[issue32038] Add API to intercept socket.close()

STINNER Victor added the comment: >From my example: "finally: sock._io_refs -= 1" Oh, sock._decref_socketios() must be uesd here to really close the socket if it was closed in the meanwhile: def _decref_socketios(self): if self._io_refs > 0:

[issue32050] Deprecated python3 -x option

STINNER Victor added the comment: > @path\to\python -x %0 %1 %2 %3 %4 %5 %6 %7 %8 %9 Oh wow, I never saw that before. There is now a "py" launcher on Windows, do we still need such "hack"? -- components: +Windows -Interpreter Core nosy: +paul.moore,

[issue32038] Add API to intercept socket.close()

Yury Selivanov added the comment: > OK, then maybe socket objects should get a dup() method that just increments the ref counter, and that would be the public API you're looking for? "dup()" returns a new socket object, but here we only want to protect the original one.

[issue31701] faulthandler dumps 'Windows fatal exception: code 0xe06d7363'

STINNER Victor added the comment: New changeset 33217d22131ab1ffaa09a6944bae704f722f25a9 by Victor Stinner in branch '3.6': bpo-31701: faulthandler: ignore MSC and COM Windows exception (#3929) (#4416)

[issue31324] support._match_test() used by test.bisect is very inefficient

Change by Serhiy Storchaka : -- keywords: +patch pull_requests: +4368 stage: -> patch review ___ Python tracker ___

[issue30148] Pathological regex behaviour

Change by Serhiy Storchaka : -- nosy: +serhiy.storchaka resolution: -> wont fix stage: -> resolved status: open -> closed ___ Python tracker

[issue32038] Add API to intercept socket.close()

Yury Selivanov added the comment: > It's worse than a resource leak - the same file descriptor number could be > reused for a different file/socket, and then depending on the selector in > use, you could see the data from a completely different connection. I actually

[issue32051] Possible issue in multiprocessing doc

New submission from jason : in multiprocessing doc https://docs.python.org/3.6/library/multiprocessing.html under 17.2.2.7.2. Using a remote manager, >>> from multiprocessing.managers import BaseManager >>> import queue >>> queue = queue.Queue() >>> class

[issue32038] Add API to intercept socket.close()

STINNER Victor added the comment: The socket.socket (Python type) has a private _io_refs counter. close() does nothing until _io_refs reachs zero. Maybe we can develop an API to temporary increase the _io_refs counter to prevent a real close? Pseudo-code:

[issue32038] Add API to intercept socket.close()

Change by Yury Selivanov : -- Removed message: https://bugs.python.org/msg306374 ___ Python tracker ___

[issue32038] Add API to intercept socket.close()

Yury Selivanov added the comment: > Oh, sock._decref_socketios() must be uesd here to really close the socket if > it was closed in the meanwhile: There's also 'sock._closed' attribute that sock.close() and sock. _decref_socketios() interact with.. Alright, let me play

[issue32050] Deprecated python3 -x option

Serhiy Storchaka added the comment: The same as on DOS. Change the extension of the Python script from .py to .bat, and insert the following first line: @path\to\python -x %0 %1 %2 %3 %4 %5 %6 %7 %8 %9 Now you can run it as a bat-file. The command in the first

[issue32038] Add API to intercept socket.close()

Yury Selivanov added the comment: > The socket.socket (Python type) has a private _io_refs counter. close() does > nothing until _io_refs reachs zero. I didn't know about this. Looks like I can use '_io_refs' to fix some problems in uvloop, thanks Victor! The only

[issue31324] support._match_test() used by test.bisect is very inefficient

Serhiy Storchaka added the comment: I know :-) Unfortunately it is not obviously that these optimization actually optimize regular expressions. Due to the implementation particularities sometimes "unoptimized" code is faster than "optimized". I'm going to

[issue32050] Deprecated python3 -x option

New submission from STINNER Victor : Python has a -x option documented as a "DOS specific hack only": https://docs.python.org/dev/using/cmdline.html#cmdoption-x Python doesn't support MS-DOS since Python 2.1:

[issue32050] Deprecated python3 -x option

STINNER Victor added the comment: The note in the documentation is wrong: the line number is correct when using the -x option, even if the first line is skipped. Example with Python 2.7: --- $ cat x.py print(1) print(2) x $ python2 x.py 1 2 Traceback (most recent

