Add strxfrm and strcoll to locale mod (#4603)

Lib/test/test_locale.py

@@ -343,8 +343,7 @@ class TestFrFRNumberFormatting(FrFRCookedTest, BaseFormattingTest):
 class TestCollation(unittest.TestCase):
     # Test string collation functions
 
-    # TODO: RUSTPYTHON
-    @unittest.expectedFailure
+    @unittest.expectedFailureIfWindows("TODO: RUSTPYTHON")
     def test_strcoll(self):
         self.assertLess(locale.strcoll('a', 'b'), 0)
         self.assertEqual(locale.strcoll('a', 'a'), 0)
@@ -353,8 +352,7 @@ class TestCollation(unittest.TestCase):
         self.assertRaises(ValueError, locale.strcoll, 'a\0', 'a')
         self.assertRaises(ValueError, locale.strcoll, 'a', 'a\0')
 
-    # TODO: RUSTPYTHON
-    @unittest.expectedFailure
+    @unittest.expectedFailureIfWindows("TODO: RUSTPYTHON")
     def test_strxfrm(self):
         self.assertLess(locale.strxfrm('a'), locale.strxfrm('b'))
         # embedded null character
@@ -381,8 +379,6 @@ class TestEnUSCollation(BaseLocalizedTest, TestCollation):
         is_emscripten or is_wasi,
         "musl libc issue on Emscripten/WASI, bpo-46390"
     )
-    # TODO: RUSTPYTHON", strcoll has not been implemented
-    @unittest.expectedFailure
     def test_strcoll_with_diacritic(self):
         self.assertLess(locale.strcoll('à', 'b'), 0)
 
@@ -392,8 +388,6 @@ class TestEnUSCollation(BaseLocalizedTest, TestCollation):
         is_emscripten or is_wasi,
         "musl libc issue on Emscripten/WASI, bpo-46390"
     )
-    # TODO: RUSTPYTHON", strxfrm has not been implemented
-    @unittest.expectedFailure
     def test_strxfrm_with_diacritic(self):
         self.assertLess(locale.strxfrm('à'), locale.strxfrm('b'))
 

stdlib/src/locale.rs

@@ -67,6 +67,25 @@ mod _locale {
         )
     }
 
+    #[pyfunction]
+    fn strcoll(string1: PyStrRef, string2: PyStrRef, vm: &VirtualMachine) -> PyResult {
+        let cstr1 = CString::new(string1.as_str()).map_err(|e| e.to_pyexception(vm))?;
+        let cstr2 = CString::new(string2.as_str()).map_err(|e| e.to_pyexception(vm))?;
+        Ok(vm.new_pyobj(unsafe { libc::strcoll(cstr1.as_ptr(), cstr2.as_ptr()) }))
+    }
+
+    #[pyfunction]
+    fn strxfrm(string: PyStrRef, vm: &VirtualMachine) -> PyResult {
+        // https://github.com/python/cpython/blob/eaae563b6878aa050b4ad406b67728b6b066220e/Modules/_localemodule.c#L390-L442
+        let n1 = string.byte_len() + 1;
+        let mut buff: Vec<u8> = vec![0; n1];
+
+        let cstr = CString::new(string.as_str()).map_err(|e| e.to_pyexception(vm))?;
+        let n2 = unsafe { libc::strxfrm(buff.as_mut_ptr() as _, cstr.as_ptr(), n1) };
+        buff.truncate(n2);
+        Ok(vm.new_pyobj(String::from_utf8(buff).expect("strxfrm returned invalid utf-8 string")))
+    }
+
     #[pyfunction]
     fn localeconv(vm: &VirtualMachine) -> PyResult<PyDictRef> {
         let result = vm.ctx.new_dict();