From 2f3eeb7a5a258f21942c82ba47c47edb6b3284f6 Mon Sep 17 00:00:00 2001
From: Titus Wormer <tituswormer@gmail.com>
Date: Tue, 13 Dec 2022 18:01:51 +0400
Subject: [PATCH] Remove unneeded escapes of `-`, `*`

Related-to: GH-57.
---
 lib/unsafe.js |  4 ++--
 test/index.js | 54 ++++++++++++++++++++++++++++++++++++++-------------
 2 files changed, 42 insertions(+), 16 deletions(-)

diff --git a/lib/unsafe.js b/lib/unsafe.js
index 4c96ae9..832d18f 100644
--- a/lib/unsafe.js
+++ b/lib/unsafe.js
@@ -85,13 +85,13 @@ export const unsafe = [
   {atBreak: true, before: '\\d+', character: ')'},
   {character: ')', inConstruct: 'destinationRaw'},
   // An asterisk can start thematic breaks, list items, emphasis, strong.
-  {atBreak: true, character: '*'},
+  {atBreak: true, character: '*', after: '(?:[ \t\r\n*])'},
   {character: '*', inConstruct: 'phrasing', notInConstruct: fullPhrasingSpans},
   // A plus sign could start a list item.
   {atBreak: true, character: '+', after: '(?:[ \t\r\n])'},
   // A dash can start thematic breaks, list items, and setext heading
   // underlines.
-  {atBreak: true, character: '-'},
+  {atBreak: true, character: '-', after: '(?:[ \t\r\n-])'},
   // A dot could start a list item.
   {atBreak: true, before: '\\d+', character: '.', after: '(?:[ \t\r\n]|$)'},
   // Slash, colon, and semicolon are not used in markdown for constructs.
diff --git a/test/index.js b/test/index.js
index fd56080..11273e4 100644
--- a/test/index.js
+++ b/test/index.js
@@ -1646,8 +1646,8 @@ test('code (text)', (t) => {
   )
 
   t.equal(
-    to({type: 'inlineCode', value: 'a\n-'}),
-    '`a -`\n',
+    to({type: 'inlineCode', value: 'a\n- b'}),
+    '`a - b`\n',
     'should prevent breaking out of code (-)'
   )
 
@@ -1664,14 +1664,14 @@ test('code (text)', (t) => {
   )
 
   t.equal(
-    to({type: 'inlineCode', value: 'a\r-'}),
-    '`a -`\n',
+    to({type: 'inlineCode', value: 'a\r- b'}),
+    '`a - b`\n',
     'should prevent breaking out of code (cr)'
   )
 
   t.equal(
-    to({type: 'inlineCode', value: 'a\r\n-'}),
-    '`a -`\n',
+    to({type: 'inlineCode', value: 'a\r\n- b'}),
+    '`a - b`\n',
     'should prevent breaking out of code (crlf)'
   )
 
@@ -3031,12 +3031,6 @@ test('escape', (t) => {
     'should escape what would otherwise be a character escape'
   )
 
-  t.equal(
-    to({type: 'paragraph', children: [{type: 'text', value: '+a'}]}),
-    '+a\n',
-    'should not escape a `+` if it is not followed by a whitespace'
-  )
-
   t.equal(
     to({
       type: 'paragraph',
@@ -3171,6 +3165,12 @@ test('escape', (t) => {
     'should escape what would otherwise be a list item (plus)'
   )
 
+  t.equal(
+    to({type: 'paragraph', children: [{type: 'text', value: '+a'}]}),
+    '+a\n',
+    'should not escape `+` when not followed by whitespace'
+  )
+
   t.equal(
     to({type: 'paragraph', children: [{type: 'text', value: '- a\n- b'}]}),
     '\\- a\n\\- b\n',
@@ -3178,11 +3178,37 @@ test('escape', (t) => {
   )
 
   t.equal(
-    to({type: 'paragraph', children: [{type: 'text', value: '* a\n* b'}]}),
-    '\\* a\n\\* b\n',
+    to({type: 'paragraph', children: [{type: 'text', value: '-a'}]}),
+    '-a\n',
+    'should not escape `-` when not followed by whitespace'
+  )
+
+  t.equal(
+    to({type: 'paragraph', children: [{type: 'text', value: '--a'}]}),
+    '\\--a\n',
+    'should escape `-` when followed by another `-` (as it looks like a thematic break, setext underline)'
+  )
+
+  // Note: these are in titles, because the `*` case here is about flow nodes,
+  // not phrasing (emphasis).
+  t.equal(
+    to({type: 'definition', identifier: 'x', url: 'y', title: 'a\n* b\n* c'}),
+    '[x]: y "a\n\\* b\n\\* c"\n',
     'should escape what would otherwise be a list item (asterisk)'
   )
 
+  t.equal(
+    to({type: 'definition', identifier: 'x', url: 'y', title: 'a\n*b'}),
+    '[x]: y "a\n*b"\n',
+    'should not escape `*` when not followed by whitespace'
+  )
+
+  t.equal(
+    to({type: 'definition', identifier: 'x', url: 'y', title: 'a\n**b'}),
+    '[x]: y "a\n\\**b"\n',
+    'should escape `*` when followed by another `*` (as it looks like a thematic break)'
+  )
+
   t.equal(
     to({type: 'paragraph', children: [{type: 'text', value: '1. a\n2. b'}]}),
     '1\\. a\n2\\. b\n',