markdown: Fix merging of separate quotations.

Initally, when writing two or more quotes, having
a blank line in between them, merges those quotes.
This created confusion especially in "quote and reply".

This commit fixes such issues. Now two or more quotes
having a blank line in between them, will not get merged.

This change is correct both for usability and for improving our
compatibility with CommonMark.

Fixes #14379.

static/shared/js/fenced_code.js

@@ -55,22 +55,17 @@ export function wrap_code(code, lang) {
 }
 
 function wrap_quote(text) {
-    const paragraphs = text.split("\n\n");
+    const paragraphs = text.split("\n");
     const quoted_paragraphs = [];
 
     // Prefix each quoted paragraph with > at the
     // beginning of each line
     for (const paragraph of paragraphs) {
         const lines = paragraph.split("\n");
-        quoted_paragraphs.push(
-            lines
-                .filter((line) => line !== "")
-                .map((line) => "> " + line)
-                .join("\n"),
-        );
+        quoted_paragraphs.push(lines.map((line) => "> " + line).join("\n"));
     }
 
-    return quoted_paragraphs.join("\n\n");
+    return quoted_paragraphs.join("\n");
 }
 
 function wrap_tex(tex) {

static/third/marked/lib/marked.js

@@ -17,7 +17,7 @@ var block = {
   hr: /^( *[-*_]){3,} *(?:\n+|$)/,
   heading: /^ {0,3}(#{1,6}) +([^\n]*?)(?: +#+)? *(?:\n+|$)/,
   nptable: noop,
-  blockquote: /^(?!( *>\s*($|\n))*($|\n))( *>[^\n]*(\n(?!def)[^\n]+)*\n*)+/,
+  blockquote: /^(?!( *>\s*($|\n))*($|\n))( *>[^\n]*(\n(?!def)[^\n]+)*)+/,
   list: /^( *)(bull) [\s\S]+?(?:hr|def|\n{2,}(?! )(?!\1bull )\n*|\s*$)/,
   html: /^ *(?:comment *(?:\n|\s*$)|closed *(?:\n{2,}|\s*$)|closing *(?:\n{2,}|\s*$))/,
   def: /^ *\[([^\]]+)\]: *<?([^\s>]+)>?(?: +["(]([^\n]+)[")])? *(?:\n+|$)/,

zerver/lib/markdown/__init__.py

@@ -1546,6 +1546,34 @@ class BlockQuoteProcessor(markdown.blockprocessors.BlockQuoteProcessor):
                     r'[ ]{0,3}>[ ]?(.*)')
     mention_re = re.compile(mention.find_mentions)
 
+    # run() is very slightly forked from the base class; see notes below.
+    def run(self, parent: Element, blocks: List[Any]) -> None:
+        block = blocks.pop(0)
+        m = self.RE.search(block)
+        if m:
+            before = block[:m.start()]  # Lines before blockquote
+            # Pass lines before blockquote in recursively for parsing forst.
+            self.parser.parseBlocks(parent, [before])
+            # Remove ``> `` from beginning of each line.
+            block = '\n'.join(
+                [self.clean(line) for line in block[m.start():].split('\n')]
+            )
+
+        # Zulip modification: The next line is patched to match
+        # CommonMark rather than original Markdown.  In original
+        # Markdown, blockquotes with a blank line between them were
+        # merged, which makes it impossible to break a blockquote with
+        # a blank line intentionally.
+        #
+        # This is a new blockquote. Create a new parent element.
+        quote = etree.SubElement(parent, 'blockquote')
+
+        # Recursively parse block with blockquote as parent.
+        # change parser state so blockquotes embedded in lists use p tags
+        self.parser.state.set('blockquote')
+        self.parser.parseChunk(quote, block)
+        self.parser.state.reset()
+
     def clean(self, line: str) -> str:
         # Silence all the mentions inside blockquotes
         line = re.sub(self.mention_re, lambda m: "@_{}".format(m.group('match')), line)

zerver/lib/markdown/fenced_code.py

@@ -427,12 +427,12 @@ class FencedBlockPreprocessor(markdown.preprocessors.Preprocessor):
         return code
 
     def format_quote(self, text: str) -> str:
-        paragraphs = text.split("\n\n")
+        paragraphs = text.split("\n")
         quoted_paragraphs = []
         for paragraph in paragraphs:
             lines = paragraph.split("\n")
-            quoted_paragraphs.append("\n".join("> " + line for line in lines if line != ''))
-        return "\n\n".join(quoted_paragraphs)
+            quoted_paragraphs.append("\n".join("> " + line for line in lines))
+        return "\n".join(quoted_paragraphs)
 
     def format_spoiler(self, header: str, text: str) -> str:
         output = []

zerver/tests/fixtures/markdown_test_cases.json

@@ -111,6 +111,22 @@
       "input": ">\n>\ntext",
       "expected_output": "<blockquote>\n<p>text</p>\n</blockquote>"
     },
+    {
+      "name": "multiple_blockquote",
+      "input": "> Hello World!\n\n> Hello Universe!",
+      "expected_output": "<blockquote>\n<p>Hello World!</p>\n</blockquote>\n<blockquote>\n<p>Hello Universe!</p>\n</blockquote>"
+    },
+    {
+      "name": "multiple_blockquote",
+      "input": "```quote\nText in\nmultiple lines.\n```\n\n> Text in a single line.",
+      "expected_output": "<blockquote>\n<p>Text in<br>\nmultiple lines.</p>\n</blockquote>\n<blockquote>\n<p>Text in a single line.</p>\n</blockquote>"
+    },
+    {
+      "name": "multiple_blockquote_with_quote_and_reply",
+      "input": "Hamlet said:\n~~~quote\nCan we talk ?\n~~~\n> Yes !!\n\nWelcome to Open Source.",
+      "expected_output": "<p>Hamlet said:</p>\n<blockquote>\n<p>Can we talk ?</p>\n</blockquote>\n<blockquote>\n<p>Yes !!</p>\n</blockquote>\n<p>Welcome to Open Source.</p>",
+      "text_content": "Hamlet said:\n> Can we talk ?\n\n> Yes !!\n\nWelcome to Open Source."
+    },
     {
       "name": "fenced_quote_with_hashtag",
       "input": "```quote\n# line 1\n#line 2\n```",

zerver/tests/test_markdown.py

@@ -87,7 +87,8 @@ class FencedBlockPreprocessorTest(ZulipTestCase):
             '',
             '> hi',
             '> bye',
-            '',
+            '> ',
+            '> ',
             '',
             '',
         ]
@@ -113,7 +114,8 @@ class FencedBlockPreprocessorTest(ZulipTestCase):
             '',
             '',
             '> bye',
-            '',
+            '> ',
+            '> ',
             '',
             '',
         ]
@@ -177,9 +179,11 @@ class FencedBlockPreprocessorTest(ZulipTestCase):
         expected = [
             '',
             '> hi',
-            '',
+            '> ',
             '> **py:hello()**',
-            '',
+            '> ',
+            '> ',
+            '> ',
             '',
             '',
         ]