SIM905: Fix handling of U+001C..U+001F whitespace (#19849)

Fixes #19845

## Summary

The linked issue explains it well, Rust and Python do not agree on what
whitespace is for the purposes of `str.split`.

crates/ruff_linter/resources/test/fixtures/flake8_simplify/SIM905.py

@@ -161,3 +161,8 @@ r"""first
 'no need' to escape
 "swap" quote style
 "use' ugly triple quotes""".split("\n")
+
+# https://github.com/astral-sh/ruff/issues/19845
+print("S\x1cP\x1dL\x1eI\x1fT".split())
+print("\x1c\x1d\x1e\x1f>".split(maxsplit=0))
+print("<\x1c\x1d\x1e\x1f".rsplit(maxsplit=0))

crates/ruff_linter/src/rules/flake8_simplify/rules/split_static_string.rs

@@ -199,9 +199,9 @@ fn split_default(
             //   - "".split(maxsplit=0) -> []
             //   - " ".split(maxsplit=0) -> []
             let processed_str = if direction == Direction::Left {
-                string_val.trim_start()
+                string_val.trim_start_matches(py_unicode_is_whitespace)
             } else {
-                string_val.trim_end()
+                string_val.trim_end_matches(py_unicode_is_whitespace)
             };
             let list_items: &[_] = if processed_str.is_empty() {
                 &[]
@@ -214,7 +214,10 @@ fn split_default(
             ))
         }
         Ordering::Less => {
-            let list_items: Vec<&str> = string_val.split_whitespace().collect();
+            let list_items: Vec<&str> = string_val
+                .split(py_unicode_is_whitespace)
+                .filter(|s| !s.is_empty())
+                .collect();
             Some(construct_replacement(
                 &list_items,
                 str_value.first_literal_flags(),
@@ -292,3 +295,34 @@ enum Direction {
     Left,
     Right,
 }
+
+/// Like [`char::is_whitespace`] but with Python's notion of whitespace.
+///
+/// <https://github.com/astral-sh/ruff/issues/19845>
+/// <https://github.com/python/cpython/blob/v3.14.0rc1/Objects/unicodetype_db.h#L6673-L6711>
+#[rustfmt::skip]
+#[inline]
+const fn py_unicode_is_whitespace(ch: char) -> bool {
+    matches!(
+        ch,
+        | '\u{0009}'
+        | '\u{000A}'
+        | '\u{000B}'
+        | '\u{000C}'
+        | '\u{000D}'
+        | '\u{001C}'
+        | '\u{001D}'
+        | '\u{001E}'
+        | '\u{001F}'
+        | '\u{0020}'
+        | '\u{0085}'
+        | '\u{00A0}'
+        | '\u{1680}'
+        | '\u{2000}'..='\u{200A}'
+        | '\u{2028}'
+        | '\u{2029}'
+        | '\u{202F}'
+        | '\u{205F}'
+        | '\u{3000}'
+    )
+}

crates/ruff_linter/src/rules/flake8_simplify/snapshots/ruff_linter__rules__flake8_simplify__tests__SIM905_SIM905.py.snap

@@ -1402,6 +1402,8 @@ SIM905 [*] Consider using a list literal instead of `str.split`
 162 | | "swap" quote style
 163 | | "use' ugly triple quotes""".split("\n")
     | |_______________________________________^
+164 |
+165 |   # https://github.com/astral-sh/ruff/issues/19845
     |
 help: Replace with list literal
 
@@ -1414,3 +1416,62 @@ help: Replace with list literal
 162     |-"swap" quote style
 163     |-"use' ugly triple quotes""".split("\n")
     160 |+[r"first", r"'no need' to escape", r'"swap" quote style', r""""use' ugly triple quotes"""]
+164 161 | 
+165 162 | # https://github.com/astral-sh/ruff/issues/19845
+166 163 | print("S\x1cP\x1dL\x1eI\x1fT".split())
+
+SIM905 [*] Consider using a list literal instead of `str.split`
+   --> SIM905.py:166:7
+    |
+165 | # https://github.com/astral-sh/ruff/issues/19845
+166 | print("S\x1cP\x1dL\x1eI\x1fT".split())
+    |       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+167 | print("\x1c\x1d\x1e\x1f>".split(maxsplit=0))
+168 | print("<\x1c\x1d\x1e\x1f".rsplit(maxsplit=0))
+    |
+help: Replace with list literal
+
+ℹ Safe fix
+163 163 | "use' ugly triple quotes""".split("\n")
+164 164 | 
+165 165 | # https://github.com/astral-sh/ruff/issues/19845
+166     |-print("S\x1cP\x1dL\x1eI\x1fT".split())
+    166 |+print(["S", "P", "L", "I", "T"])
+167 167 | print("\x1c\x1d\x1e\x1f>".split(maxsplit=0))
+168 168 | print("<\x1c\x1d\x1e\x1f".rsplit(maxsplit=0))
+
+SIM905 [*] Consider using a list literal instead of `str.split`
+   --> SIM905.py:167:7
+    |
+165 | # https://github.com/astral-sh/ruff/issues/19845
+166 | print("S\x1cP\x1dL\x1eI\x1fT".split())
+167 | print("\x1c\x1d\x1e\x1f>".split(maxsplit=0))
+    |       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+168 | print("<\x1c\x1d\x1e\x1f".rsplit(maxsplit=0))
+    |
+help: Replace with list literal
+
+ℹ Safe fix
+164 164 | 
+165 165 | # https://github.com/astral-sh/ruff/issues/19845
+166 166 | print("S\x1cP\x1dL\x1eI\x1fT".split())
+167     |-print("\x1c\x1d\x1e\x1f>".split(maxsplit=0))
+    167 |+print([">"])
+168 168 | print("<\x1c\x1d\x1e\x1f".rsplit(maxsplit=0))
+
+SIM905 [*] Consider using a list literal instead of `str.split`
+   --> SIM905.py:168:7
+    |
+166 | print("S\x1cP\x1dL\x1eI\x1fT".split())
+167 | print("\x1c\x1d\x1e\x1f>".split(maxsplit=0))
+168 | print("<\x1c\x1d\x1e\x1f".rsplit(maxsplit=0))
+    |       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+    |
+help: Replace with list literal
+
+ℹ Safe fix
+165 165 | # https://github.com/astral-sh/ruff/issues/19845
+166 166 | print("S\x1cP\x1dL\x1eI\x1fT".split())
+167 167 | print("\x1c\x1d\x1e\x1f>".split(maxsplit=0))
+168     |-print("<\x1c\x1d\x1e\x1f".rsplit(maxsplit=0))
+    168 |+print(["<"])