exceptions that can't be raised any further, because (for instance) they
occur in __del__ methods. The coroutine tests in test_generators was
triggering this leak. Remove the leakers' testcase, and add a simpler
testcase that explicitly tests this leak to test_generators.
test_generators now no longer leaks at all, on my machine. This fix may also
solve other leaks, but my full refleakhunting run is still busy, so who
knows?
not be tracked by GC. This fixes 254 of test_generators' refleaks on my
machine, but I'm sure something else will make them come back :>
Not adding a separate test for this kind of cycle, since the existing
fib/m235 already test them in more extensive ways than any 'minimal' test
has been able to manage.
examples no longer require any explicit closing to avoid
leaking.
That the tee-based examples still do is (I think) still a
mystery. Part of the mystery is that gc.garbage remains
empty: if it were the case that some generator in a trash
cycle said it needed finalization, suppressing collection
of that cycle, that generator _would_ show up in gc.garbage.
So this is acting more like, e.g., some tp_traverse slot
isn't visiting all the pointers it should (in which case
the skipped pointer(s) would act like an external root,
silently suppressing collection of everything reachable
from it(them)).
problems: first, PyGen_NeedsFinalizing() had an off-by-one bug that
prevented it from ever saying a generator didn't need finalizing, and
second, frame objects cleared themselves in a way that caused their
owning generator to think they were still executable, causing a double
deallocation of objects on the value stack if there was still a loop
on the block stack. This revision also removes some unnecessary
close() operations from test_generators that are now appropriately
handled by the cycle collector.
an incremental encoder that must retain part of the data between calls
to the encode() method.
Fix the incremental encoder and decoder for the IDNA encoding.
This closes SF patch #1453235.
The test case came from test_generators, not test_itertools.
Ensure there's no cyclic garbage we are counting.
This is weird because it leaks, then reaches a limit:
python.exe -i test_tee.py
>>> leak()
0
[26633 refs]
>>> leak()
0
[26658 refs]
>>> leak()
0
[26683 refs]
>>> leak()
0
[26708 refs]
>>> leak()
0
[26708 refs]
>>> leak()
0
[26708 refs]
>>> leak()
0
appear. Get rid of them by nuking doctest's default DocTestRunner
instance as part of cleanup(). Also cleanup() before running the
first test repetition (the test was run once before we get into
the -R branch).
This may be causing the debian sparc buildbot to fail.
Print a little message to let the user ^w buildbot know it's still thinking.
We may want to adjust the time period which is currently 5 minutes.
Will backport.
prepends the exception's module name to non-builtin exceptions, like
the interpreter itself does.
broke a number of doctests. should be discussed before checking in (see
discussion on python-dev).
bsddb.*open() methods cachesize parameter wouldn't work (raised an
internal bsddb.db exception when it was given). The set_cachesize
call needed to be moved from the DB object to the DBEnv since the env
was introduced to allow for threading.
(will backport to 2.4)
Using None for a filename with the 'n' flag when calling bsddb.btopen
would cause an error while checking if the file None existed. error
not likely to be seen as anyone using None for a filename would likely
use the 'c' flag in the first place.
tests. Alas, because only the "x86 OpenBSD trunk" buildbot fails
these tests, and test_descr stops after the first failure, there's
no sane way for me to fix these short of fixing one and then
waiting for the buildbot to reveal the next one.
to that id() can now return a Python long on a 32-bit box that allocates
addresses "with the sign bit set".
test_set.py test_subclass_with_custom_hash(): it's never been portably
legal for a __hash__() method to return id(self), but on 32-bit boxes
that never caused a problem before it became possible for id() to
return a Python long. Changed __hash__ here to return a Python int
regardless of platform.
test_descr.py specials():
vereq(hash(c1), id(c1))
has never been a correct test -- just removed it (hash() is always
a Python int; id() may be a Python long).
default decimal context, causing test_tokenize to fail
if it ran after test_contextlib. Changed to restore
the decimal context in effect at the test's start.
soon after because the gmail address it connects to started timing
out on all the buildbot slaves. Rewrote the test to produce a
warning message (instead of failing) when the address times out.
Also removed the special case for Windows -- this test started to
work on Windows as soon as bug 1462352 was fixed.
mouse events. This makes the test fail. Catch that case and don't run
the tests. Should make the debian/ubuntu buildbots that run in a chroot
work again.
