong*, and so *is this*.
-An __link:/url[emphasized link]__.
+An _link:/url[emphasized link]_.
*_This is strong and em._*
@@ -451,7 +451,7 @@ So is *_this_* word.
This is code: `>`, `$`, `\`, `\$`, `<html>`.
-[line-through]*This is __strikeout__.*
+[line-through]*This is _strikeout_.*
Superscripts: a^bc^d a^_hello_^ a^hello there^.