Files
RustPython/Lib/test/test_unicode_identifiers.py
Shahar Naveh 620faab734 Update some tests to 3.14.5 (#7990)
* Update `test_embed.py` to 3.14.5

* Update `test_cmd_line*` to 3.14.5

* Update unicode related tests

* Update `test_script_helper.py`

* Update `test_gc.py`
2026-05-30 20:16:10 +09:00

34 lines
1.0 KiB
Python
Vendored
Raw Permalink Blame History

This file contains invisible Unicode characters
This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
import unittest
class PEP3131Test(unittest.TestCase):
def test_valid(self):
class T:
ä = 1
µ = 2 # this is a compatibility character
蟒 = 3
x󠄀 = 4
self.assertEqual(getattr(T, "\xe4"), 1)
self.assertEqual(getattr(T, "\u03bc"), 2)
self.assertEqual(getattr(T, '\u87d2'), 3)
self.assertEqual(getattr(T, 'x\U000E0100'), 4)
def test_non_bmp_normalized(self):
𝔘𝔫𝔦𝔠𝔬𝔡𝔢 = 1
self.assertIn("Unicode", dir())
@unittest.expectedFailure # TODO: RUSTPYTHON
def test_invalid(self):
try:
from test.tokenizedata import badsyntax_3131 # noqa: F401
except SyntaxError as err:
self.assertEqual(str(err),
"invalid character '€' (U+20AC) (badsyntax_3131.py, line 2)")
self.assertEqual(err.lineno, 2)
self.assertEqual(err.offset, 1)
else:
self.fail("expected exception didn't occur")
if __name__ == "__main__":
unittest.main()