From 8ccc1f1547ac9ab5d492461b25dcb8470d548ff8 Mon Sep 17 00:00:00 2001
From: Tony Brix <tony@brix.ninja>
Date: Mon, 1 Jul 2019 08:47:06 -0500
Subject: [PATCH] improve paragraph

---
 lib/marked.js                                 | 33 +++++++++++--------
 test/specs/commonmark/commonmark.0.29.json    | 18 ++++------
 test/specs/gfm/commonmark.0.29.json           | 18 ++++------
 ...pedantic_heading_interrupts_paragraph.html |  5 +++
 .../pedantic_heading_interrupts_paragraph.md  | 10 ++++++
 test/specs/new/toplevel_paragraphs.html       |  6 ----
 test/specs/new/toplevel_paragraphs.md         |  7 ----
 7 files changed, 47 insertions(+), 50 deletions(-)
 create mode 100644 test/specs/new/pedantic_heading_interrupts_paragraph.html
 create mode 100644 test/specs/new/pedantic_heading_interrupts_paragraph.md

diff --git a/lib/marked.js b/lib/marked.js
index 8543c6b288..6145a8cf65 100644
--- a/lib/marked.js
+++ b/lib/marked.js
@@ -33,7 +33,9 @@ var block = {
   nptable: noop,
   table: noop,
   lheading: /^([^\n]+)\n {0,3}(=+|-+) *(?:\n+|$)/,
-  paragraph: /^([^\n]+(?:\n(?!hr|heading|lheading| {0,3}>|<\/?(?:tag)(?: +|\n|\/?>)|<(?:script|pre|style|!--))[^\n]+)*)/,
+  // regex template, placeholders will be replaced according to different paragraph
+  // interruption rules of commonmark and the original markdown spec:
+  _paragraph: /^([^\n]+(?:\n(?!hr|heading|lheading|blockquote|fences|list|html)[^\n]+)*)/,
   text: /^[^\n]+/
 };
 
@@ -69,10 +71,14 @@ block.html = edit(block.html, 'i')
   .replace('attribute', / +[a-zA-Z:_][\w.:-]*(?: *= *"[^"\n]*"| *= *'[^'\n]*'| *= *[^\s"'=<>`]+)?/)
   .getRegex();
 
-block.paragraph = edit(block.paragraph)
+block.paragraph = edit(block._paragraph)
   .replace('hr', block.hr)
-  .replace('heading', block.heading)
-  .replace('lheading', block.lheading)
+  .replace('heading', ' {0,3}#{1,6} +')
+  .replace('|lheading', '') // setex headings don't interrupt commonmark paragraphs
+  .replace('blockquote', ' {0,3}>')
+  .replace('fences', ' {0,3}(?:`{3,}|~{3,})[^`\\n]*\\n')
+  .replace('list', ' {0,3}(?:[*+-]|1[.)]) ') // only lists starting from 1 can interrupt
+  .replace('html', '</?(?:tag)(?: +|\\n|/?>)|<(?:script|pre|style|!--)')
   .replace('tag', block._tag) // pars can be interrupted by type (6) html blocks
   .getRegex();
 
@@ -91,16 +97,8 @@ block.normal = merge({}, block);
  */
 
 block.gfm = merge({}, block.normal, {
-  paragraph: /^/,
-  heading: /^ *(#{1,6}) +([^\n]+?) *#* *(?:\n+|$)/
 });
 
-block.gfm.paragraph = edit(block.paragraph)
-  .replace('(?!', '(?!'
-    + block.fences.source.replace('\\1', '\\2') + '|'
-    + block.list.source.replace('\\1', '\\3') + '|')
-  .getRegex();
-
 /**
  * GFM + Tables Block Grammar
  */
@@ -127,7 +125,16 @@ block.pedantic = merge({}, block.normal, {
     .getRegex(),
   def: /^ *\[([^\]]+)\]: *<?([^\s>]+)>?(?: +(["(][^\n]+[")]))? *(?:\n+|$)/,
   heading: /^ *(#{1,6}) *([^\n]+?) *(?:#+ *)?(?:\n+|$)/,
-  fences: noop // fences not supported
+  fences: noop, // fences not supported
+  paragraph: edit(block.normal._paragraph)
+    .replace('hr', block.hr)
+    .replace('heading', ' *#{1,6} *[^\n]')
+    .replace('lheading', block.lheading)
+    .replace('blockquote', ' {0,3}>')
+    .replace('|fences', '')
+    .replace('|list', '')
+    .replace('|html', '')
+    .getRegex()
 });
 