[issue32040] Sorting pahtlib.Paths does give the same order as sorting the (string) filenames of that pathlib.Paths

QbLearningPython added the comment: Thanks, serhiy.storchaka, for your answer. I am not fully convinced. You have described the current behaviour of the pathlib package. But let me ask: should be this the desired behaviour? Since string filenames and pathlib.Paths are

[issue32050] Deprecated python3 -x option

STINNER Victor added the comment: > Python still supports Windows. What is the use case for this option on Windows? On Unix, the first line is usually used for the shebang, a line like: #!/usr/bin/env python3 But this shebang is seen as a comment in Python and

[issue32050] Deprecated python3 -x option

Change by STINNER Victor : -- pull_requests: +4371 ___ Python tracker ___ ___

[issue32038] Add API to intercept socket.close()

STINNER Victor added the comment: > Hmm... _decref_socketios() is not really a public API. I think it would be > nice to have an official way to deal with this, and a socket close callback > sounds reasonable to me. I dislike the idea of an event when a socket is

[issue31324] support._match_test() used by test.bisect is very inefficient

STINNER Victor added the comment: > There is good opportunity for optimizing the regular expression. All cases > have the literal 'test.test_asyncio.' prefix, and even longer literal > prefixes are common for tens cases. I implemented such code long time ago :-)

[issue32038] Add API to intercept socket.close()

Guido van Rossum added the comment: Perhaps you can just dup() the socket? That's what the ref counter is for IIRC. -- ___ Python tracker

[issue32038] Add API to intercept socket.close()

STINNER Victor added the comment: > Perhaps you can just dup() the socket? That's what the ref counter is for > IIRC. Sure. That's another option. But for asyncio with a lot of concurrent users, I'm not sure that it's ok to create a temporary file descriptor, since

[issue31324] support._match_test() used by test.bisect is very inefficient

Change by STINNER Victor : -- pull_requests: +4369 ___ Python tracker ___ ___

[issue32050] Deprecated python3 -x option

Serhiy Storchaka added the comment: Yes, of course. The py launcher is not an alternative to this "hack". It just allows you to specify just "py" (with possible options -2, -3, -3.6) instead of hardcoding the full name to the Python binary or add the path to

[issue32038] Add API to intercept socket.close()

Yury Selivanov added the comment: > Perhaps you can just dup() the socket? That's what the ref counter is for > IIRC. I already dup them for loop.create_server(sock=sock) and loop.create_connection(sock=sock) cases. Duping for sock_recv & friends will add too much

[issue32038] Add API to intercept socket.close()

Guido van Rossum added the comment: OK, then maybe socket objects should get a dup() method that just increments the ref counter, and that would be the public API you're looking for? -- ___ Python tracker

[issue32050] Deprecated python3 -x option

Change by STINNER Victor : -- keywords: +patch pull_requests: +4370 stage: -> patch review ___ Python tracker ___

[issue32050] Deprecated python3 -x option

Serhiy Storchaka added the comment: Python still supports Windows. -- nosy: +serhiy.storchaka ___ Python tracker ___

[issue32038] Add API to intercept socket.close()

Yury Selivanov added the comment: > Oh, sock._decref_socketios() must be uesd here to really close the socket if > it was closed in the meanwhile: For it to work correctly, socket objects should be initialized with "self._io_refs = 1". --

[issue32038] Add API to intercept socket.close()

Antoine Pitrou added the comment: Hmm... _decref_socketios() is not really a public API. I think it would be nice to have an official way to deal with this, and a socket close callback sounds reasonable to me. -- ___ Python

[issue32038] Add API to intercept socket.close()

Yury Selivanov added the comment: > Hmm... _decref_socketios() is not really a public API. I think it would be > nice to have an official way to deal with this, and a socket close callback > sounds reasonable to me. We can just make it public (after some renames) :)

[issue31701] faulthandler dumps 'Windows fatal exception: code 0xe06d7363'

STINNER Victor added the comment: I backported the fix to Python 3.6. @Steve Dower: What do you think of using SetUnhandledExceptionFilter? -- versions: +Python 3.7 ___ Python tracker