Will backport to release24-maint.
least as big as a long. I believe this to be a safe assumption that is being
made in many parts of CPython, but a check could be added.
len(xrange(sys.maxint)) works now, so fix the testsuite's odd exception for
64-bit platforms too. It also fixes 'zip(xrange(sys.maxint), it)' as a
portable-ish (if expensive) alternative to enumerate(it); since zip() now
calls len(), this was breaking on (real) 64-bit platforms. No additional
test was added for that behaviour.
- The buildbot "fetch it" step failed at the end, due to
using Unix syntax in the final "copy the DLL" step.
test_sqlite was skipped as a result.
- test_sqlite is no longer an expected skip on Windows.
Re-enable all the tests in test_trace.py except one. Still not sure that these tests test what they used to test, but they pass. One failing test seems to be caused by undocumented line number table behavior in Python 2.4.
tracing/line number table in except blocks.
Reflow long lines introduced by col_offset changes. Update test_ast
to handle new fields in excepthandler.
As note in Python.asdl says, we might want to rethink how attributes
are handled. Perhaps they should be the same as other fields, with
the primary difference being how they are defined for all types within
a sum.
Also fix asdl_c so that constructors with int fields don't fail when
passed a zero value.
converted to CR CR NL. There may be a way to fix this with tcsetattr,
but I couldn't find it. There was a similar problem on IRIX.
Just normalize the output and compare that.
Will backport.
like cause the interpreter to exit abruptly. If there's a way to fix this,
it would be good to really fix it. It could just be the operation of the
std C library and we just aren't supposed to do that.
When the test case is skipped, we print a message so the user can check
for themselves.
This is based on pysqlite2.1.3, and provides a DB-API interface in
the standard library. You'll need sqlite 3.2.2 or later to build
this - if you have an earlier version, the C extension module will
not be built.
we are about to leave behind. An example of the cause of this leak can be
found in the leakers directory, in case we ever want to tackle the
underlying problem.
effect at the time test_decimal was imported. Else
running test_decimal had the bad side effect of
permanently changing the decimal context in effect.
That caused text_tokenize to fail if it ran after
test_decimal.
- The doctests in decistmt() weren't run at all when
test_tokenize was run via regrtest.py.
- Some expected output in decistmt() was Windows-specific
(but nobody noticed because the doctests weren't getting
run).
- test_roundtrip() didn't actually test anything when
running the tests with -O. Now it does.
- Changed test_roundtrip() to show the name of the input
file when it fails. That would have saved a lot of
time earlier today.
- Added a bunch of comments.
adds the following API calls: PySet_Clear(), _PySet_Next(), and
_PySet_Update(). The latter two are considered non-public. Tests and
documentation (for the public API) are included.
itertools.tee->instance->attribute->itertools.tee and
itertools.tee->teedataobject->itertools.tee cycles, which can be found now
that itertools.tee and its teedataobject participate in GC, remain findable
and cleanable. The test won't fail when they aren't, but at least the
frequent hunt-refleaks runs would spot the rise in refleaks.
def foo((x)): was getting recognized as requiring tuple unpacking
which is not correct.
Add tests for this case and the proper way to unpack a tuple of one:
def foo((x,)):
test_inpsect was incorrect before. I'm not sure why it was passing,
but that has been corrected with a test for both functions above.
This means the test (and therefore inspect.getargspec()) are broken in 2.4.
that yields after a throw(). Make @contextmanager not reraise
exceptions, but return a false value in that case instead. Add test
cases for both behaviors.
as diagnosed by Nick Coghlan.
test_capi.py: A test module should never spawn a thread as
a side effect of being imported. Because this one did, the
segfault one of its thread tests caused didn't occur until
a few tests after test_regrtest.py thought test_capi was
finished. Repair that. Also join() the thread spawned
at the end, so that test_capi is truly finished when
regrtest reports that it's done.
_testcapimodule.c test_thread_state(): this spawns a
couple of non-threading.py threads, passing them a PyObject*
argument, but did nothing to ensure that those threads
finished before returning. As a result, the PyObject*
_could_ (although this was unlikely) get decref'ed out of
existence before the threads got around to using it.
Added explicit synchronization (via a Python mutex) so
that test_thread_state can reliably wait for its spawned
threads to finish.
This was a fair amount of rework of the patch. Refactored test_fork1 so it
could be reused by the new tests for wait3/4. Also made them into new style
unittests (derive from unittest.TestCase).
This patch adds a-LAW encoding to audioop and replaces the old
u-LAW encoding/decoding code with the current code from sox.
Possible issues: the code from sox uses int16_t.
