Commit e3c2ad6d authored by Tim Peters's avatar Tim Peters

Merge rev 29448 from ZODB trunk.

Convert some XXXs.  More to come.
parent 4fe3a674
......@@ -760,8 +760,7 @@ earlier names:
The in-memory, per-connection object cache (pickle cache) was changed
to participate in garbage collection. This should reduce the number
of memory leaks, although we are still tracking a few problems. [XXX
might say more here]
of memory leaks, although we are still tracking a few problems.
Multi-version concurrency control
---------------------------------
......@@ -1309,7 +1308,7 @@ The ZEO cache verification protocol was revised to require many fewer
messages in cases where a client or server restarts quickly.
The performance of full cache verification has improved dramatically.
XXX Get measurements from Jim -- somewhere in 2x-5x recall. The
Measurements from Jim were somewhere in 2x-5x. The
implementation was fixed to use the very-fast getSerial() method on
the storage instead of the comparatively slow load().
......@@ -1667,7 +1666,7 @@ The ZEO version number was bumped to 2.0.2 on account of the following
minor feature additions.
The performance of full cache verification has improved dramatically.
XXX Get measurements from Jim -- somewhere in 2x-5x recall. The
Measurements from Jim were somewhere in 2x-5x. The
implementation was fixed to use the very-fast getSerial() method on
the storage instead of the comparatively slow load().
......
......@@ -96,7 +96,7 @@ type VALUE_TYPE. The macro returns an int, with value
< 0 if X < Y
== 0 if X == Y
> 0 if X > Y
XXX There is no provision for determining whether the comparison
Bug: There is no provision for determining whether the comparison
attempt failed (set a Python exception).
DECREF_VALUE(K)
......
......@@ -79,9 +79,9 @@ merge_error(int p1, int p2, int p3, int reason)
* Mapping value modification: s2 or s3 can modify the value associated
* with a key in s1, provided the other transaction doesn't make a
* modification of the same key to a different value. It's OK if s2 and s3
* both give the same new value to the key (XXX while it's hard to be
* precise about why, this doesn't seem consistent with that it's *not* OK
* for both to add a new key mapping to the same value).
* both give the same new value to the key while it's hard to be precise about
* why, this doesn't seem consistent with that it's *not* OK for both to add
* a new key mapping to the same value).
*/
static PyObject *
bucket_merge(Bucket *s1, Bucket *s2, Bucket *s3)
......@@ -118,7 +118,7 @@ bucket_merge(Bucket *s1, Bucket *s2, Bucket *s3)
/* Consult zodb/btrees/interfaces.py for the meaning of the last
* argument passed to merge_error().
*/
/* XXX This isn't passing on errors raised by value comparisons. */
/* TODO: This isn't passing on errors raised by value comparisons. */
while (i1.position >= 0 && i2.position >= 0 && i3.position >= 0)
{
TEST_KEY_SET_OR(cmp12, i1.key, i2.key) goto err;
......
......@@ -38,12 +38,12 @@ class Base(TestCase):
def _getRoot(self):
if self.db is None:
# XXX On the next line, the ZODB4 flavor of this routine
# XXX passes a cache_size argument:
# Unclear: On the next line, the ZODB4 flavor of this routine
# [asses a cache_size argument:
# self.db = DB(MappingStorage(), cache_size=1)
# XXX If that's done here, though, testLoadAndStore() and
# XXX testGhostUnghost() both nail the CPU and seemingly
# XXX never finish.
# If that's done here, though, testLoadAndStore() and
# testGhostUnghost() both nail the CPU and seemingly
# never finish.
self.db = DB(MappingStorage())
return self.db.open().root()
......@@ -137,9 +137,9 @@ class MappingBase(Base):
for i in range(1000):
self.t[i] = i
r = repr(self.t)
# make sure the repr is 10000 bytes long for a bucket
# XXX since we the test is also run for btrees, skip the length
# XXX check if the repr starts with '<'
# Make sure the repr is 10000 bytes long for a bucket.
# But since the test is also run for btrees, skip the length
# check if the repr starts with '<'
if not r.startswith('<'):
self.assert_(len(r) > 10000)
......