[issue23691] re.finditer iterator is not reentrant, but doesn't protect against nested calls to __next__

Change by Serhiy Storchaka : -- assignee: -> serhiy.storchaka nosy: +serhiy.storchaka stage: -> needs patch versions: +Python 2.7, Python 3.6, Python 3.7 -Python 3.4 ___ Python tracker

[issue23689] Memory leak in Modules/sre_lib.h

Change by Serhiy Storchaka : -- versions: +Python 3.6, Python 3.7 -Python 3.4, Python 3.5 ___ Python tracker ___

[issue32049] 2.7.14 does not uninstall cleanly if installation was run as SYSTEM account (SCCM)

New submission from Sven Pagel : If I install Python 2.7.14 on Windows (using the MSI) using the SYSTEM account, as is done by SCCM, uninstalling fails to clean up the registry so Python is still showing up as installed in the control panel (and still

[issue32046] 2to3 fix for operator.isCallable()

Change by Dong-hee Na : -- nosy: +corona10 ___ Python tracker ___ ___

[issue32047] asyncio: enable debug mode when -X dev is used

Change by STINNER Victor : -- keywords: +patch pull_requests: +4367 stage: -> patch review ___ Python tracker ___

[issue32047] asyncio: enable debug mode when -X dev is used

New submission from STINNER Victor : The bpo-32043 added a new "developer mode" enabled with the new -X dev command line option. I propose to enable asyncio debug mode with the global Python "developer mode". If an asyncio application is correctly written, the single

[issue24555] Python logic error when deal with re and muti-threading

Change by Serhiy Storchaka : -- resolution: -> duplicate stage: -> resolved status: open -> closed superseder: -> re functions never release GIL ___ Python tracker

[issue28951] re.flags not documented in Module Contents as promised.

Change by Serhiy Storchaka : Removed file: https://bugs.python.org/file45941/unnamed ___ Python tracker ___

[issue28951] re.flags not documented in Module Contents as promised.

Change by Serhiy Storchaka : -- components: +Regular Expressions nosy: +ezio.melotti, mrabarnett stage: -> needs patch versions: +Python 2.7, Python 3.6, Python 3.7 -Python 3.5 ___ Python tracker

[issue28951] re.flags not documented in Module Contents as promised.

Change by Serhiy Storchaka : Removed file: https://bugs.python.org/file45940/unnamed ___ Python tracker ___

[issue28450] Misleading/inaccurate documentation about unknown escape sequences in regular expressions

Serhiy Storchaka added the comment: Barry, could you please improve the documentation about unknown escape sequences in regular expressions? My skills is not enough for this. -- ___ Python tracker

[issue31701] faulthandler dumps 'Windows fatal exception: code 0xe06d7363'

STINNER Victor added the comment: > May I ask why AddVectoredExceptionHandler is used instead of > SetUnhandledExceptionFilter? Honestly, I don't know well Windows API and so I'm not sure which one is the best.

[issue32043] Add a new -X dev option: "developer mode"

STINNER Victor added the comment: Antoine Pitrou asked me to document the performance and memory overhead of -X dev: https://mail.python.org/pipermail/python-dev/2017-November/150578.html For example, I measured an increase of +50% on the peak memory usage (measured

[issue25054] Capturing start of line '^'

Change by Serhiy Storchaka : -- assignee: -> serhiy.storchaka nosy: +serhiy.storchaka versions: +Python 2.7, Python 3.7 -Python 3.5 ___ Python tracker

[issue32047] asyncio: enable debug mode when -X dev is used

STINNER Victor added the comment: I sent an email about this change to python-dev: https://mail.python.org/pipermail/python-dev/2017-November/150572.html -- ___ Python tracker

[issue28951] re.flags not documented in Module Contents as promised.

Change by Serhiy Storchaka : Removed file: https://bugs.python.org/file45934/unnamed ___ Python tracker ___

[issue32048] Misprint in the unittest.mock documentation.

