cpython/Lib/importlib/test/source/test_file_loader.py

371 lines
15 KiB
Python
Raw Normal View History

import importlib
from importlib import _bootstrap
from .. import abc
from . import util as source_util
import imp
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
import marshal
import os
import py_compile
import stat
import sys
import unittest
2010-05-18 11:15:20 -03:00
from test.support import make_legacy_pyc
class SimpleTest(unittest.TestCase):
"""Should have no issue importing a source module [basic]. And if there is
a syntax error, it should raise a SyntaxError [syntax error].
"""
# [basic]
def test_module(self):
with source_util.create_modules('_temp') as mapping:
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
loader = _bootstrap._SourceFileLoader('_temp', mapping['_temp'])
module = loader.load_module('_temp')
self.assertTrue('_temp' in sys.modules)
check = {'__name__': '_temp', '__file__': mapping['_temp'],
'__package__': ''}
for attr, value in check.items():
self.assertEqual(getattr(module, attr), value)
def test_package(self):
with source_util.create_modules('_pkg.__init__') as mapping:
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
loader = _bootstrap._SourceFileLoader('_pkg',
mapping['_pkg.__init__'])
module = loader.load_module('_pkg')
self.assertTrue('_pkg' in sys.modules)
check = {'__name__': '_pkg', '__file__': mapping['_pkg.__init__'],
'__path__': [os.path.dirname(mapping['_pkg.__init__'])],
'__package__': '_pkg'}
for attr, value in check.items():
self.assertEqual(getattr(module, attr), value)
def test_lacking_parent(self):
with source_util.create_modules('_pkg.__init__', '_pkg.mod')as mapping:
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
loader = _bootstrap._SourceFileLoader('_pkg.mod',
mapping['_pkg.mod'])
module = loader.load_module('_pkg.mod')
self.assertTrue('_pkg.mod' in sys.modules)
check = {'__name__': '_pkg.mod', '__file__': mapping['_pkg.mod'],
'__package__': '_pkg'}
for attr, value in check.items():
self.assertEqual(getattr(module, attr), value)
def fake_mtime(self, fxn):
"""Fake mtime to always be higher than expected."""
return lambda name: fxn(name) + 1
def test_module_reuse(self):
with source_util.create_modules('_temp') as mapping:
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
loader = _bootstrap._SourceFileLoader('_temp', mapping['_temp'])
module = loader.load_module('_temp')
module_id = id(module)
module_dict_id = id(module.__dict__)
with open(mapping['_temp'], 'w') as file:
file.write("testing_var = 42\n")
# For filesystems where the mtime is only to a second granularity,
# everything that has happened above can be too fast;
# force an mtime on the source that is guaranteed to be different
# than the original mtime.
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
loader.path_mtime = self.fake_mtime(loader.path_mtime)
module = loader.load_module('_temp')
self.assertTrue('testing_var' in module.__dict__,
"'testing_var' not in "
"{0}".format(list(module.__dict__.keys())))
self.assertEqual(module, sys.modules['_temp'])
self.assertEqual(id(module), module_id)
self.assertEqual(id(module.__dict__), module_dict_id)
def test_state_after_failure(self):
# A failed reload should leave the original module intact.
attributes = ('__file__', '__path__', '__package__')
value = '<test>'
name = '_temp'
with source_util.create_modules(name) as mapping:
orig_module = imp.new_module(name)
for attr in attributes:
setattr(orig_module, attr, value)
with open(mapping[name], 'w') as file:
file.write('+++ bad syntax +++')
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
loader = _bootstrap._SourceFileLoader('_temp', mapping['_temp'])
with self.assertRaises(SyntaxError):
loader.load_module(name)
for attr in attributes:
self.assertEqual(getattr(orig_module, attr), value)
# [syntax error]
def test_bad_syntax(self):
with source_util.create_modules('_temp') as mapping:
with open(mapping['_temp'], 'w') as file:
file.write('=')
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
loader = _bootstrap._SourceFileLoader('_temp', mapping['_temp'])
with self.assertRaises(SyntaxError):
loader.load_module('_temp')
self.assertTrue('_temp' not in sys.modules)
class BadBytecodeTest(unittest.TestCase):
def import_(self, file, module_name):
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
loader = self.loader(module_name, file)
module = loader.load_module(module_name)
self.assertTrue(module_name in sys.modules)
def manipulate_bytecode(self, name, mapping, manipulator, *,
del_source=False):
"""Manipulate the bytecode of a module by passing it into a callable
that returns what to use as the new bytecode."""