......@@ -28,10 +28,10 @@ class SubclassTest(unittest.TestCase):
# of that type
t = T()
# XXX there's no good way to get a bucket at the moment.
# XXX __getstate__() is as good as it gets, but the default
# XXX getstate explicitly includes the pickle of the bucket
# XXX for small trees, so we have to be clever :-(
# There's no good way to get a bucket at the moment.
# __getstate__() is as good as it gets, but the default
# getstate explicitly includes the pickle of the bucket
# for small trees, so we have to be clever :-(
# make sure there is more than one bucket in the tree
for i in range(1000):
......
......@@ -216,7 +216,7 @@ class PersistenceTest(unittest.TestCase):
self.assertEqual(obj.curly, 2)
self.assertEqual(obj.moe, 3)
# XXX Need to decide how __setattr__ and __delattr__ should work,
# TODO: Need to decide how __setattr__ and __delattr__ should work,
# then write tests.
......
......@@ -189,5 +189,4 @@ def loop(timeout=30.0, use_poll=0, map=None):
##asyncore.loop = deprecated_loop
# XXX Remove this once we've updated ZODB4 since they share this package
asyncore.loop = loop
......@@ -16,7 +16,7 @@
This server acts as a front-end for one or more real storages, like
file storage or Berkeley storage.
XXX Need some basic access control-- a declaration of the methods
TODO: Need some basic access control-- a declaration of the methods
exported for invocation by the server.
"""
......@@ -47,8 +47,8 @@ from ZODB.loglevels import BLATHER
logger = logging.getLogger('ZEO.StorageServer')
# XXX This used to say "ZSS", which is now implied in the logger name, can this
# be either set to str(os.getpid()) (if that makes sense) or removed?
# TODO: This used to say "ZSS", which is now implied in the logger name.
# Can this be either set to str(os.getpid()) (if that makes sense) or removed?
_label = "" # default label used for logging.
def set_label():
......@@ -298,7 +298,7 @@ class ZEOStorage:
os = self.storage.getTid(oid)
except KeyError:
self.client.invalidateVerify((oid, ''))
# XXX It's not clear what we should do now. The KeyError
# It's not clear what we should do now. The KeyError
# could be caused by an object uncreation, in which case
# invalidation is right. It could be an application bug
# that left a dangling reference, in which case it's bad.
......@@ -564,7 +564,7 @@ class ZEOStorage:
# finishes as a transaction and finds another instance is in the
# _waiting list.
# XXX It might be better to have a mechanism to explicitly send
# It might be better to have a mechanism to explicitly send
# the finishing transaction's reply before restarting the waiting
# transaction. If the restart takes a long time, the previous
# client will be blocked until it finishes.
......
......@@ -103,7 +103,6 @@ class StatsClient(asyncore.dispatcher):
return len(self.buf)
def readable(self):
# XXX what goes here?
return 0
def handle_write(self):
......
......@@ -52,7 +52,7 @@ def log(msg, level=logging.INFO, exc_info=False):
def parse_address(arg):
# XXX Not part of the official ZConfig API
# Caution: Not part of the official ZConfig API.
obj = ZConfig.datatypes.SocketAddress(arg)
return obj.family, obj.address
......@@ -203,7 +203,7 @@ class ZEOServer:
transaction_timeout=self.options.transaction_timeout,
monitor_address=self.options.monitor_address,
auth_protocol=self.options.auth_protocol,
auth_database=self.options.auth_database, # XXX option spelling
auth_database=self.options.auth_database,
auth_realm=self.options.auth_realm)
def loop_forever(self):
......@@ -223,9 +223,9 @@ class ZEOServer:
sys.exit(1)
def handle_sigusr2(self):
# XXX this used to reinitialize zLOG. How do I achieve
# the same effect with Python's logging package?
# Should we restart as with SIGHUP?
# TODO: this used to reinitialize zLOG. How do I achieve
# the same effect with Python's logging package?
# Should we restart as with SIGHUP?
log("received SIGUSR2, but it was not handled!", level=logging.WARNING)
def close_storages(self):
......