Change by Артемий Родионов : -- assignee: docs@python components: Documentation nosy: docs@python, Артемий Родионов priority: normal severity: normal status: open title: Misprint in the unittest.mock documentation. type: enhancement versions: Python 3.7

[issue30004] in regex-howto, improve example on grouping

Serhiy Storchaka added the comment: Do you mind to create a pull request on GitHub Cristian? -- components: +Regular Expressions nosy: +ezio.melotti, mrabarnett, serhiy.storchaka stage: -> needs patch versions: -Python 3.3, Python 3.4, Python 3.5

[issue31324] support._match_test() used by test.bisect is very inefficient

STINNER Victor added the comment: I tested manually Serhiy's PR 4420: it works as expected! haypo@selma$ ./python -m test test_os --list-cases -m test.test_os.FileTests.test_access test.test_os.FileTests.test_access haypo@selma$ ./python -m test test_os --list-cases

[issue9134] sre bug: lastmark_save/restore

Change by Serhiy Storchaka : -- assignee: -> serhiy.storchaka versions: +Python 3.6, Python 3.7 -Python 3.4, Python 3.5 ___ Python tracker

[issue32048] Misprint in the unittest.mock documentation.

Change by Артемий Родионов : -- resolution: -> not a bug stage: -> resolved status: open -> closed ___ Python tracker ___

[issue30133] Strings that end with properly escaped backslashes cause error to be thrown in re.search/sub/etc. functions.

Serhiy Storchaka added the comment: The proper way of escaping the replacement string has been documented by issue31714. -- resolution: -> out of date stage: -> resolved status: open -> closed superseder: -> Improve re documentation

[issue31324] support._match_test() used by test.bisect is very inefficient

Serhiy Storchaka added the comment: $ time ./python -m test.bisect --fail-env-changed -o bisect test_asyncio -N1 ... Bisection failed after 2 iterations and 0:00:11 real0m10,441s user0m8,726s sys 0m0,292s There is good opportunity for optimizing the

[issue31956] Add start and stop parameters to the array.index()

Change by Roundup Robot : -- keywords: +patch pull_requests: +4379 stage: -> patch review ___ Python tracker ___

[issue32055] Reconsider comparison chaining for containment tests

New submission from Nick Coghlan : (I thought there was an open low priority issue for this, but I can't find it, so filing a new one) Currently, "in" and "not in" are classified as comparison operations in the language grammar, even though they're not actually documented

[issue28617] Why isn't "in" called a comparison operation?

Nick Coghlan added the comment: Also see https://bugs.python.org/issue32055 regarding the prospect of bringing the implementation into line with the intuitive semantics, and preventing implicit comparison chaining for containment tests. -- nosy: +ncoghlan

[issue32055] Reconsider comparison chaining for containment tests

Nick Coghlan added the comment: https://bugs.python.org/issue14247 is another older issue related to this point (again related to a user finding the current behaviour genuinely unintuitive) -- ___ Python tracker

[issue32045] Does json.dumps have a memory leak?

Rohan D'Sa added the comment: you are right. i realized later i actually had a leak in a com instantiated object, assumed it was a leak in the python and tried to find it using the gc module. The gc documentation led me down the garden path. QUOTE gc.garbage A list of

[issue32046] 2to3 fix for operator.isCallable()

Éric Araujo added the comment: Why about just removing the check, now that callable has been back in 3.x for a few releases? -- nosy: +eric.araujo ___ Python tracker

[issue31701] faulthandler dumps 'Windows fatal exception: code 0xe06d7363'

Ilya Kulakov added the comment: I think faulthandler should use both. E.g. in [1] you can read about an exception that can be handled by AddVectoredExceptionHandler but not SetUnhandledExceptionFilter. Perhaps implementation should use SetUnhandledExceptionFilter for

[issue32052] Provide access to buffer of asyncio.StreamReader

New submission from Bruce Merry : While asyncio.StreamReader.readuntil is an improvement on only having readline, it is still quite limited e.g. you cannot have multiple possible terminators. The real problem is that it's not possible to roll your own without accessing

[issue32025] Add time.thread_time()

Antoine Pitrou added the comment: Should I leave this open for you to work on the macOS implementation, Victor? -- ___ Python tracker

[issue32038] Add API to intercept socket.close()

Ben Darnell added the comment: Note that a guard on socket objects can only solve part of the problem: in the case where i've seen this bug, it was with raw file descriptors from libcurl, and there's nothing python can do about that because there are no (python) socket

