From d20152e01135a27decfd08dbda097fd1fda75354 Mon Sep 17 00:00:00 2001
From: John MacFarlane <jgm@berkeley.edu>
Date: Sat, 18 Apr 2015 10:45:46 -0700
Subject: [PATCH] Markdown writer: improved escaping.

`<` should not be escaped as `\<`, for compatibility with
original Markdown.  We now escape `<` and `>` with entities.
Also, we now backslash-escape square brackets.

Closes #2086.
---
 src/Text/Pandoc/Writers/Markdown.hs |  5 +++--
 tests/Tests/Writers/Markdown.hs     |  4 ++--
 tests/writer.markdown               | 22 +++++++++++-----------
 tests/writer.opml                   |  8 ++++----
 4 files changed, 20 insertions(+), 19 deletions(-)

diff --git a/src/Text/Pandoc/Writers/Markdown.hs b/src/Text/Pandoc/Writers/Markdown.hs
index dee4d56a4..ef87a8603 100644
--- a/src/Text/Pandoc/Writers/Markdown.hs
+++ b/src/Text/Pandoc/Writers/Markdown.hs
@@ -246,7 +246,8 @@ noteToMarkdown opts num blocks = do
 -- | Escape special characters for Markdown.
 escapeString :: WriterOptions -> String -> String
 escapeString opts = escapeStringUsing markdownEscapes
-  where markdownEscapes = backslashEscapes specialChars
+  where markdownEscapes = ('<', "&lt;") : ('>', "&gt;") :
+                          backslashEscapes specialChars
         specialChars =
                 (if isEnabled Ext_superscript opts
                     then ('^':)
@@ -257,7 +258,7 @@ escapeString opts = escapeStringUsing markdownEscapes
                 (if isEnabled Ext_tex_math_dollars opts
                     then ('$':)
                     else id) $
-                "\\`*_<>#"
+                "\\`*_[]#"
 
 -- | Construct table of contents from list of header blocks.
 tableOfContents :: WriterOptions -> [Block] -> Doc
diff --git a/tests/Tests/Writers/Markdown.hs b/tests/Tests/Writers/Markdown.hs
index dce40ddcb..8ffd7a8c0 100644
--- a/tests/Tests/Writers/Markdown.hs
+++ b/tests/Tests/Writers/Markdown.hs
@@ -84,13 +84,13 @@ shortcutLinkRefsTests =
                        ]
      , "Reference link is followed by text in brackets"
           =:  (para ((link "/url" "" "link") <> "[text in brackets]"))
-          =?> unlines [ "[link][][text in brackets]"
+          =?> unlines [ "[link][]\\[text in brackets\\]"
                       , ""
                       , "  [link]: /url"
                       ]
      , "Reference link is followed by space and text in brackets"
           =:  (para ((link "/url" "" "link") <> " [text in brackets]"))
-          =?> unlines [ "[link][] [text in brackets]"
+          =?> unlines [ "[link][] \\[text in brackets\\]"
                       , ""
                       , "  [link]: /url"
                       ]
diff --git a/tests/writer.markdown b/tests/writer.markdown
index 7276b31c7..6a0b9801c 100644
--- a/tests/writer.markdown
+++ b/tests/writer.markdown
@@ -80,7 +80,7 @@ E-mail style:
 >
 > > nested
 
-This should not be a block quote: 2 \> 1.
+This should not be a block quote: 2 &gt; 1.
 
 And a following paragraph.
 
@@ -581,9 +581,9 @@ AT&T is another way to write it.
 
 This & that.
 
-4 \< 5.
+4 &lt; 5.
 
-6 \> 5.
+6 &gt; 5.
 
 Backslash: \\
 
@@ -597,15 +597,15 @@ Left brace: {
 
 Right brace: }
 
-Left bracket: [
+Left bracket: \[
 
-Right bracket: ]
+Right bracket: \]
 
 Left paren: (
 
 Right paren: )
 
-Greater-than: \>
+Greater-than: &gt;
 
 Hash: \#
 
@@ -652,7 +652,7 @@ Foo [bar](/url/).
 
 Foo [bar](/url/).
 
-With [embedded [brackets]](/url/).
+With [embedded \[brackets\]](/url/).
 
