blob: 9a969e8ec047419f80d888a189a5dd7d35c1eeb4 [file] [log] [blame]
"""Test suite for 2to3's parser and grammar files.
This is the place to add tests for changes to 2to3's grammar, such as those
merging the grammars for Python 2 and 3. In addition to specific tests for
parts of the grammar we've changed, we also make sure we can parse the
test_grammar.py files from both Python 2 and Python 3.
"""
# Testing imports
from . import support
from .support import driver
from test.support import verbose
# Python imports
import os
import shutil
import subprocess
import sys
import tempfile
import unittest
import warnings
# Local imports
from lib2to3.pgen2 import driver as pgen2_driver
from lib2to3.pgen2 import tokenize
from ..pgen2.parse import ParseError
from lib2to3.pygram import python_symbols as syms
class TestDriver(support.TestCase):
def test_formfeed(self):
s = """print 1\n\x0Cprint 2\n"""
t = driver.parse_string(s)
self.assertEqual(t.children[0].children[0].type, syms.print_stmt)
self.assertEqual(t.children[1].children[0].type, syms.print_stmt)
class TestPgen2Caching(support.TestCase):
def test_load_grammar_from_txt_file(self):
pgen2_driver.load_grammar(support.grammar_path, save=False, force=True)
def test_load_grammar_from_pickle(self):
# Make a copy of the grammar file in a temp directory we are
# guaranteed to be able to write to.
tmpdir = tempfile.mkdtemp()
try:
grammar_copy = os.path.join(
tmpdir, os.path.basename(support.grammar_path))
shutil.copy(support.grammar_path, grammar_copy)
pickle_name = pgen2_driver._generate_pickle_name(grammar_copy)
pgen2_driver.load_grammar(grammar_copy, save=True, force=True)
self.assertTrue(os.path.exists(pickle_name))
os.unlink(grammar_copy) # Only the pickle remains...
pgen2_driver.load_grammar(grammar_copy, save=False, force=False)
finally:
shutil.rmtree(tmpdir)
@unittest.skipIf(sys.executable is None, 'sys.executable required')
def test_load_grammar_from_subprocess(self):
tmpdir = tempfile.mkdtemp()
tmpsubdir = os.path.join(tmpdir, 'subdir')
try:
os.mkdir(tmpsubdir)
grammar_base = os.path.basename(support.grammar_path)
grammar_copy = os.path.join(tmpdir, grammar_base)
grammar_sub_copy = os.path.join(tmpsubdir, grammar_base)
shutil.copy(support.grammar_path, grammar_copy)
shutil.copy(support.grammar_path, grammar_sub_copy)
pickle_name = pgen2_driver._generate_pickle_name(grammar_copy)
pickle_sub_name = pgen2_driver._generate_pickle_name(
grammar_sub_copy)
self.assertNotEqual(pickle_name, pickle_sub_name)
# Generate a pickle file from this process.
pgen2_driver.load_grammar(grammar_copy, save=True, force=True)
self.assertTrue(os.path.exists(pickle_name))
# Generate a new pickle file in a subprocess with a most likely
# different hash randomization seed.
sub_env = dict(os.environ)
sub_env['PYTHONHASHSEED'] = 'random'
subprocess.check_call(
[sys.executable, '-c', """
from lib2to3.pgen2 import driver as pgen2_driver
pgen2_driver.load_grammar(%r, save=True, force=True)
""" % (grammar_sub_copy,)],
env=sub_env)
self.assertTrue(os.path.exists(pickle_sub_name))
with open(pickle_name, 'rb') as pickle_f_1, \
open(pickle_sub_name, 'rb') as pickle_f_2:
self.assertEqual(
pickle_f_1.read(), pickle_f_2.read(),
msg='Grammar caches generated using different hash seeds'
' were not identical.')