[issue32038] Add API to intercept socket.close()

Yury Selivanov added the comment: > Note that a guard on socket objects can only solve part of the problem: in > the case where i've seen this bug, it was with raw file descriptors from > libcurl, and there's nothing python can do about that because there are no >

[issue32040] Sorting pahtlib.Paths does give the same order as sorting the (string) filenames of that pathlib.Paths

R. David Murray added the comment: To put it another way, think about sorting a list of tuples. Same behavior. -- ___ Python tracker

[issue32040] Sorting pahtlib.Paths does give the same order as sorting the (string) filenames of that pathlib.Paths

R. David Murray added the comment: It is "obvious by inspection". Paths are paths instead of strings because they are formed out of discrete path components instead of strings. If you sorted each directory in the paths from the top down, and then sorted the

[issue31701] faulthandler dumps 'Windows fatal exception: code 0xe06d7363'

Ilya Kulakov added the comment: Another option is to use AddVectoredContinueHandler [1]. It seems to be called if both VEH and SEH failed to handle the error. 1: https://msdn.microsoft.com/en-us/library/windows/desktop/ms679273(v=vs.85).aspx --

[issue31911] Use malloc_usable_size() in pymalloc for realloc

Antoine Pitrou added the comment: What would be the point of this? This would complicate the code with platform-specific code and it's not obvious that it would change runtime performance in any significant way. -- nosy: +pitrou

[issue31691] Include missing info on required build steps and how to build installer

Change by Roundup Robot : -- pull_requests: +4372 stage: -> patch review ___ Python tracker ___

[issue32051] Possible issue in multiprocessing doc

Berker Peksag added the comment: The example itself works fine, but I agree that it would be better to use a better name. We could replace ``('foo.bar.org', 5)`` in the following examples with ``('', 5)`` too. Would you like to send a PR? The documentation

[issue32051] Possible issue in multiprocessing doc

jason added the comment: Thank you for answering! Yes, I would be very happy to do that, this is going to be my first contribution! :) -- ___ Python tracker

[issue29753] Ctypes Packing Bitfields Incorrectly - Linux

Charles Machalow added the comment: Antti, is there a place in the ctypes documentation that explicitly says ctypes is not meant to be used cross-platform? If not, shouldn't that be mentioned? I think ultimately ctypes should default to standard OS/compiler behavior, but

[issue31701] faulthandler dumps 'Windows fatal exception: code 0xe06d7363'

Change by STINNER Victor : -- pull_requests: +4365 ___ Python tracker ___ ___

[issue32046] 2to3 fix for operator.isCallable()

Change by Dong-hee Na : -- keywords: +patch pull_requests: +4366 stage: -> patch review ___ Python tracker ___

[issue32043] Add a new -X dev option: "developer mode"

STINNER Victor added the comment: New changeset ccb0442a338066bf40fe417455e5a374e5238afb by Victor Stinner in branch 'master': bpo-32043: New "developer mode": "-X dev" option (#4413) https://github.com/python/cpython/commit/ccb0442a338066bf40fe417455e5a374e5238afb

[issue31702] Allow to specify the number of rounds for SHA-* hashing in crypt

Change by Serhiy Storchaka : -- resolution: -> fixed stage: patch review -> resolved status: open -> closed ___ Python tracker

[issue31702] Allow to specify the number of rounds for SHA-* hashing in crypt

Serhiy Storchaka added the comment: New changeset cede8c9edb408321b493d8d5e73be9e1018020e4 by Serhiy Storchaka in branch 'master': bpo-31702: Allow to specify rounds for SHA-2 hashing in crypt.mksalt(). (#4110)

[issue26752] Mock(2.0.0).assert_has_calls() raise AssertionError in two same calls

Guillaume Boily added the comment: As pointed here: https://github.com/testing-cabal/mock/issues/353, this issue is related to the method assert_has_calls or probably any calls that use the method _call_matcher to match calls. Given that you mock a class and spec it, since

[issue32046] 2to3 fix for operator.isCallable()

Change by Éric Araujo : -- Removed message: https://bugs.python.org/msg306393 ___ Python tracker ___

[issue32053] Inconsistent use of tabs and spaces in indentation not always detected

New submission from Serhiy Storchaka : >>> exec('''if 1: ... print(1) ...\tprint(2) ... ''') 1 2 The first indented line uses 8 spaces. The second indented line uses 7 spaces + tabulation. Indentations are different, but TabError is not raised.

