From f9fca5b46952f3238531ea33019ff64b62341ede Mon Sep 17 00:00:00 2001
From: Prakhar Pratyush <prakhar841301@gmail.com>
Date: Fri, 31 May 2024 00:02:52 +0530
Subject: [PATCH] remove_single_newlines: Fix regex to parse list syntax
 correctly.

Earlier, the 'remove_single_newlines' function wasn't working
correctly when '\n' was followed by asterix, hyphen, or number.
Specifically, they were not added as a syntax for ordered or
unordered list in markdown.

For example, see the workaround fix in 330439a83b7c9d01bfc608a7e82b427f53bae164.

This commit updates the function to replace '\n' with ' ', when
"\n" is not preceded by "\n" and not followed by:
* Another newline (\n)
* A hyphen followed by a space
* An asterisk followed by a space
* A number followed by a dot and a space

We won't have to do fixes like 330439a83b7c9d01bfc608a7e82b427f53bae164 in the future.
---
 zerver/lib/message.py                         |  2 +-
 .../tests/test_zulip_update_announcements.py  | 19 +++++++++++++++++++
 2 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/zerver/lib/message.py b/zerver/lib/message.py
index 3027d87bb7..6cbc9b8533 100644
--- a/zerver/lib/message.py
+++ b/zerver/lib/message.py
@@ -1435,4 +1435,4 @@ def set_visibility_policy_possible(user_profile: UserProfile, message: Message)
 
 def remove_single_newlines(content: str) -> str:
     content = content.strip("\n")
-    return re.sub(r"(?<!\n)\n(?![\n0-9*-])", " ", content)
+    return re.sub(r"(?<!\n)\n(?!\n|[-*] |[0-9]+\. )", " ", content)
diff --git a/zerver/tests/test_zulip_update_announcements.py b/zerver/tests/test_zulip_update_announcements.py
index e8c23d6d11..c9d5cb2743 100644
--- a/zerver/tests/test_zulip_update_announcements.py
+++ b/zerver/tests/test_zulip_update_announcements.py
@@ -313,6 +313,25 @@ class ZulipUpdateAnnouncementsTest(ZulipTestCase):
         expected_output = "- This is a bullet.\n- This is another bullet.\n\n1. This is a list\n1. This is more list."
         self.assertEqual(remove_single_newlines(input_text), expected_output)
 
+        # Asterisks after newline for bold.
+        input_text = "* This is a bullet.\n**word in bold** on the same line.\n* Another bullet."
+        expected_output = (
+            "* This is a bullet. **word in bold** on the same line.\n* Another bullet."
+        )
+        self.assertEqual(remove_single_newlines(input_text), expected_output)
+
+        # Digit after newline.
+        input_text = "1. This is a numbered list.\n2. Second list element.\n3.5 is a decimal.\n3. Third list element."
+        expected_output = "1. This is a numbered list.\n2. Second list element. 3.5 is a decimal.\n3. Third list element."
+        self.assertEqual(remove_single_newlines(input_text), expected_output)
+
+        # Hyphen after newline.
+        input_text = "- This is a list.\n-C-C- organic molecule structure.\n- Another list element."
+        expected_output = (
+            "- This is a list. -C-C- organic molecule structure.\n- Another list element."
+        )
+        self.assertEqual(remove_single_newlines(input_text), expected_output)
+
     def test_zulip_updates_for_realm_imported_from_other_product(self) -> None:
         with mock.patch(
             "zerver.lib.zulip_update_announcements.zulip_update_announcements",