 [b](/url/) by itself should be a link.
 
@@ -662,7 +662,7 @@ Indented [twice](/url).
 
 Indented [thrice](/url).
 
-This should [not][] be a link.
+This should \[not\]\[\] be a link.
 
     [not]: /url
 
@@ -716,8 +716,8 @@ Footnotes
 =========
 
 Here is a footnote reference,[^1] and another.[^2] This should *not* be a
-footnote reference, because it contains a space.[\^my note] Here is an inline
-note.[^3]
+footnote reference, because it contains a space.\[\^my note\] Here is an
+inline note.[^3]
 
 > Notes can go in quotes.[^4]
 
@@ -740,7 +740,7 @@ This paragraph should not be part of the note, as it is not indented.
 
 [^3]: This is *easier* to type. Inline notes may contain
     [links](http://google.com) and `]` verbatim characters, as well as
-    [bracketed text].
+    \[bracketed text\].
 
 [^4]: In quote.
 
diff --git a/tests/writer.opml b/tests/writer.opml
index 8f79e842c..c10ca44c8 100644
--- a/tests/writer.opml
+++ b/tests/writer.opml
@@ -26,7 +26,7 @@
 </outline>
 <outline text="Paragraphs" _note="Here’s a regular paragraph.&#10;&#10;In Markdown 1.0.0 and earlier. Version 8. This line turns into a list&#10;item. Because a hard-wrapped line in the middle of a paragraph looked&#10;like a list item.&#10;&#10;Here’s one with a bullet. \* criminey.&#10;&#10;There should be a hard line break\&#10;here.&#10;&#10;------------------------------------------------------------------------">
 </outline>
-<outline text="Block Quotes" _note="E-mail style:&#10;&#10;&gt; This is a block quote. It is pretty short.&#10;&#10;&gt; Code in a block quote:&#10;&gt;&#10;&gt;     sub status {&#10;&gt;         print &quot;working&quot;;&#10;&gt;     }&#10;&gt;&#10;&gt; A list:&#10;&gt;&#10;&gt; 1.  item one&#10;&gt; 2.  item two&#10;&gt;&#10;&gt; Nested block quotes:&#10;&gt;&#10;&gt; &gt; nested&#10;&gt;&#10;&gt; &gt; nested&#10;&#10;This should not be a block quote: 2 \&gt; 1.&#10;&#10;And a following paragraph.&#10;&#10;------------------------------------------------------------------------">
+<outline text="Block Quotes" _note="E-mail style:&#10;&#10;&gt; This is a block quote. It is pretty short.&#10;&#10;&gt; Code in a block quote:&#10;&gt;&#10;&gt;     sub status {&#10;&gt;         print &quot;working&quot;;&#10;&gt;     }&#10;&gt;&#10;&gt; A list:&#10;&gt;&#10;&gt; 1.  item one&#10;&gt; 2.  item two&#10;&gt;&#10;&gt; Nested block quotes:&#10;&gt;&#10;&gt; &gt; nested&#10;&gt;&#10;&gt; &gt; nested&#10;&#10;This should not be a block quote: 2 &amp;gt; 1.&#10;&#10;And a following paragraph.&#10;&#10;------------------------------------------------------------------------">
 </outline>
 <outline text="Code Blocks" _note="Code:&#10;&#10;    ---- (should be four hyphens)&#10;&#10;    sub status {&#10;        print &quot;working&quot;;&#10;    }&#10;&#10;    this code block is indented by one tab&#10;&#10;And:&#10;&#10;        this code block is indented by two tabs&#10;&#10;    These should not be escaped:  \$ \\ \&gt; \[ \{&#10;&#10;------------------------------------------------------------------------">
 </outline>
@@ -52,12 +52,12 @@
 </outline>