Code by Lars Immisch
of tuple) that provides incremental decoders and encoders (a way to use
stateful codecs without the stream API). Functions
codecs.getincrementaldecoder() and codecs.getincrementalencoder() have
been added.
Since it's never intended that this script be run by
regrtest.py, it shouldn't have been named with a "test_"
prefix to begin with. A consequence is that we shouldn't
see useless:
test_hashlib_speed skipped -- not a unit test (stand alone benchmark)
lines in regrtest output anymore.
Anyway, this is the changes to the with-statement
so that __exit__ must return a true value in order
for a pending exception to be ignored.
The PEP (343) is already updated.
an error code, this let `self` leak. This is a disaster
on Windows, since `self` already points to a newly-opened
file object, and it was impossible for Python code to
close the thing since the only reference to it was in a
blob of leaked C memory.
test_hotshot test_bad_sys_path(): This new test provoked
the C bug above. This test passed, but left an open
"@test" file behind, which caused a massive cascade of
bogus test failures in later, unrelated tests on Windows.
Changed the test code to remove the @test file it leaves
behind, which relies on the change above to close that
file first.
The failure definitely seems timing related. This change *seems* to work.
Since the failure isn't doesn't occur consistently, it's hard to tell.
Running these tests on Solaris in this order:
test_urllibnet test_operator test_cgi \
test_isinstance test_future test_ast test_logging
generally caused a failure (about 50% of the time) before the sleep.
I couldn't provoke the failure with the sleep.
This should really be cleaned up by using threading.Events or something
so it is not timing dependent and doesn't hang forever on failure.
test_codecmaps_tw test_importhooks test_socket_ssl
I don't completely understand the cause, but there's a lot of import magic
going on and this is the smallest change which fixes the problem.
want to wait forever if we don't receive the last message. But we also
don't want the test to fail if we shutdown too quickly. I can't reliably
reproduce this failure, so I'm kinda guessing this is the problem.
We'll see if this band-aid helps.
The culprit was an expression-less yield -- the first apparently in
the standard library. I added a unit test for this.
Also removed the hack to force compilation of test_with.py.
added message attribute compared to the previous version of Exception. It is
also a new-style class, making all exceptions now new-style. KeyboardInterrupt
and SystemExit inherit from BaseException directly. String exceptions now
raise DeprecationWarning.
Applies patch 1104669, and closes bugs 1012952 and 518846.
- New semantics for __exit__() -- it must re-raise the exception
if type is not None; the with-statement itself doesn't do this.
(See the updated PEP for motivation.)
- Added context managers to:
- file
- thread.LockType
- threading.{Lock,RLock,Condition,Semaphore,BoundedSemaphore}
- decimal.Context
- Added contextlib.py, which defines @contextmanager, nested(), closing().
- Unit tests all around; bot no docs yet.
- IMPORT_NAME takes an extra argument from the stack: the relativeness of
the import. Only passed to __import__ when it's not -1.
- __import__() takes an optional 5th argument for the same thing; it
__defaults to -1 (old semantics: try relative, then absolute)
- 'from . import name' imports name (be it module or regular attribute)
from the current module's *package*. Likewise, 'from .module import name'
will import name from a sibling to the current module.
- Importing from outside a package is not allowed; 'from . import sys' in a
toplevel module will not work, nor will 'from .. import sys' in a
(single-level) package.
- 'from __future__ import absolute_import' will turn on the new semantics
for import and from-import: imports will be absolute, except for
from-import with dots.
Includes tests for regular imports and importhooks, parser changes and a
NEWS item, but no compiler-package changes or documentation changes.
This was started by Mike Bland and completed by Guido
(with help from Neal).
This still needs a __future__ statement added;
Thomas is working on Michael's patch for that aspect.
There's a small amount of code cleanup and refactoring
in ast.c, compile.c and ceval.c (I fixed the lltrace
behavior when EXT_POP is used -- however I had to make
lltrace a static global).
breaks the parser module, because it adds the if/else construct as well as
two new grammar rules for backward compatibility. If no one else fixes
parsermodule, I guess I'll go ahead and fix it later this week.
The TeX code was checked with texcheck.py, but not rendered. There is
actually a slight incompatibility:
>>> (x for x in lambda:0)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: iteration over non-sequence
changes into
>>> (x for x in lambda: 0)
File "<stdin>", line 1
(x for x in lambda: 0)
^
SyntaxError: invalid syntax
Since there's no way the former version can be useful, it's probably a
bugfix ;)
- The copy module now "copies" function objects (as atomic objects).