......@@ -524,8 +524,8 @@ class InvalidationTests:
db1.close()
db2.close()
# XXX Temporarily disabled. I know it fails, and there's no point
# XXX getting an endless number of reports about that.
# TODO: Temporarily disabled. I know it fails, and there's no point
# getting an endless number of reports about that.
def xxxcheckConcurrentUpdatesInVersions(self):
self._storage = storage1 = self.openClientStorage()
db1 = DB(storage1)
......
......@@ -12,7 +12,7 @@
#
##############################################################################
"""A multi-client test of the ZEO storage server"""
# XXX This code is currently broken.
# TODO: This code is currently broken.
import ZODB, ZODB.DB, ZODB.FileStorage, ZODB.POSException
import persistent
......@@ -148,7 +148,7 @@ def main(client_func=None):
server.close()
os.waitpid(server_pid, 0)
# XXX Should check that the results are consistent!
# TODO: Should check that the results are consistent!
print "Total time:", t2 - t0
print "Server start time", t1 - t0
......
......@@ -16,7 +16,7 @@
The stress test should run in an infinite loop and should involve
multiple connections.
"""
# XXX This code is currently broken.
# TODO: This code is currently broken.
import transaction
import ZODB
......
......@@ -74,8 +74,8 @@ class MonitorTests(CommonSetupTearDown):
return """<mappingstorage 1/>"""
def testMonitor(self):
# just open a client to know that the server is up and running
# XXX should put this in setUp
# Just open a client to know that the server is up and running
# TODO: should put this in setUp.
self.storage = self.openClientStorage()
s = self.get_monitor_output()
self.storage.close()
......
......@@ -87,7 +87,7 @@ def get_port():
try:
s.connect(('localhost', port))
except socket.error:
# XXX check value of error?
# Perhaps we should check value of error too.
return port
finally:
s.close()
......@@ -112,7 +112,7 @@ class GenericTests(
"""Combine tests from various origins in one class."""
def setUp(self):
logger.info("setUp() %s", self.id()) # XXX is this really needed?
logger.info("setUp() %s", self.id())
port = get_port()
zconf = forker.ZEOConfig(('', port))
zport, adminaddr, pid, path = forker.start_zeo_server(self.getConfig(),
......@@ -136,7 +136,7 @@ class GenericTests(
os.waitpid(pid, 0)
def open(self, read_only=0):
# XXX Needed to support ReadOnlyStorage tests. Ought to be a
# Needed to support ReadOnlyStorage tests. Ought to be a
# cleaner way.
addr = self._storage._addr
self._storage.close()
......@@ -201,10 +201,6 @@ class MappingStorageTests(GenericTests):
def getConfig(self):
return """<mappingstorage 1/>"""
# XXX There are still a bunch of tests that fail. Are there
# still test classes in GenericTests that shouldn't be there?
# XXX Is the above comment still relevant?
test_classes = [FileStorageTests, MappingStorageTests]
def test_suite():
......
......@@ -123,7 +123,7 @@ class CacheTests(unittest.TestCase):
self.cache.store(n, "", n, None, data[51])
self.assert_(len(self.cache) < 51)
# XXX Need to make sure eviction of non-current data
# TODO: Need to make sure eviction of non-current data
# and of version data are handled correctly.
def testSerialization(self):
......
......@@ -72,7 +72,7 @@ def find_global(module, name):
if safe:
return r
# XXX what's a better way to do this? esp w/ 2.1 & 2.2
# TODO: is there a better way to do this?
if type(r) == types.ClassType and issubclass(r, Exception):
return r
......
......@@ -133,7 +133,7 @@ class SizedMessageAsyncConnection(asyncore.dispatcher):
def get_addr(self):
return self.addr
# XXX avoid expensive getattr calls? Can't remember exactly what
# TODO: avoid expensive getattr calls? Can't remember exactly what
# this comment was supposed to mean, but it has something to do
# with the way asyncore uses getattr and uses if sock:
def __nonzero__(self):
......@@ -193,7 +193,7 @@ class SizedMessageAsyncConnection(asyncore.dispatcher):
else:
msg_size = 4
state = 0
# XXX We call message_input() with __input_lock
# Obscure: We call message_input() with __input_lock
# held!!! And message_input() may end up calling
# message_output(), which has its own lock. But
# message_output() cannot call message_input(), so
......
......@@ -292,7 +292,7 @@ class InvalidObjectReference(POSError):
o A reference to an object in a different database connection.
XXX The exception ought to have a member that is the invalid object.
TODO: The exception ought to have a member that is the invalid object.
"""
class ConnectionStateError(POSError):
......
......@@ -245,7 +245,7 @@ class ITransaction(Interface):
class IConnection(Interface):
"""ZODB connection.
XXX: This interface is incomplete.
TODO: This interface is incomplete.
"""
def add(ob):
......
......@@ -454,7 +454,7 @@ class ConnectionObjectReader(BaseObjectReader):
# to create the instance w/o hitting the db, so go for it!
oid, klass = oid
obj = self._cache.get(oid, None) # XXX it's not a dict
obj = self._cache.get(oid, None)
if obj is not None:
return obj
......@@ -474,7 +474,7 @@ class ConnectionObjectReader(BaseObjectReader):
# current data in the object's actual record!
return self._conn.get(oid)
# XXX should be done by connection
# TODO: should be done by connection
obj._p_oid = oid
obj._p_jar = self._conn
# When an object is created, it is put in the UPTODATE
......
......@@ -127,12 +127,14 @@ class IteratorStorage(IteratorCompare):
self.assertEqual(count, 1)
def checkIterationIntraTransaction(self):
# XXX try this test with logging enabled. If you see something like
# TODO: Try this test with logging enabled. If you see something
# like
#
# ZODB FS FS21 warn: FileStorageTests.fs truncated, possibly due to
# damaged records at 4
#
# Then the code in FileIterator.next() hasn't yet been fixed.
# Should automate that check.
oid = self._storage.new_oid()
t = Transaction()
data = zodb_pickle(MinPO(0))
......
......@@ -188,7 +188,7 @@ class ExtStorageClientThread(StorageClientThread):
try:
iter = self.storage.iterator()
except AttributeError:
# XXX It's hard to detect that a ZEO ClientStorage
# It's hard to detect that a ZEO ClientStorage
# doesn't have this method, but does have all the others.
return
for obj in iter:
......
......@@ -32,7 +32,7 @@ class ReadOnlyStorage:
def checkReadMethods(self):
self._create_data()
self._make_readonly()
# XXX not going to bother checking all read methods
# Note that this doesn't check _all_ read methods.
for oid in self.oids.keys():
data, revid = self._storage.load(oid, '')
self.assertEqual(revid, self.oids[oid])
......
......@@ -107,7 +107,7 @@ class RevisionStorage:
self.assertEqual(end, revs[12])
# XXX Is it okay to assume everyone testing against RevisionStorage
# Unsure: Is it okay to assume everyone testing against RevisionStorage
# implements undo?
def checkLoadBeforeUndo(self):
......@@ -172,4 +172,4 @@ class RevisionStorage:
results = self._storage.loadBefore(oid2, revid2)
eq(results, None)
# XXX There are other edge cases to handle, including pack.
# TODO: There are other edge cases to handle, including pack.
......@@ -80,7 +80,7 @@ def zodb_unpickle(data):
klass_info = u.load()
if isinstance(klass_info, types.TupleType):
if isinstance(klass_info[0], type):
# XXX what is the second part of klass_info?
# Unclear: what is the second part of klass_info?
klass, xxx = klass_info
assert not xxx
else:
......@@ -144,7 +144,7 @@ def import_helper(name):
class StorageTestBase(unittest.TestCase):
# XXX It would be simpler if concrete tests didn't need to extend
# It would be simpler if concrete tests didn't need to extend
# setUp() and tearDown().
def setUp(self):
......@@ -210,7 +210,7 @@ class StorageTestBase(unittest.TestCase):
def _undo(self, tid, expected_oids=None, note=None):
# Undo a tid that affects a single object (oid).
# XXX This is very specialized
# This is very specialized.
t = transaction.Transaction()
t.note(note or "undo")
self._storage.tpc_begin(t)
......