 <outline text="LaTeX" _note="-   \cite[22-23]{smith.1899}&#10;-   $2+2=4$&#10;-   $x \in y$&#10;-   $\alpha \wedge \omega$&#10;-   $223$&#10;-   $p$-Tree&#10;-   Here’s some display math:&#10;    $$\frac{d}{dx}f(x)=\lim_{h\to 0}\frac{f(x+h)-f(x)}{h}$$&#10;-   Here’s one that has a line break in it:&#10;    $\alpha + \omega \times x^2$.&#10;&#10;These shouldn’t be math:&#10;&#10;-   To get the famous equation, write `$e = mc^2$`.&#10;-   \$22,000 is a *lot* of money. So is \$34,000. (It worked if “lot”&#10;    is emphasized.)&#10;-   Shoes (\$20) and socks (\$5).&#10;-   Escaped `$`: \$73 *this should be emphasized* 23\$.&#10;&#10;Here’s a LaTeX table:&#10;&#10;\begin{tabular}{|l|l|}\hline&#10;Animal &amp; Number \\ \hline&#10;Dog    &amp; 2      \\&#10;Cat    &amp; 1      \\ \hline&#10;\end{tabular}&#10;&#10;------------------------------------------------------------------------">
 </outline>
-<outline text="Special Characters" _note="Here is some unicode:&#10;&#10;-   I hat: Î&#10;-   o umlaut: ö&#10;-   section: §&#10;-   set membership: ∈&#10;-   copyright: ©&#10;&#10;AT&amp;T has an ampersand in their name.&#10;&#10;AT&amp;T is another way to write it.&#10;&#10;This &amp; that.&#10;&#10;4 \&lt; 5.&#10;&#10;6 \&gt; 5.&#10;&#10;Backslash: \\&#10;&#10;Backtick: \`&#10;&#10;Asterisk: \*&#10;&#10;Underscore: \_&#10;&#10;Left brace: {&#10;&#10;Right brace: }&#10;&#10;Left bracket: [&#10;&#10;Right bracket: ]&#10;&#10;Left paren: (&#10;&#10;Right paren: )&#10;&#10;Greater-than: \&gt;&#10;&#10;Hash: \#&#10;&#10;Period: .&#10;&#10;Bang: !&#10;&#10;Plus: +&#10;&#10;Minus: -&#10;&#10;------------------------------------------------------------------------">
+<outline text="Special Characters" _note="Here is some unicode:&#10;&#10;-   I hat: Î&#10;-   o umlaut: ö&#10;-   section: §&#10;-   set membership: ∈&#10;-   copyright: ©&#10;&#10;AT&amp;T has an ampersand in their name.&#10;&#10;AT&amp;T is another way to write it.&#10;&#10;This &amp; that.&#10;&#10;4 &amp;lt; 5.&#10;&#10;6 &amp;gt; 5.&#10;&#10;Backslash: \\&#10;&#10;Backtick: \`&#10;&#10;Asterisk: \*&#10;&#10;Underscore: \_&#10;&#10;Left brace: {&#10;&#10;Right brace: }&#10;&#10;Left bracket: \[&#10;&#10;Right bracket: \]&#10;&#10;Left paren: (&#10;&#10;Right paren: )&#10;&#10;Greater-than: &amp;gt;&#10;&#10;Hash: \#&#10;&#10;Period: .&#10;&#10;Bang: !&#10;&#10;Plus: +&#10;&#10;Minus: -&#10;&#10;------------------------------------------------------------------------">
 </outline>
 <outline text="Links">
   <outline text="Explicit" _note="Just a [URL](/url/).&#10;&#10;[URL and title](/url/ &quot;title&quot;).&#10;&#10;[URL and title](/url/ &quot;title preceded by two spaces&quot;).&#10;&#10;[URL and title](/url/ &quot;title preceded by a tab&quot;).&#10;&#10;[URL and title](/url/ &quot;title with &quot;quotes&quot; in it&quot;)&#10;&#10;[URL and title](/url/ &quot;title with single quotes&quot;)&#10;&#10;[with\_underscore](/url/with_underscore)&#10;&#10;[Email link](mailto:nobody@nowhere.net)&#10;&#10;[Empty]().">
   </outline>
-  <outline text="Reference" _note="Foo [bar](/url/).&#10;&#10;Foo [bar](/url/).&#10;&#10;Foo [bar](/url/).&#10;&#10;With [embedded [brackets]](/url/).&#10;&#10;[b](/url/) by itself should be a link.&#10;&#10;Indented [once](/url).&#10;&#10;Indented [twice](/url).&#10;&#10;Indented [thrice](/url).&#10;&#10;This should [not][] be a link.&#10;&#10;    [not]: /url&#10;&#10;Foo [bar](/url/ &quot;Title with &quot;quotes&quot; inside&quot;).&#10;&#10;Foo [biz](/url/ &quot;Title with &quot;quote&quot; inside&quot;).">
