From 732ad89f3d7b13664d7ccc09490b63bfa6c8392e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Adri=C3=A1n=20Oliva?= <aditooliva@gmail.com>
Date: Fri, 1 Sep 2023 15:55:49 -0600
Subject: [PATCH] markdown: Fix URL link topic skipping query.

When searching for links inside a topic name, the question mark (?)
was used to split the topic. If a URL had a query after the URL
(e.g., "?foo=bar"), then the query was trimmed from the URL.

Removing the question mark from `basic_link_splitter` is sufficient
to fix this issue. The `get_web_link_regex` function then removes
the trailing punctuation if any, including literal question marks.

Fixes #26368.
---
 zerver/lib/markdown/__init__.py |  2 +-
 zerver/tests/test_markdown.py   | 30 ++++++++++++++++++++++++++++++
 2 files changed, 31 insertions(+), 1 deletion(-)

diff --git a/zerver/lib/markdown/__init__.py b/zerver/lib/markdown/__init__.py
index 21b6f9ac7c..4808511222 100644
--- a/zerver/lib/markdown/__init__.py
+++ b/zerver/lib/markdown/__init__.py
@@ -2284,7 +2284,7 @@ def make_md_engine(linkifiers_key: int, email_gateway: bool) -> None:
 
 # Split the topic name into multiple sections so that we can easily use
 # our common single link matching regex on it.
-basic_link_splitter = re.compile(r"[ !;\?\),\'\"]")
+basic_link_splitter = re.compile(r"[ !;\),\'\"]")
 
 
 def percent_escape_format_string(format_string: str) -> str:
diff --git a/zerver/tests/test_markdown.py b/zerver/tests/test_markdown.py
index 979a29f56b..c75f129c50 100644
--- a/zerver/tests/test_markdown.py
+++ b/zerver/tests/test_markdown.py
@@ -1160,6 +1160,36 @@ class MarkdownTest(ZulipTestCase):
             ],
         )
 
+        # Query strings in a URL should be included in the link.
+        msg.set_topic_name("https://google.com/test?foo=bar")
+        converted_topic = topic_links(realm.id, msg.topic_name())
+        self.assertEqual(
+            converted_topic,
+            [
+                {
+                    "url": "https://google.com/test?foo=bar",
+                    "text": "https://google.com/test?foo=bar",
+                },
+            ],
+        )
+        # But question marks at the end of sentence are not part of the URL.
+        msg.set_topic_name("Have you seen github.com/zulip?")
+        converted_topic = topic_links(realm.id, msg.topic_name())
+        self.assertEqual(
+            converted_topic,
+            [
+                {"url": "https://github.com/zulip", "text": "github.com/zulip"},
+            ],
+        )
+        msg.set_topic_name("Do you like https://example.com? I love it.")
+        converted_topic = topic_links(realm.id, msg.topic_name())
+        self.assertEqual(
+            converted_topic,
+            [
+                {"url": "https://example.com", "text": "https://example.com"},
+            ],
+        )
+
     def check_add_linkifiers(
         self, linkifiers: List[RealmFilter], expected_linkifier_reprs: List[str]
     ) -> None: