Merge pull request #4541 from elacuesta/pickle-adjustments

Pickle: use protocol 4, update tests
This commit is contained in:
Mikhail Korobov 2020-05-11 15:16:14 +05:00 committed by GitHub
commit 892467cb8a
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
6 changed files with 43 additions and 35 deletions

View File

@ -166,6 +166,7 @@ flake8-ignore =
scrapy/signalmanager.py E501
scrapy/spiderloader.py F841 E501
scrapy/squeues.py E128
scrapy/squeues.py E501
scrapy/statscollectors.py E501
# tests
tests/__init__.py E402 E501

View File

@ -250,7 +250,7 @@ class CsvItemExporter(BaseItemExporter):
class PickleItemExporter(BaseItemExporter):
def __init__(self, file, protocol=2, **kwargs):
def __init__(self, file, protocol=4, **kwargs):
super().__init__(**kwargs)
self.file = file
self.protocol = protocol

View File

@ -250,7 +250,7 @@ class DbmCacheStorage:
'headers': dict(response.headers),
'body': response.body,
}
self.db['%s_data' % key] = pickle.dumps(data, protocol=2)
self.db['%s_data' % key] = pickle.dumps(data, protocol=4)
self.db['%s_time' % key] = str(time())
def _read_data(self, spider, request):
@ -317,7 +317,7 @@ class FilesystemCacheStorage:
with self._open(os.path.join(rpath, 'meta'), 'wb') as f:
f.write(to_bytes(repr(metadata)))
with self._open(os.path.join(rpath, 'pickled_meta'), 'wb') as f:
pickle.dump(metadata, f, protocol=2)
pickle.dump(metadata, f, protocol=4)
with self._open(os.path.join(rpath, 'response_headers'), 'wb') as f:
f.write(headers_dict_to_raw(response.headers))
with self._open(os.path.join(rpath, 'response_body'), 'wb') as f:

View File

@ -26,7 +26,7 @@ class SpiderState:
def spider_closed(self, spider):
if self.jobdir:
with open(self.statefn, 'wb') as f:
pickle.dump(spider.state, f, protocol=2)
pickle.dump(spider.state, f, protocol=4)
def spider_opened(self, spider):
if self.jobdir and os.path.exists(self.statefn):

View File

@ -81,12 +81,11 @@ def _scrapy_non_serialization_queue(queue_class):
def _pickle_serialize(obj):
try:
return pickle.dumps(obj, protocol=2)
# Python <= 3.4 raises pickle.PicklingError here while
# 3.5 <= Python < 3.6 raises AttributeError and
# Python >= 3.6 raises TypeError
return pickle.dumps(obj, protocol=4)
# Both pickle.PicklingError and AttributeError can be raised by pickle.dump(s)
# TypeError is raised from parsel.Selector
except (pickle.PicklingError, AttributeError, TypeError) as e:
raise ValueError(str(e))
raise ValueError(str(e)) from e
PickleFifoDiskQueueNonRequest = _serializable_queue(

View File

@ -1,4 +1,5 @@
import pickle
import sys
from queuelib.tests import test_queue as t
from scrapy.squeues import (
@ -28,31 +29,13 @@ class TestLoader(ItemLoader):
def nonserializable_object_test(self):
q = self.queue()
try:
pickle.dumps(lambda x: x)
except Exception:
# Trigger Twisted bug #7989
import twisted.persisted.styles # NOQA
self.assertRaises(ValueError, q.push, lambda x: x)
else:
# Use a different unpickleable object
class A:
pass
a = A()
a.__reduce__ = a.__reduce_ex__ = None
self.assertRaises(ValueError, q.push, a)
self.assertRaises(ValueError, q.push, lambda x: x)
# Selectors should fail (lxml.html.HtmlElement objects can't be pickled)
sel = Selector(text='<html><body><p>some text</p></body></html>')
self.assertRaises(ValueError, q.push, sel)
class MarshalFifoDiskQueueTest(t.FifoDiskQueueTest):
chunksize = 100000
def queue(self):
return MarshalFifoDiskQueue(self.qpath, chunksize=self.chunksize)
class FifoDiskQueueTestMixin:
def test_serialize(self):
q = self.queue()
@ -66,6 +49,13 @@ class MarshalFifoDiskQueueTest(t.FifoDiskQueueTest):
test_nonserializable_object = nonserializable_object_test
class MarshalFifoDiskQueueTest(t.FifoDiskQueueTest, FifoDiskQueueTestMixin):
chunksize = 100000
def queue(self):
return MarshalFifoDiskQueue(self.qpath, chunksize=self.chunksize)
class ChunkSize1MarshalFifoDiskQueueTest(MarshalFifoDiskQueueTest):
chunksize = 1
@ -82,7 +72,7 @@ class ChunkSize4MarshalFifoDiskQueueTest(MarshalFifoDiskQueueTest):
chunksize = 4
class PickleFifoDiskQueueTest(MarshalFifoDiskQueueTest):
class PickleFifoDiskQueueTest(t.FifoDiskQueueTest, FifoDiskQueueTestMixin):
chunksize = 100000
@ -116,6 +106,21 @@ class PickleFifoDiskQueueTest(MarshalFifoDiskQueueTest):
self.assertEqual(r.url, r2.url)
assert r2.meta['request'] is r2
def test_non_pickable_object(self):
q = self.queue()
try:
q.push(lambda x: x)
except ValueError as exc:
if hasattr(sys, "pypy_version_info"):
self.assertIsInstance(exc.__context__, pickle.PicklingError)
else:
self.assertIsInstance(exc.__context__, AttributeError)
sel = Selector(text='<html><body><p>some text</p></body></html>')
try:
q.push(sel)
except ValueError as exc:
self.assertIsInstance(exc.__context__, TypeError)
class ChunkSize1PickleFifoDiskQueueTest(PickleFifoDiskQueueTest):
chunksize = 1
@ -133,10 +138,7 @@ class ChunkSize4PickleFifoDiskQueueTest(PickleFifoDiskQueueTest):
chunksize = 4
class MarshalLifoDiskQueueTest(t.LifoDiskQueueTest):
def queue(self):
return MarshalLifoDiskQueue(self.qpath)
class LifoDiskQueueTestMixin:
def test_serialize(self):
q = self.queue()
@ -150,7 +152,13 @@ class MarshalLifoDiskQueueTest(t.LifoDiskQueueTest):
test_nonserializable_object = nonserializable_object_test
class PickleLifoDiskQueueTest(MarshalLifoDiskQueueTest):
class MarshalLifoDiskQueueTest(t.LifoDiskQueueTest, LifoDiskQueueTestMixin):
def queue(self):
return MarshalLifoDiskQueue(self.qpath)
class PickleLifoDiskQueueTest(t.LifoDiskQueueTest, LifoDiskQueueTestMixin):
def queue(self):
return PickleLifoDiskQueue(self.qpath)