1 files changed, 24 insertions, 7 deletions
diff --git a/src/Text/Pandoc/Writers/Ms.hs b/src/Text/Pandoc/Writers/Ms.hs
index 97c23f24d..53763a609 100644
--- a/src/Text/Pandoc/Writers/Ms.hs
+++ b/src/Text/Pandoc/Writers/Ms.hs
@@ -21,7 +21,7 @@ TODO:
 
 module Text.Pandoc.Writers.Ms ( writeMs ) where
 import Control.Monad.State.Strict
-import Data.Char (isLower, isUpper, ord)
+import Data.Char (isAscii, isLower, isUpper, ord)
 import Data.List (intercalate, intersperse)
 import Data.List.NonEmpty (nonEmpty)
 import qualified Data.Map as Map
@@ -46,6 +46,8 @@ import Text.Pandoc.Writers.Shared
 import Text.Pandoc.Writers.Roff
 import Text.Printf (printf)
 import Text.TeXMath (writeEqn)
+import qualified Data.Text.Encoding as TE
+import qualified Data.ByteString as B
 
 -- | Convert Pandoc to Ms.
 writeMs :: PandocMonad m => WriterOptions -> Pandoc -> m Text
@@ -88,6 +90,21 @@ escapeStr :: WriterOptions -> Text -> Text
 escapeStr opts =
   escapeString (if writerPreferAscii opts then AsciiOnly else AllowUTF8)
 
+-- In PDFs we need to escape parentheses and backslash.
+-- In PDF we need to encode as UTF-16 BE.
+escapePDFString :: Text -> Text
+escapePDFString t
+  | T.all isAscii t =
+    T.replace "(" "\\(" .  T.replace ")" "\\)" . T.replace "\\" "\\\\" $ t
+  | otherwise = ("\\376\\377" <>) .  -- add bom
+    mconcat . map encodeChar .  T.unpack $ t
+ where
+  encodeChar c =
+    if isAscii c && c /= '\\' && c /= '(' && c /= ')'
+       then "\\000" <> T.singleton c
+       else mconcat . map toOctal . B.unpack . TE.encodeUtf16BE $ T.singleton c
+  toOctal n = "\\" <> T.pack (printf "%03o" n)
+
 escapeUri :: Text -> Text
 escapeUri = T.pack . escapeURIString (\c -> c /= '@' && isAllowedInURI c) . T.unpack
 
@@ -143,7 +160,7 @@ blockToMs opts (Div (ident,cls,kvs) bs) = do
        setFirstPara
        return $ anchor $$ res
 blockToMs opts (Plain inlines) =
-  liftM vcat $ mapM (inlineListToMs' opts) $ splitSentences inlines
+  splitSentences <$> inlineListToMs' opts inlines
 blockToMs opts (Para [Image attr alt (src,_tit)])
   | let ext = takeExtension (T.unpack src) in (ext == ".ps" || ext == ".eps") = do
   let (mbW,mbH) = (inPoints opts <$> dimension Width attr,
@@ -156,7 +173,7 @@ blockToMs opts (Para [Image attr alt (src,_tit)])
                               space <>
                               doubleQuotes (literal (tshow (floor hp :: Int)))
                        _ -> empty
-  capt <- inlineListToMs' opts alt
+  capt <- splitSentences <$> inlineListToMs' opts alt
   return $ nowrap (literal ".PSPIC -C " <>
              doubleQuotes (literal (escapeStr opts src)) <>
              sizeAttrs) $$
@@ -166,9 +183,9 @@ blockToMs opts (Para [Image attr alt (src,_tit)])
 blockToMs opts (Para inlines) = do
   firstPara <- gets stFirstPara
   resetFirstPara
-  contents <- liftM vcat $ mapM (inlineListToMs' opts) $
-    splitSentences inlines
-  return $ literal (if firstPara then ".LP" else ".PP") $$ contents
+  contents <- inlineListToMs' opts inlines
+  return $ literal (if firstPara then ".LP" else ".PP") $$
+           splitSentences contents
 blockToMs _ b@(RawBlock f str)
   | f == Format "ms" = return $ literal str
   | otherwise        = do
@@ -196,7 +213,7 @@ blockToMs opts (Header level (ident,classes,_) inlines) = do
                                       (if T.null secnum
                                           then ""
                                           else "  ") <>
-                                      escapeStr opts (stringify inlines))
+                                      escapePDFString (stringify inlines))
   let backlink = nowrap (literal ".pdfhref L -D " <>
        doubleQuotes (literal (toAscii ident)) <> space <> literal "\\") <> cr <>
        literal " -- "