try:
del sys.modules['_temp']
except KeyError:
pass
py_compile.compile(mapping[name])
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
if not del_source:
bytecode_path = imp.cache_from_source(mapping[name])
else:
os.unlink(mapping[name])
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
bytecode_path = make_legacy_pyc(mapping[name])
if manipulator:
with open(bytecode_path, 'rb') as file:
bc = file.read()
new_bc = manipulator(bc)
with open(bytecode_path, 'wb') as file:
if new_bc is not None:
file.write(new_bc)
return bytecode_path
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
def _test_empty_file(self, test, *, del_source=False):
with source_util.create_modules('_temp') as mapping:
bc_path = self.manipulate_bytecode('_temp', mapping,
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
lambda bc: b'',
del_source=del_source)
test('_temp', mapping, bc_path)
@source_util.writes_bytecode_files
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
def _test_partial_magic(self, test, *, del_source=False):
# When their are less than 4 bytes to a .pyc, regenerate it if
# possible, else raise ImportError.
with source_util.create_modules('_temp') as mapping:
bc_path = self.manipulate_bytecode('_temp', mapping,
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
lambda bc: bc[:3],
del_source=del_source)
test('_temp', mapping, bc_path)
def _test_magic_only(self, test, *, del_source=False):
with source_util.create_modules('_temp') as mapping:
bc_path = self.manipulate_bytecode('_temp', mapping,
lambda bc: bc[:4],
del_source=del_source)
test('_temp', mapping, bc_path)
def _test_partial_timestamp(self, test, *, del_source=False):
with source_util.create_modules('_temp') as mapping:
bc_path = self.manipulate_bytecode('_temp', mapping,
lambda bc: bc[:7],
del_source=del_source)
test('_temp', mapping, bc_path)
def _test_no_marshal(self, *, del_source=False):
with source_util.create_modules('_temp') as mapping:
bc_path = self.manipulate_bytecode('_temp', mapping,
lambda bc: bc[:8],
del_source=del_source)
file_path = mapping['_temp'] if not del_source else bc_path
with self.assertRaises(EOFError):
self.import_(file_path, '_temp')
def _test_non_code_marshal(self, *, del_source=False):
with source_util.create_modules('_temp') as mapping:
bytecode_path = self.manipulate_bytecode('_temp', mapping,
lambda bc: bc[:8] + marshal.dumps(b'abcd'),
del_source=del_source)
file_path = mapping['_temp'] if not del_source else bytecode_path
with self.assertRaises(ImportError):
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
self.import_(file_path, '_temp')
def _test_bad_marshal(self, *, del_source=False):
with source_util.create_modules('_temp') as mapping:
bytecode_path = self.manipulate_bytecode('_temp', mapping,
lambda bc: bc[:8] + b'<test>',
del_source=del_source)
file_path = mapping['_temp'] if not del_source else bytecode_path
with self.assertRaises(ValueError):
self.import_(file_path, '_temp')
def _test_bad_magic(self, test, *, del_source=False):
with source_util.create_modules('_temp') as mapping:
bc_path = self.manipulate_bytecode('_temp', mapping,
lambda bc: b'\x00\x00\x00\x00' + bc[4:])
test('_temp', mapping, bc_path)
class SourceLoaderBadBytecodeTest(BadBytecodeTest):
loader = _bootstrap._SourceFileLoader
@source_util.writes_bytecode_files
def test_empty_file(self):
# When a .pyc is empty, regenerate it if possible, else raise
# ImportError.
def test(name, mapping, bytecode_path):
self.import_(mapping[name], name)
with open(bytecode_path, 'rb') as file:
self.assertGreater(len(file.read()), 8)
self._test_empty_file(test)
def test_partial_magic(self):
def test(name, mapping, bytecode_path):
self.import_(mapping[name], name)
with open(bytecode_path, 'rb') as file:
self.assertGreater(len(file.read()), 8)
self._test_partial_magic(test)
@source_util.writes_bytecode_files
def test_magic_only(self):
# When there is only the magic number, regenerate the .pyc if possible,
# else raise EOFError.
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
def test(name, mapping, bytecode_path):
self.import_(mapping[name], name)
with open(bytecode_path, 'rb') as file:
self.assertGreater(len(file.read()), 8)
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
@source_util.writes_bytecode_files
def test_bad_magic(self):
# When the magic number is different, the bytecode should be
# regenerated.
def test(name, mapping, bytecode_path):
self.import_(mapping[name], name)
with open(bytecode_path, 'rb') as bytecode_file:
self.assertEqual(bytecode_file.read(4), imp.get_magic())
self._test_bad_magic(test)
@source_util.writes_bytecode_files
def test_partial_timestamp(self):
# When the timestamp is partial, regenerate the .pyc, else
# raise EOFError.
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
def test(name, mapping, bc_path):
self.import_(mapping[name], name)
with open(bc_path, 'rb') as file:
self.assertGreater(len(file.read()), 8)
@source_util.writes_bytecode_files
def test_no_marshal(self):
# When there is only the magic number and timestamp, raise EOFError.
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
self._test_no_marshal()
@source_util.writes_bytecode_files
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
def test_non_code_marshal(self):
self._test_non_code_marshal()
# XXX ImportError when sourceless
# [bad marshal]
@source_util.writes_bytecode_files
def test_bad_marshal(self):
# Bad marshal data should raise a ValueError.
self._test_bad_marshal()
# [bad timestamp]
@source_util.writes_bytecode_files
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
def test_old_timestamp(self):
# When the timestamp is older than the source, bytecode should be
# regenerated.
zeros = b'\x00\x00\x00\x00'
with source_util.create_modules('_temp') as mapping:
py_compile.compile(mapping['_temp'])
2010-04-16 21:19:56 -03:00
bytecode_path = imp.cache_from_source(mapping['_temp'])
with open(bytecode_path, 'r+b') as bytecode_file:
bytecode_file.seek(4)
bytecode_file.write(zeros)
self.import_(mapping['_temp'], '_temp')
source_mtime = os.path.getmtime(mapping['_temp'])
source_timestamp = importlib._w_long(source_mtime)
with open(bytecode_path, 'rb') as bytecode_file:
bytecode_file.seek(4)
self.assertEqual(bytecode_file.read(4), source_timestamp)
# [bytecode read-only]
@source_util.writes_bytecode_files
def test_read_only_bytecode(self):
# When bytecode is read-only but should be rewritten, fail silently.
with source_util.create_modules('_temp') as mapping:
# Create bytecode that will need to be re-created.
py_compile.compile(mapping['_temp'])
2010-04-16 21:19:56 -03:00
bytecode_path = imp.cache_from_source(mapping['_temp'])
with open(bytecode_path, 'r+b') as bytecode_file:
bytecode_file.seek(0)
bytecode_file.write(b'\x00\x00\x00\x00')
# Make the bytecode read-only.
os.chmod(bytecode_path,
stat.S_IRUSR | stat.S_IRGRP | stat.S_IROTH)
try:
# Should not raise IOError!
self.import_(mapping['_temp'], '_temp')
finally:
# Make writable for eventual clean-up.
os.chmod(bytecode_path, stat.S_IWUSR)
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
class SourcelessLoaderBadBytecodeTest(BadBytecodeTest):
loader = _bootstrap._SourcelessFileLoader
def test_empty_file(self):
def test(name, mapping, bytecode_path):
with self.assertRaises(ImportError):
self.import_(bytecode_path, name)
self._test_empty_file(test, del_source=True)
def test_partial_magic(self):
def test(name, mapping, bytecode_path):
with self.assertRaises(ImportError):
self.import_(bytecode_path, name)
self._test_partial_magic(test, del_source=True)
def test_magic_only(self):
def test(name, mapping, bytecode_path):
with self.assertRaises(EOFError):
self.import_(bytecode_path, name)
self._test_magic_only(test, del_source=True)
def test_bad_magic(self):
def test(name, mapping, bytecode_path):
with self.assertRaises(ImportError):
self.import_(bytecode_path, name)
self._test_bad_magic(test, del_source=True)
def test_partial_timestamp(self):
def test(name, mapping, bytecode_path):
with self.assertRaises(EOFError):
self.import_(bytecode_path, name)
self._test_partial_timestamp(test, del_source=True)
def test_no_marshal(self):
self._test_no_marshal(del_source=True)
def test_non_code_marshal(self):
self._test_non_code_marshal(del_source=True)
def test_main():
from test.support import run_unittest
Make importlib.abc.SourceLoader the primary mechanism for importlib. This required moving the class from importlib/abc.py into importlib/_bootstrap.py and jiggering some code to work better with the class. This included changing how the file finder worked to better meet import semantics. This also led to fixing importlib to handle the empty string from sys.path as import currently does (and making me wish we didn't support that instead just required people to insert '.' instead to represent cwd). It also required making the new set_data abstractmethod create any needed subdirectories implicitly thanks to __pycache__ (it was either this or grow the SourceLoader ABC to gain an 'exists' method and either a mkdir method or have set_data with no data arg mean to create a directory). Lastly, as an optimization the file loaders cache the file path where the finder found something to use for loading (this is thanks to having a sourceless loader separate from the source loader to simplify the code and cut out stat calls). Unfortunately test_runpy assumed a loader would always work for a module, even if you changed from underneath it what it was expected to work with. By simply dropping the previous loader in test_runpy so the proper loader can be returned by the finder fixed the failure. At this point importlib deviates from import on two points: 1. The exception raised when trying to import a file is different (import does an explicit file check to print a special message, importlib just says the path cannot be imported as if it was just some module name). 2. the co_filename on a code object is not being set to where bytecode was actually loaded from instead of where the marshalled code object originally came from (a solution for this has already been agreed upon on python-dev but has not been implemented yet; issue8611).
2010-07-03 18:48:25 -03:00
run_unittest(SimpleTest,
SourceLoaderBadBytecodeTest,
SourcelessLoaderBadBytecodeTest
)
if __name__ == '__main__':
test_main()