finally:
shutil.rmtree(tmpdir)
class GrammarTest(support.TestCase):
def validate(self, code):
support.parse_string(code)
def invalid_syntax(self, code):
try:
self.validate(code)
except ParseError:
pass
else:
raise AssertionError("Syntax shouldn't have been valid")
class TestMatrixMultiplication(GrammarTest):
def test_matrix_multiplication_operator(self):
self.validate("a @ b")
self.validate("a @= b")
class TestYieldFrom(GrammarTest):
def test_yield_from(self):
self.validate("yield from x")
self.validate("(yield from x) + y")
self.invalid_syntax("yield from")
class TestAsyncAwait(GrammarTest):
def test_await_expr(self):
self.validate("""async def foo():
await x
""")
self.validate("""async def foo():
[i async for i in b]
""")
self.validate("""async def foo():
{i for i in b
async for i in a if await i
for b in i}
""")
self.validate("""async def foo():
[await i for i in b if await c]
""")
self.validate("""async def foo():
[ i for i in b if c]
""")
self.validate("""async def foo():
def foo(): pass
def foo(): pass
await x
""")
self.validate("""async def foo(): return await a""")
self.validate("""def foo():
def foo(): pass
async def foo(): await x
""")
self.invalid_syntax("await x")
self.invalid_syntax("""def foo():
await x""")
self.invalid_syntax("""def foo():
def foo(): pass
async def foo(): pass
await x
""")
def test_async_var(self):
self.validate("""async = 1""")
self.validate("""await = 1""")
self.validate("""def async(): pass""")
def test_async_with(self):
self.validate("""async def foo():
async for a in b: pass""")
self.invalid_syntax("""def foo():
async for a in b: pass""")
def test_async_for(self):
self.validate("""async def foo():
async with a: pass""")
self.invalid_syntax("""def foo():
async with a: pass""")
class TestRaiseChanges(GrammarTest):
def test_2x_style_1(self):
self.validate("raise")
def test_2x_style_2(self):
self.validate("raise E, V")
def test_2x_style_3(self):
self.validate("raise E, V, T")
def test_2x_style_invalid_1(self):
self.invalid_syntax("raise E, V, T, Z")
def test_3x_style(self):
self.validate("raise E1 from E2")
def test_3x_style_invalid_1(self):
self.invalid_syntax("raise E, V from E1")
def test_3x_style_invalid_2(self):
self.invalid_syntax("raise E from E1, E2")
def test_3x_style_invalid_3(self):
self.invalid_syntax("raise from E1, E2")
def test_3x_style_invalid_4(self):
self.invalid_syntax("raise E from")
# Modelled after Lib/test/test_grammar.py:TokenTests.test_funcdef issue2292
# and Lib/test/text_parser.py test_list_displays, test_set_displays,
# test_dict_displays, test_argument_unpacking, ... changes.
class TestUnpackingGeneralizations(GrammarTest):
def test_mid_positional_star(self):
self.validate("""func(1, *(2, 3), 4)""")
def test_double_star_dict_literal(self):
self.validate("""func(**{'eggs':'scrambled', 'spam':'fried'})""")
def test_double_star_dict_literal_after_keywords(self):
self.validate("""func(spam='fried', **{'eggs':'scrambled'})""")
def test_list_display(self):
self.validate("""[*{2}, 3, *[4]]""")
def test_set_display(self):
self.validate("""{*{2}, 3, *[4]}""")
def test_dict_display_1(self):
self.validate("""{**{}}""")
def test_dict_display_2(self):
self.validate("""{**{}, 3:4, **{5:6, 7:8}}""")
def test_argument_unpacking_1(self):
self.validate("""f(a, *b, *c, d)""")
def test_argument_unpacking_2(self):
self.validate("""f(**a, **b)""")
def test_argument_unpacking_3(self):
self.validate("""f(2, *a, *b, **b, **c, **d)""")
# Adapted from Python 3's Lib/test/test_grammar.py:GrammarTests.testFuncdef
class TestFunctionAnnotations(GrammarTest):
def test_1(self):
self.validate("""def f(x) -> list: pass""")
def test_2(self):
self.validate("""def f(x:int): pass""")
def test_3(self):
self.validate("""def f(*x:str): pass""")
def test_4(self):
self.validate("""def f(**x:float): pass""")
