Fix escaping of - in ms writer.

In 5132f1e we added `-` to
the list of characters needing backslash escaping, to accommodate
a change in groff man's behavior, described here:
https://lwn.net/Articles/947941/

This change also led `-` to be escaped in ms output, but that
is wrong; `\-` in ms is a unicode minus sign.

To fix this, we add a Boolean parameter to `escapeString` in
Text.Pandoc.Writers.Roff that determines whether `-` is to
be escaped.  (NB: This is not an exported function in the API.)

The list `standardEscapes` in Text.Pandoc.RoffChar no longer
contains `-`.

Closes #10536.

Author: jgm

src/Text/Pandoc/RoffChar.hs

@@ -34,7 +34,6 @@ standardEscapes =
   , ('`', "\\[ga]")
   , ('^', "\\[ha]")
   , ('~', "\\[ti]")
-  , ('-', "\\-")
   , ('\\', "\\[rs]")
   , ('@', "\\[at]") -- because we use @ as a table and math delimiter
   , ('\x2026', "\\&...")  -- because u2026 doesn't render on tty

src/Text/Pandoc/Writers/Man.hs

@@ -86,9 +86,9 @@ pandocToMan opts (Pandoc meta blocks) = do
        Just tpl -> renderTemplate tpl context
 
 escString :: WriterOptions -> Text -> Text
-escString opts = escapeString (if writerPreferAscii opts
-                                  then AsciiOnly
-                                  else AllowUTF8)
+escString opts = escapeString True (if writerPreferAscii opts
+                                       then AsciiOnly
+                                       else AllowUTF8)
 
 -- | Return man representation of notes.
 notesToMan :: PandocMonad m => WriterOptions -> [[Block]] -> StateT WriterState m (Doc Text)

src/Text/Pandoc/Writers/Ms.hs

@@ -92,7 +92,7 @@ pandocToMs opts (Pandoc meta blocks) = do
 
 escapeStr :: WriterOptions -> Text -> Text
 escapeStr opts =
-  escapeString (if writerPreferAscii opts then AsciiOnly else AllowUTF8)
+  escapeString False (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.

src/Text/Pandoc/Writers/Roff.hs

@@ -74,13 +74,18 @@ combiningAccentsMap = Map.fromList combiningAccents
 essentialEscapes :: Map.Map Char Text
 essentialEscapes = Map.fromList standardEscapes
 
--- | Escape special characters for roff.
-escapeString :: EscapeMode -> Text -> Text
-escapeString e = Text.concat . escapeString' e . Text.unpack
+-- | Escape special characters for roff. If the first parameter is
+-- True, escape @-@ as @\-@, as required by current versions of groff man;
+-- otherwise leave it unescaped, as neededfor ms.
+escapeString :: Bool -> EscapeMode -> Text -> Text
+escapeString escapeHyphen e = Text.concat . escapeString' e . Text.unpack
   where
     escapeString' _ [] = []
     escapeString' escapeMode ('\n':'.':xs) =
       "\n\\&." : escapeString' escapeMode xs
+    -- see #10533; we need to escape hyphens as \- in man but not in ms:
+    escapeString' escapeMode ('-':xs) | escapeHyphen =
+      "\\-" : escapeString' escapeMode xs
     escapeString' escapeMode (x:xs) =
       case Map.lookup x essentialEscapes of
         Just s  -> s : escapeString' escapeMode xs

test/writer.ms

@@ -134,7 +134,7 @@ Here\[cq]s a regular paragraph.
 In Markdown 1.0.0 and earlier.
 Version 8.
 This line turns into a list item.
-Because a hard\-wrapped line in the middle of a paragraph looked like a list
+Because a hard-wrapped line in the middle of a paragraph looked like a list
 item.
 .PP
 Here\[cq]s one with a bullet.
@@ -149,7 +149,7 @@ Block Quotes
 .pdfhref O 1 "Block Quotes"
 .pdfhref M "block-quotes"
 .LP
-E\-mail style:
+E-mail style:
 .QS
 .LP
 This is a block quote.
@@ -197,7 +197,7 @@ Code:
 .IP
 .nf
 \f[C]
-\-\-\-\- (should be four hyphens)
+---- (should be four hyphens)
 
 sub status {
     print \[dq]working\[dq];
@@ -605,7 +605,7 @@ Code block:
 .IP
 .nf
 \f[C]
-<!\-\- Comment \-\->
+<!-- Comment -->
 \f[]
 .fi
 .LP
@@ -695,7 +695,7 @@ LaTeX
 .IP \[bu] 3
 @223@
 .IP \[bu] 3
-@p@\-Tree
+@p@-Tree
 .IP \[bu] 3
 Here\[cq]s some display math:
 .EQ
@@ -765,7 +765,7 @@ Left paren: (
 .PP
 Right paren: )
 .PP
-Greater\-than: >
+Greater-than: >
 .PP
 Hash: #
 .PP
@@ -775,7 +775,7 @@ Bang: !
 .PP
 Plus: +
 .PP
-Minus: \-
+Minus: -
 .HLINE
 .SH 1
 Links
@@ -925,7 +925,7 @@ In a list?
 .IP \[bu] 3
 It should.
 .LP
-An e\-mail address: \c
+An e-mail address: \c
 .pdfhref W -D "mailto:nobody%40nowhere.net" -A "\c" \
  -- "nobody\[at]nowhere.net"
 \&
@@ -937,7 +937,7 @@ Blockquoted: \c
 \&
 .QE
 .LP
-Auto\-links should not occur here: \f[CR]<http://example.com/>\f[R]
+Auto-links should not occur here: \f[CR]<http://example.com/>\f[R]
 .IP
 .nf
 \f[C]