[issue29753] Ctypes Packing Bitfields Incorrectly - Linux

Change by Marc Le Roy : -- nosy: +mleroy003 ___ Python tracker ___ ___ Python-bugs-list

[issue32054] Creating RPM on Python 2 works, but Python 3 fails because of sys.implementation.cache_tag

New submission from Pedro : I am trying to create an RPM on SLES12 SP2. (I think that corresponds to OpenSUSE 42.2.) This is my setup.py file, nothing special: import setuptools setuptools.setup(name='MyApp', version='1.2.3',

[issue32045] Does json.dumps have a memory leak?

Serhiy Storchaka added the comment: indent=True just makes json to use Python implementation instead of C implementation. Python implementation uses closures which reference one other. Simple example not involving json is: import gc def f(): def g():

[issue32027] argparse allow_abbrev option also controls short flag combinations

Berker Peksag added the comment: Thank you for the report and for the PR. I think this is a duplicate of issue 26967. > This issue is either a doc issue only or an unintended combination of > long option shortening and (the way more common) flag combinations. This is

[issue32045] Does json.dumps have a memory leak?

New submission from Rohan D'Sa : import gc, json class leak(object): def __init__(self): pass gc.set_debug(gc.DEBUG_LEAK) while True: leak_ = leak() json.dumps(leak_.__dict__, indent=True) gc.collect() print(f"garbage count: {len(gc.garbage)}")

[issue31808] tarfile.extractall fails to overwrite symlinks

Change by Dong-hee Na : -- keywords: +patch pull_requests: +4364 stage: needs patch -> patch review ___ Python tracker ___

[issue31808] tarfile.extractall fails to overwrite symlinks

Change by Dong-hee Na : -- nosy: +corona10 ___ Python tracker ___ ___

[issue31691] Include missing info on required build steps and how to build installer

Change by Roundup Robot : -- pull_requests: +4373 ___ Python tracker ___

[issue31691] Include missing info on required build steps and how to build installer

Steve Dower added the comment: New changeset c69611990365ae2f2efffd191de0aba9ee98d95c by Steve Dower (native-api) in branch '3.6': bpo-31691: Specify where to find installer build instructions for Windows (#4427)

[issue31691] Include missing info on required build steps and how to build installer

Change by Roundup Robot : -- pull_requests: +4374 ___ Python tracker ___

[issue31691] Include missing info on required build steps and how to build installer

Change by Roundup Robot : -- pull_requests: +4377 ___ Python tracker ___

[issue31691] Include missing info on required build steps and how to build installer

Change by Roundup Robot : -- pull_requests: +4375 ___ Python tracker ___

[issue31691] Include missing info on required build steps and how to build installer

Steve Dower added the comment: New changeset fd0fa67464f75cebca67cd14d050476a0c73bd53 by Steve Dower (native-api) in branch 'master': bpo-31691: Specify where to find build instructions for the Windows installer (#4426)

[issue31691] Include missing info on required build steps and how to build installer

Change by Roundup Robot : -- pull_requests: +4376 ___ Python tracker ___

[issue31691] Include missing info on required build steps and how to build installer

Steve Dower added the comment: New changeset c02307e9634583d3f2660e46b5200bf0b282695a by Steve Dower (native-api) in branch '2.7': bpo-31691: Specify how to build installer in Windows (#4430)

[issue31691] Include missing info on required build steps and how to build installer

Steve Dower added the comment: Thanks for the PR! -- resolution: -> fixed stage: patch review -> resolved status: open -> closed type: -> crash ___ Python tracker

[issue31691] Include missing info on required build steps and how to build installer

Steve Dower added the comment: New changeset 0a6158ca392adfdfbeb7eb43d8d8109c02b00d23 by Steve Dower (Miss Islington (bot)) in branch '3.6': bpo-31691: Specify where to find build instructions for the Windows installer (GH-4426) (#4431)

  1   2   >