def test_5(self):
self.validate("""def f(x, y:1+2): pass""")
def test_6(self):
self.validate("""def f(a, (b:1, c:2, d)): pass""")
def test_7(self):
self.validate("""def f(a, (b:1, c:2, d), e:3=4, f=5, *g:6): pass""")
def test_8(self):
s = """def f(a, (b:1, c:2, d), e:3=4, f=5,
*g:6, h:7, i=8, j:9=10, **k:11) -> 12: pass"""
self.validate(s)
# Adapted from Python 3's Lib/test/test_grammar.py:GrammarTests.test_var_annot
class TestVarAnnotations(GrammarTest):
def test_1(self):
self.validate("var1: int = 5")
def test_2(self):
self.validate("var2: [int, str]")
def test_3(self):
self.validate("def f():\n"
" st: str = 'Hello'\n"
" a.b: int = (1, 2)\n"
" return st\n")
def test_4(self):
self.validate("def fbad():\n"
" x: int\n"
" print(x)\n")
def test_5(self):
self.validate("class C:\n"
" x: int\n"
" s: str = 'attr'\n"
" z = 2\n"
" def __init__(self, x):\n"
" self.x: int = x\n")
def test_6(self):
self.validate("lst: List[int] = []")
class TestExcept(GrammarTest):
def test_new(self):
s = """
try:
x
except E as N:
y"""
self.validate(s)
def test_old(self):
s = """
try:
x
except E, N:
y"""
self.validate(s)
# Adapted from Python 3's Lib/test/test_grammar.py:GrammarTests.testAtoms
class TestSetLiteral(GrammarTest):
def test_1(self):
self.validate("""x = {'one'}""")
def test_2(self):
self.validate("""x = {'one', 1,}""")
def test_3(self):
self.validate("""x = {'one', 'two', 'three'}""")
def test_4(self):
self.validate("""x = {2, 3, 4,}""")
class TestNumericLiterals(GrammarTest):
def test_new_octal_notation(self):
self.validate("""0o7777777777777""")
self.invalid_syntax("""0o7324528887""")
def test_new_binary_notation(self):
self.validate("""0b101010""")
self.invalid_syntax("""0b0101021""")
class TestClassDef(GrammarTest):
def test_new_syntax(self):
self.validate("class B(t=7): pass")
self.validate("class B(t, *args): pass")
self.validate("class B(t, **kwargs): pass")
self.validate("class B(t, *args, **kwargs): pass")
self.validate("class B(t, y=9, *args, **kwargs): pass")
class TestParserIdempotency(support.TestCase):
"""A cut-down version of pytree_idempotency.py."""
# Issue 13125
@unittest.expectedFailure
def test_all_project_files(self):
for filepath in support.all_project_files():
with open(filepath, "rb") as fp:
encoding = tokenize.detect_encoding(fp.readline)[0]
self.assertIsNotNone(encoding,
"can't detect encoding for %s" % filepath)
with open(filepath, "r", encoding=encoding) as fp:
source = fp.read()
try:
tree = driver.parse_string(source)
except ParseError as err:
if verbose > 0:
warnings.warn('ParseError on file %s (%s)' % (filepath, err))
continue
new = str(tree)
x = diff(filepath, new)
if x:
self.fail("Idempotency failed: %s" % filepath)
def test_extended_unpacking(self):
driver.parse_string("a, *b, c = x\n")
driver.parse_string("[*a, b] = x\n")
driver.parse_string("(z, *y, w) = m\n")
driver.parse_string("for *z, m in d: pass\n")
class TestLiterals(GrammarTest):
def validate(self, s):
driver.parse_string(support.dedent(s) + "\n\n")
def test_multiline_bytes_literals(self):
s = """
md5test(b"\xaa" * 80,
(b"Test Using Larger Than Block-Size Key "
b"and Larger Than One Block-Size Data"),
"6f630fad67cda0ee1fb1f562db3aa53e")
"""
self.validate(s)
def test_multiline_bytes_tripquote_literals(self):
s = '''
b"""
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN">
"""
'''
self.validate(s)
def test_multiline_str_literals(self):
s = """
md5test("\xaa" * 80,
("Test Using Larger Than Block-Size Key "
"and Larger Than One Block-Size Data"),
"6f630fad67cda0ee1fb1f562db3aa53e")
"""
self.validate(s)
def diff(fn, result):
try:
with open('@', 'w') as f:
f.write(str(result))
fn = fn.replace('"', '\\"')
return subprocess.call(['diff', '-u', fn, '@'], stdout=(subprocess.DEVNULL if verbose < 1 else None))
finally:
try:
os.remove("@")
except OSError:
pass