Avoid overlong-line errors for lines that end with URLs (#3663)

crates/ruff/resources/test/fixtures/pycodestyle/E501.py

@@ -56,7 +56,29 @@ sit amet  consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labor
 # OK
 # https://loooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooong.url.com
 
-# Not OK
+# OK
 _ = """
 Source: https://github.com/PyCQA/pycodestyle/pull/258/files#diff-841c622497a8033d10152bfdfb15b20b92437ecdea21a260944ea86b77b51533
 """
+
+# OK
+_ = """
+[this-is-ok](https://github.com/PyCQA/pycodestyle/pull/258/files#diff-841c622497a8033d10152bfdfb15b20b92437ecdea21a260944ea86b77b51533)
+[this is ok](https://github.com/PyCQA/pycodestyle/pull/258/files#diff-841c622497a8033d10152bfdfb15b20b92437ecdea21a260944ea86b77b51533)
+"""
+
+
+# OK
+class Foo:
+    """
+    @see https://looooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooong.url.com
+
+    :param dynamodb_scan_kwargs: kwargs pass to <https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/dynamodb.html#DynamoDB.Table.scan>
+    """
+
+
+# Error
+class Bar:
+    """
+    This is a long sentence that ends with a shortened URL and, therefore, could easily be broken across multiple lines ([source](https://ruff.rs))
+    """

crates/ruff/src/rules/pycodestyle/helpers.rs

@@ -1,7 +1,6 @@
-use once_cell::sync::Lazy;
-use regex::Regex;
 use rustpython_parser::ast::{Cmpop, Expr, ExprKind};
 use rustpython_parser::Tok;
+use unicode_width::UnicodeWidthStr;
 
 use ruff_python_ast::helpers::{create_expr, unparse_expr};
 use ruff_python_ast::source_code::Stylist;
@@ -21,8 +20,6 @@ pub fn compare(left: &Expr, ops: &[Cmpop], comparators: &[Expr], stylist: &Styli
     )
 }
 
-static URL_REGEX: Lazy<Regex> = Lazy::new(|| Regex::new(r"^https?://\S+$").unwrap());
-
 pub fn is_overlong(
     line: &str,
     line_width: usize,
@@ -35,22 +32,25 @@ pub fn is_overlong(
     }
 
     let mut chunks = line.split_whitespace();
-    let (Some(first), Some(second)) = (chunks.next(), chunks.next()) else {
+    let (Some(first_chunk), Some(second_chunk)) = (chunks.next(), chunks.next()) else {
         // Single word / no printable chars - no way to make the line shorter
         return false;
     };
 
-    if first == "#" {
+    if first_chunk == "#" {
         if ignore_overlong_task_comments {
-            let second = second.trim_end_matches(':');
-            if task_tags.iter().any(|tag| tag == second) {
+            let second = second_chunk.trim_end_matches(':');
+            if task_tags.iter().any(|task_tag| task_tag == second) {
                 return false;
             }
         }
+    }
 
-        // Do not enforce the line length for commented lines that end with a URL
-        // or contain only a single word.
-        if chunks.last().map_or(true, |c| URL_REGEX.is_match(c)) {
+    // Do not enforce the line length for lines that end with a URL, as long as the URL
+    // begins before the limit.
+    let last_chunk = chunks.last().unwrap_or(second_chunk);
+    if last_chunk.contains("://") {
+        if line_width - last_chunk.width() <= limit {
             return false;
         }
     }

crates/ruff/src/rules/pycodestyle/snapshots/ruff__rules__pycodestyle__tests__E501_E501.py.snap

@@ -69,15 +69,15 @@ expression: diagnostics
   parent: ~
 - kind:
     name: LineTooLong
-    body: Line too long (129 > 88 characters)
+    body: Line too long (147 > 88 characters)
     suggestion: ~
     fixable: false
   location:
-    row: 61
+    row: 83
     column: 88
   end_location:
-    row: 61
-    column: 129
+    row: 83
+    column: 147
   fix: ~
   parent: ~