- dict.__getitem__ now looks for a __missing__ hook before raising
KeyError.
- Added a new type, defaultdict, to the collections module.
This uses the new __missing__ hook behavior added to dict (see above).
Google for: eu_ES decimal point
shows that BSD locales had the eu_ES decimal point as
a single quote (') instead of a comma (,).
This was seems to have been fixed 15 months ago, but it's not on our
Mac and presumably others. So skip this broken locale.
test suite.
For urllib2, move the import of gopherlib into the
only function that uses it: users (including the
test suite) certainly shouldn't see a deprecation
warning just because they import urllib2! If they
actually use gopher_open(), fine, _then_ they should
see a deprecation warning.
test_file to fail on Windows in reality (can't delete
a still-open file), but a new bare "except:" hid that
test_file failed on Windows, and leaving behind the
still-open TESTFN caused a cascade of bogus failures
in later tests.
So, close the file, and stop hiding failure to unlink.
readline/readlines/read/readinto, loudly break by raising ValueError, rather
than silently deliver data out of order or hitting EOF prematurely.
Probably not a bugfix candidate, even though it affects no 'working' code.
Based on lsprof (patch #1212837) by Brett Rosen and Ted Czotter.
With further editing by Michael Hudson and myself.
History in svn repo: http://codespeak.net/svn/user/arigo/hack/misc/lsprof
* Module/_lsprof.c is the internal C module, Lib/cProfile.py a wrapper.
* pstats.py updated to display cProfile's caller/callee timings if available.
* setup.py and NEWS updated.
* documentation updates in the profiler section:
- explain the differences between the three profilers that we have now
- profile and cProfile can use a unified documentation, like (c)Pickle
- mention that hotshot is "for specialized usage" now
- removed references to the "old profiler" that no longer exists
* test updates:
- extended test_profile to cover delicate cases like recursion
- added tests for the caller/callee displays
- added test_cProfile, performing the same tests for cProfile
* TO-DO:
- cProfile gives a nicer name to built-in, particularly built-in methods,
which could be backported to profile.
- not tested on Windows recently!
Not sure why/how _handlers/_handlerList is out of sync. This could
indicate a deeper problem.
In test_logging, the only absolutely necessary change to get working
was tcpserver.abort = 1. But we don't want to wait infinitely
to join the threads, so give a 2.0 second timeout.
There doesn't appear to be a need for a local abort variable
in serve_until_stopped, so just use the instance member.
Note the problem is only on HEAD, not in 2.4.
on both Unix (SVR4 and BSD) and Windows. Restores behaviour of passing -1
for anonymous memory on Unix. Use MAP_ANONYMOUS instead of _ANON since
the latter is deprecated according to Linux (gentoo) man pages.
Should we continue to allow mmap.mmap(0, length) to work on Windows?
0 is a valid fd.
Will backport bugfix portions.
tty opened by os.openpty() isn't always a tty according to os.isatty(), when
it's tested inside the process that opened it. Doesn't affect actual
functionality, as using a tty this way is rarely, if ever, useful. Ignoring
the failure allows the test for actual functionality to continue.
Will backport to 2.4-maint.
Expand set of errors caught in set_context(). Some new errors, some
old error messages changed for consistency.
Fixed error checking in generator expression code. The first set of
tests were impossible condition given the grammar. In general, the
ast code uses REQ() for those sanity checks.
Fix some error handling for augmented assignments. As comments in the
code explain, set_context() ought to work here, but I got unexpected
crashes when I tried it. Should come back to this.
Add note to Grammar that yield expression is a special case.
Add doctest cases for SyntaxErrors raised by ast.c.
before the listener was ready (on gentoo x86 buildslave). This
caused the listener to not exit normally since nobody connected to it
(waited in accept()). The exception was raised in the other thread
and the test failed.
This fix doesn't completely eliminate the race, but should make it
near impossible to trigger. Hopefully it's good enough.
Thomas Wouters
Tim Peters
Phillip J. Eby
Walter Dörwald
Armin Rigo
Neal Norwitz
Georg Brandl
Anthony Baxter
Gregory P. Smith
Thomas Heller
Jeremy Hylton
Martin v. Löwis
Fred Drake
Barry Warsaw
Raymond Hettinger
Hye-Shik Chang
Trent Mick
Nick Coghlan
Guido van Rossum
Brett Cannon
Marc-André Lemburg
Vinay Sajip
Gustavo Niemeyer