......@@ -142,4 +142,4 @@ class SynchronizedStorage:
self._storage.tpc_begin(t)
self._storage.tpc_abort(t)
# XXX how to check undo?
# TODO: how to check undo?
......@@ -107,8 +107,8 @@ class DBMethods(CacheTestBase):
for k, v in dict.items():
self.assert_(k in expected)
# XXX not really sure how to do a black box test of the cache.
# should the full sweep and minimize calls always remove things?
# TODO: not really sure how to do a black box test of the cache.
# Should the full sweep and minimize calls always remove things?
def checkFullSweep(self):
old_size = self.db.cacheSize()
......@@ -175,10 +175,9 @@ class DBMethods(CacheTestBase):
self.fail("cacheMinimize still running after 30 seconds -- "
"almost certainly in an infinite loop")
# XXX don't have an explicit test for incrgc, because the
# connection and database call it internally
# XXX same for the get and invalidate methods
# TODO: don't have an explicit test for incrgc, because the
# connection and database call it internally.
# Same for the get and invalidate methods.
def checkLRUitems(self):
# get a cache
......@@ -264,7 +263,7 @@ class LRUCacheTests(CacheTestBase):
gc.collect()
# XXX The above gc.collect call is necessary to make this test
# Obscure: The above gc.collect call is necessary to make this test
# pass.
#
# This test then only works because the order of computations
......
......@@ -142,10 +142,10 @@ class ConnectionDotAdd(unittest.TestCase):
class UserMethodTests(unittest.TestCase):
# XXX add isn't tested here, because there are a bunch of traditional
# add isn't tested here, because there are a bunch of traditional
# unit tests for it.
# XXX the version tests would require a storage that supports versions
# The version tests would require a storage that supports versions
# which is a bit more work.
def test_root(self):
......
......@@ -56,7 +56,7 @@ class DBTests(unittest.TestCase):
self.db.setCacheDeactivateAfter, 12)
self.assertRaises(DeprecationWarning,
self.db.setVersionCacheDeactivateAfter, 12)
# XXX There is no API call for removing the warning we just
# Obscure: There is no API call for removing the warning we just
# added, but filters appears to be a public variable.
del warnings.filters[0]
self.db.setCacheSize(15)
......
......@@ -36,11 +36,11 @@ class DemoStorageTests(StorageTestBase.StorageTestBase,
pass
def checkAbortVersionNonCurrent(self):
# XXX Need to implement a real loadBefore for DemoStorage?
# TODO: Need to implement a real loadBefore for DemoStorage?
pass
def checkLoadBeforeVersion(self):
# XXX Need to implement a real loadBefore for DemoStorage?
# TODO: Need to implement a real loadBefore for DemoStorage?
pass
# the next three pack tests depend on undo
......
......@@ -104,7 +104,7 @@ class TimeStampTests(unittest.TestCase):
ts2 = ts.laterThan(ts)
self.assert_(ts2 > ts)
# XXX should test for bogus inputs to TimeStamp constructor
# TODO: should test for bogus inputs to TimeStamp constructor
def checkTimeStamp(self):
# Alternate test suite
......
......@@ -11,8 +11,7 @@
# FOR A PARTICULAR PURPOSE.
#
##############################################################################
"""XXX short summary goes here.
"""
$Id$
"""
import unittest
......
......@@ -89,7 +89,7 @@ monotonically increasing, so the first one seen during the current
transaction remains the high-water mark for the duration of the
transaction.
XXX We'd like simple abort and commit calls to make txn boundaries,
We'd like simple abort and commit calls to make txn boundaries,
but that doesn't work unless an object is modified. sync() will abort
a transaction and process invalidations.
......
......@@ -254,7 +254,7 @@ class IPersistentNoReadConflicts(IPersistent):
conflicts for this object.
"""
# XXX TODO: document conflict resolution.
# TODO: document conflict resolution.
class IPersistentDataManager(Interface):
"""Provide services for managing persistent state.
......