 /**
diff --git a/test/specs/commonmark/commonmark.0.29.json b/test/specs/commonmark/commonmark.0.29.json
index e203324415..e4f58e8aec 100644
--- a/test/specs/commonmark/commonmark.0.29.json
+++ b/test/specs/commonmark/commonmark.0.29.json
@@ -505,8 +505,7 @@
     "example": 63,
     "start_line": 1186,
     "end_line": 1196,
-    "section": "Setext headings",
-    "shouldFail": true
+    "section": "Setext headings"
   },
   {
     "markdown": "- Foo\n---\n",
@@ -883,8 +882,7 @@
     "example": 110,
     "start_line": 1908,
     "end_line": 1919,
-    "section": "Fenced code blocks",
-    "shouldFail": true
+    "section": "Fenced code blocks"
   },
   {
     "markdown": "foo\n---\n~~~\nbar\n~~~\n# baz\n",
@@ -1648,8 +1646,7 @@
     "example": 205,
     "start_line": 3464,
     "end_line": 3476,
-    "section": "Block quotes",
-    "shouldFail": true
+    "section": "Block quotes"
   },
   {
     "markdown": ">     foo\n    bar\n",
@@ -2210,8 +2207,7 @@
     "example": 273,
     "start_line": 4928,
     "end_line": 4938,
-    "section": "Lists",
-    "shouldFail": true
+    "section": "Lists"
   },
   {
     "markdown": "The number of windows in my house is\n14.  The number of doors is 6.\n",
@@ -2227,8 +2223,7 @@
     "example": 275,
     "start_line": 5015,
     "end_line": 5023,
-    "section": "Lists",
-    "shouldFail": true
+    "section": "Lists"
   },
   {
     "markdown": "- foo\n\n- bar\n\n\n- baz\n",
@@ -2365,8 +2360,7 @@
     "example": 291,
     "start_line": 5379,
     "end_line": 5397,
-    "section": "Lists",
-    "shouldFail": true
+    "section": "Lists"
   },
   {
     "markdown": "- a\n",
diff --git a/test/specs/gfm/commonmark.0.29.json b/test/specs/gfm/commonmark.0.29.json
index f577c7c5cd..885da2d835 100644
--- a/test/specs/gfm/commonmark.0.29.json
+++ b/test/specs/gfm/commonmark.0.29.json
@@ -358,8 +358,7 @@
     "example": 45,
     "start_line": 915,
     "end_line": 919,
-    "section": "ATX headings",
-    "shouldFail": true
+    "section": "ATX headings"
   },
   {
     "markdown": "### foo \\###\n## foo #\\##\n# foo \\#\n",
@@ -367,8 +366,7 @@
     "example": 46,
     "start_line": 925,
     "end_line": 933,
-    "section": "ATX headings",
-    "shouldFail": true
+    "section": "ATX headings"
   },
   {
     "markdown": "****\n## foo\n****\n",
@@ -507,8 +505,7 @@
     "example": 63,
     "start_line": 1186,
     "end_line": 1196,
-    "section": "Setext headings",
-    "shouldFail": true
+    "section": "Setext headings"
   },
   {
     "markdown": "- Foo\n---\n",
@@ -1649,8 +1646,7 @@
     "example": 205,
     "start_line": 3464,
     "end_line": 3476,
-    "section": "Block quotes",
-    "shouldFail": true
+    "section": "Block quotes"
   },
   {
     "markdown": ">     foo\n    bar\n",
@@ -2219,8 +2215,7 @@
     "example": 274,
     "start_line": 5005,
     "end_line": 5011,
-    "section": "Lists",
-    "shouldFail": true
+    "section": "Lists"
   },
   {
     "markdown": "The number of windows in my house is\n1.  The number of doors is 6.\n",
@@ -2365,8 +2360,7 @@
     "example": 291,
     "start_line": 5379,
     "end_line": 5397,
-    "section": "Lists",
-    "shouldFail": true
+    "section": "Lists"
   },
   {
     "markdown": "- a\n",
diff --git a/test/specs/new/pedantic_heading_interrupts_paragraph.html b/test/specs/new/pedantic_heading_interrupts_paragraph.html
new file mode 100644
index 0000000000..9e232fd65d
--- /dev/null
+++ b/test/specs/new/pedantic_heading_interrupts_paragraph.html
@@ -0,0 +1,5 @@
+<p>paragraph before head with hash</p>
+<h1 id="how-are-you">how are you</h1>
+
+<p>paragraph before head with equals</p>
+<h1 id="how-are-you-again">how are you again</h1>
diff --git a/test/specs/new/pedantic_heading_interrupts_paragraph.md b/test/specs/new/pedantic_heading_interrupts_paragraph.md
new file mode 100644
index 0000000000..13300a7a1b
--- /dev/null
+++ b/test/specs/new/pedantic_heading_interrupts_paragraph.md
@@ -0,0 +1,10 @@
+---
+pedantic: true
+---
+
+paragraph before head with hash
+#how are you
+
+paragraph before head with equals
+how are you again
+===========
diff --git a/test/specs/new/toplevel_paragraphs.html b/test/specs/new/toplevel_paragraphs.html
index 7fd509470e..dec8d1d772 100644
--- a/test/specs/new/toplevel_paragraphs.html
+++ b/test/specs/new/toplevel_paragraphs.html
@@ -8,12 +8,6 @@
 <p>paragraph before hr</p>
 <hr>
 
-<p>paragraph before head with hash</p>
-<h1 id="how-are-you">how are you</h1>
-
-<p>paragraph before head with equals</p>
-<h1 id="how-are-you-again">how are you again</h1>
-
 <p>paragraph before blockquote</p>
 <blockquote><p>text for blockquote</p></blockquote>
 
diff --git a/test/specs/new/toplevel_paragraphs.md b/test/specs/new/toplevel_paragraphs.md
index f344fbe349..9e92e44f9c 100644
--- a/test/specs/new/toplevel_paragraphs.md
+++ b/test/specs/new/toplevel_paragraphs.md
@@ -13,13 +13,6 @@ text inside block code
 paragraph before hr
 * * *
 
-paragraph before head with hash
-# how are you
-
-paragraph before head with equals
-how are you again
-===========
-
 paragraph before blockquote
 > text for blockquote