+  <outline text="Reference" _note="Foo [bar](/url/).&#10;&#10;Foo [bar](/url/).&#10;&#10;Foo [bar](/url/).&#10;&#10;With [embedded \[brackets\]](/url/).&#10;&#10;[b](/url/) by itself should be a link.&#10;&#10;Indented [once](/url).&#10;&#10;Indented [twice](/url).&#10;&#10;Indented [thrice](/url).&#10;&#10;This should \[not\]\[\] be a link.&#10;&#10;    [not]: /url&#10;&#10;Foo [bar](/url/ &quot;Title with &quot;quotes&quot; inside&quot;).&#10;&#10;Foo [biz](/url/ &quot;Title with &quot;quote&quot; inside&quot;).">
   </outline>
   <outline text="With ampersands" _note="Here’s a [link with an ampersand in the&#10;URL](http://example.com/?foo=1&amp;bar=2).&#10;&#10;Here’s a link with an amersand in the link text:&#10;[AT&amp;T](http://att.com/ &quot;AT&amp;T&quot;).&#10;&#10;Here’s an [inline link](/script?foo=1&amp;bar=2).&#10;&#10;Here’s an [inline link in pointy braces](/script?foo=1&amp;bar=2).">
   </outline>
@@ -66,7 +66,7 @@
 </outline>
 <outline text="Images" _note="From “Voyage dans la Lune” by Georges Melies (1902):&#10;&#10;![lalune](lalune.jpg &quot;Voyage dans la Lune&quot;)&#10;&#10;Here is a movie ![movie](movie.jpg) icon.&#10;&#10;------------------------------------------------------------------------">
 </outline>
-<outline text="Footnotes" _note="Here is a footnote reference,[^1] and another.[^2] This should *not* be&#10;a footnote reference, because it contains a space.[\^my note] Here is an&#10;inline note.[^3]&#10;&#10;&gt; Notes can go in quotes.[^4]&#10;&#10;1.  And in list items.[^5]&#10;&#10;This paragraph should not be part of the note, as it is not indented.&#10;&#10;[^1]: Here is the footnote. It can go anywhere after the footnote&#10;    reference. It need not be placed at the end of the document.&#10;&#10;[^2]: Here’s the long note. This one contains multiple blocks.&#10;&#10;    Subsequent blocks are indented to show that they belong to the&#10;    footnote (as with list items).&#10;&#10;          { &lt;code&gt; }&#10;&#10;    If you want, you can indent every line, but you can also be lazy and&#10;    just indent the first line of each block.&#10;&#10;[^3]: This is *easier* to type. Inline notes may contain&#10;    [links](http://google.com) and `]` verbatim characters, as well as&#10;    [bracketed text].&#10;&#10;[^4]: In quote.&#10;&#10;[^5]: In list.">
+<outline text="Footnotes" _note="Here is a footnote reference,[^1] and another.[^2] This should *not* be&#10;a footnote reference, because it contains a space.\[\^my note\] Here is&#10;an inline note.[^3]&#10;&#10;&gt; Notes can go in quotes.[^4]&#10;&#10;1.  And in list items.[^5]&#10;&#10;This paragraph should not be part of the note, as it is not indented.&#10;&#10;[^1]: Here is the footnote. It can go anywhere after the footnote&#10;    reference. It need not be placed at the end of the document.&#10;&#10;[^2]: Here’s the long note. This one contains multiple blocks.&#10;&#10;    Subsequent blocks are indented to show that they belong to the&#10;    footnote (as with list items).&#10;&#10;          { &lt;code&gt; }&#10;&#10;    If you want, you can indent every line, but you can also be lazy and&#10;    just indent the first line of each block.&#10;&#10;[^3]: This is *easier* to type. Inline notes may contain&#10;    [links](http://google.com) and `]` verbatim characters, as well as&#10;    \[bracketed text\].&#10;&#10;[^4]: In quote.&#10;&#10;[^5]: In list.">
 </outline>
   </body>
 </opml>