......@@ -47,7 +47,7 @@ class Test(unittest.TestCase):
self.assertEqual(p._p_changed, 1)
self.assertEqual(dm.called, 1)
del p._p_changed
# XXX deal with current cPersistence implementation
# deal with current cPersistence implementation
if p._p_changed != 3:
self.assertEqual(p._p_changed, None)
self.assertEqual(dm.called, 1)
......@@ -164,7 +164,7 @@ class Test(unittest.TestCase):
p._p_jar = dm
p._p_changed = 0
p._p_deactivate()
# XXX does this really test the activate method?
# Unsure: does this really test the activate method?
p._p_activate()
self.assertEqual(p._p_changed, 0)
self.assertEqual(p.x, 42)
......@@ -266,10 +266,10 @@ class Test(unittest.TestCase):
self.assert_(P.__dictoffset__ < P.__weakrefoffset__)
self.assert_(P.__basicsize__ > Persistent.__basicsize__)
# XXX Can anyone defend/explain the test below? The tests classes defined here
# don't define __call__, so this weird test will always pass, but to what
# end? What the heck is the point. If a klass is given that happens
# to define __call__, the test *may* mysteriously fail. Who cares?
# Unsure: Can anyone defend/explain the test below? The tests classes defined
# here don't define __call__, so this weird test will always pass, but to what
# end? If a klass is given that happens to define __call__, the test *may*
# mysteriously fail. Who cares?
## def testDeactivateErrors(self):
## p = self.klass()
......
......@@ -251,7 +251,7 @@ class PersistenceTest(unittest.TestCase):
self.assertEqual(obj.curly, 2)
self.assertEqual(obj.moe, 3)
# XXX Need to decide how __setattr__ and __delattr__ should work,
# TODO: Need to decide how __setattr__ and __delattr__ should work,
# then write tests.
......
......@@ -20,10 +20,10 @@ def find_paths(root, maxdist):
from the root, looking for persistent objects. Return a dict
mapping oids to traversal paths.
XXX Assumes that the keys of the root are not themselves
TODO: Assumes that the keys of the root are not themselves
persistent objects.
XXX Doesn't traverse containers.
TODO: Doesn't traverse containers.
"""
paths = {}
......
......@@ -473,9 +473,6 @@ def do_backup(options):
# then perhaps the file was packed at some point (or a
# non-transactional undo was performed, but this is deprecated). Only
# do a full backup if forced to.
#
# XXX For ZODB4, this needs to take into account the storage metadata
# header that FileStorage has grown at the front of the file.
if reposum == srcsum_backedup:
log('doing incremental, starting at: %s', reposz)
do_incremental_backup(options, reposz, repofiles)
......
......@@ -27,7 +27,7 @@ class IDataManager(Interface):
This is currently implemented by ZODB database connections.
XXX This exists to document ZODB4 behavior, to help create some
This exists to document ZODB4 behavior, to help create some
backward-compatability support for Zope 3. New classes shouldn't
implement this. They should implement ZODB.interfaces.IDataManager
for now. Our hope is that there will eventually be an interface
......
......@@ -93,7 +93,7 @@ Test harness.
This requires that Python was built --with-pydebug.
-T
Use the trace module from Python for code coverage. XXX This only works
Use the trace module from Python for code coverage. This only works
if trace.py is explicitly added to PYTHONPATH. The current utility writes
coverage files to a directory named `coverage' that is parallel to
`build'. It also prints a summary to stdout.
......@@ -208,8 +208,8 @@ class ImmediateTestResult(unittest._TextTestResult):
print "The following test left garbage:"
print test
print gc.garbage
# XXX Perhaps eat the garbage here, so that the garbage isn't
# printed for every subsequent test.
# TODO: Perhaps eat the garbage here, so that the garbage isn't
# |printed for every subsequent test.
def print_times(self, stream, count=None):
results = self._testtimes.items()
......@@ -244,7 +244,7 @@ class ImmediateTestResult(unittest._TextTestResult):
if self.showAll:
self.stream.write(": ")
elif self._progressWithNames:
# XXX will break with multibyte strings
# TODO: will break with multibyte strings.
name = self.getShortDescription(test)
width = len(name)
if width < self._lastWidth:
......@@ -641,7 +641,7 @@ def configure_logging():
import logging.config
# Get the log.ini file from the current directory instead of possibly
# buried in the build directory. XXX This isn't perfect because if
# buried in the build directory. This isn't perfect because if
# log.ini specifies a log file, it'll be relative to the build directory.
# Hmm...
logini = os.path.abspath("log.ini")
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment