Commit Graph

35217 Commits

Author SHA1 Message Date
Victor Stinner 14d45c06eb Issue #23009: Skip test_selectors.test_empty_select() on Windows 2015-01-22 09:07:36 +01:00
Victor Stinner 1ca9392c70 Issue #23095, asyncio: IocpProactor.close() must not cancel pending
_WaitCancelFuture futures
2015-01-22 00:17:54 +01:00
Victor Stinner 0c2e408833 asyncio: BaseEventLoop._create_connection_transport() catchs any exception, not
only Exception
2015-01-22 00:17:41 +01:00
Victor Stinner d0a28dee78 Issue #23095, asyncio: Rewrite _WaitHandleFuture.cancel()
This change fixes a race conditon related to _WaitHandleFuture.cancel() leading
to Python crash or "GetQueuedCompletionStatus() returned an unexpected event"
logs. Before, the overlapped object was destroyed too early, it was possible
that the wait completed whereas the overlapped object was already destroyed.
Sometimes, a different overlapped was allocated at the same address, leading to
unexpected completition.

_WaitHandleFuture.cancel() now waits until the wait is cancelled to clear its
reference to the overlapped object. To wait until the cancellation is done,
UnregisterWaitEx() is used with an event instead of UnregisterWait().

To wait for this event, a new _WaitCancelFuture class was added. It's a
simplified version of _WaitCancelFuture. For example, its cancel() method calls
UnregisterWait(), not UnregisterWaitEx(). _WaitCancelFuture should not be
cancelled.

The overlapped object is kept alive in _WaitHandleFuture until the wait is
unregistered.

Other changes:

* Add _overlapped.UnregisterWaitEx()
* Remove fast-path in IocpProactor.wait_for_handle() to immediatly set the
  result if the wait already completed. I'm not sure that it's safe to
  call immediatly UnregisterWaitEx() before the completion was signaled.
* Add IocpProactor._unregistered() to forget an overlapped which may never be
  signaled, but may be signaled for the next loop iteration. It avoids to
  block forever IocpProactor.close() if a wait was cancelled, and it may also
  avoid some "... unexpected event ..." warnings.
2015-01-21 23:39:51 +01:00
Victor Stinner 442b0adccd asyncio: pyflakes, remove unused import
tests: Remove unused function; inline another function
2015-01-21 23:39:16 +01:00
Victor Stinner cd0f7f9832 asyncio: Enhance BaseProactorEventLoop._loop_self_reading()
* Handle correctly CancelledError: just exit
* On error, log the exception and exit

Don't try to close the event loop, it is probably running and so it cannot be
closed.
2015-01-21 23:38:37 +01:00
Gregory P. Smith b5684c48e1 Add the command line to the AssertionError raised by test.script_helper's
Python subprocess failure assertion error messages for easier debugging.

Adds a unittest for test.script_helper to confirm that this code works as
it is otherwise uncovered by an already passing test suite that uses it. :)
2015-01-20 17:19:47 -08:00
Zachary Ware b176d40398 Issue #23280: Fix docstrings for binascii.(un)hexlify 2015-01-20 13:59:46 -06:00
Serhiy Storchaka d3faf43f9b Issue #23181: More "codepoint" -> "code point". 2015-01-18 11:28:37 +02:00
Ned Deily ccb416fee2 Issue #23180: Rename IDLE "Windows" menu item to "Window".
Patch by Al Sweigart.
2015-01-17 21:06:27 -08:00
Ned Deily 3bbb37e09b Issue #23211: Fix patch for 3.4 differences. 2015-01-17 17:31:13 -08:00
Ned Deily 1e012e6c7f Issue #23211: Workaround test_logging failure on some OS X 10.6 systems:
getaddrinfo("localhost") can fail depending on the name server configuration,
use "127.0.0.0" instead.
2015-01-17 16:57:19 -08:00
Zachary Ware 465b057f91 Closes #23256: Avoid a crash in test_ctypes
Only happened with oddly capitalized debug executables on Windows.
Patch by Claudiu Popa.
2015-01-17 08:50:42 -06:00
Benjamin Peterson bd341629b0 capitialize "HttpOnly" and "Secure" as they appear in the standard and other impls (closes #23250)
Patch by Jon Dufresne.
2015-01-16 20:43:55 -05:00
Victor Stinner f716d8b7a5 Issue #22685: Fix test_pause_reading() of asyncio/test_subprocess
Override the connect_read_pipe() method of the loop to mock immediatly
pause_reading() and resume_reading() methods.

The test failed randomly on FreeBSD 9 buildbot and on Windows using trollius.
2015-01-15 22:52:59 +01:00
Victor Stinner 406204c8c2 Backout changeset 6ab2575bc12b
StreamWriter: close() now clears the reference to the transport

StreamWriter now raises an exception if it is closed: write(), writelines(),
write_eof(), can_write_eof(), get_extra_info(), drain().
2015-01-15 21:50:19 +01:00
Victor Stinner 922bc2ca12 Closes #23219: cancelling asyncio.wait_for() now cancels the task 2015-01-15 16:29:10 +01:00
Victor Stinner ab8848bc2a asyncio: Close transports in tests
* Use test_utils.run_briefly() to execute pending calls to really close
  transports
* sslproto: mock also _SSLPipe.shutdown(), it's need to close the transport
* pipe test: the test doesn't close explicitly the PipeHandle, so ignore
  the warning instead
* test_popen: use the context manager ("with p:") to explicitly close pipes
2015-01-15 14:24:55 +01:00
Victor Stinner 4bf22e033e asyncio: Close the transport on subprocess creation failure 2015-01-15 14:24:22 +01:00
Victor Stinner fcd58de78f asyncio: Fix _ProactorBasePipeTransport.close()
Set the _read_fut attribute to None after cancelling it.

This change should fix a race condition with
_ProactorWritePipeTransport._pipe_closed().
2015-01-15 13:40:27 +01:00
Victor Stinner 79c93ba47b asyncio: Fix _ProactorBasePipeTransport.__repr__()
Check if the _sock attribute is None to check if the transport is closed.
2015-01-15 13:32:28 +01:00
Victor Stinner 02392c9282 Issue #23243: Close explicitly transports in asyncio tests 2015-01-15 13:18:32 +01:00
Victor Stinner 33cb0396e8 Issue #23243: Close explicitly event loops in asyncio tests 2015-01-15 13:17:34 +01:00
Victor Stinner 41ed958ee6 Issue #23243: Fix asyncio._UnixWritePipeTransport.close()
Do nothing if the transport is already closed. Before it was not possible to
close the transport twice.
2015-01-15 13:16:50 +01:00
Victor Stinner 7e222f411c SSLProtocol: set the _transport attribute in the constructor 2015-01-15 13:16:27 +01:00
Victor Stinner 791009bb89 Issue #23242: asyncio.SubprocessStreamProtocol now closes the subprocess
transport at subprocess exit.

Clear also its reference to the transport.
2015-01-15 13:16:02 +01:00
Victor Stinner 72bdefb086 Issue #22560: Fix typo: call -> call_soon 2015-01-15 09:44:13 +01:00
Victor Stinner 042dad7232 Issue #22560: Fix SSLProtocol._on_handshake_complete()
Don't call immediatly self._process_write_backlog() but schedule the call using
call_soon(). _on_handshake_complete() can be called indirectly from
_process_write_backlog(), and _process_write_backlog() is not reentrant.
2015-01-15 09:41:48 +01:00
Victor Stinner e7a35717d2 StreamWriter: close() now clears the reference to the transport
StreamWriter now raises an exception if it is closed: write(), writelines(),
write_eof(), can_write_eof(), get_extra_info(), drain().
2015-01-15 09:33:50 +01:00
Benjamin Peterson 562b7cbff9 fix parsing reST with code or code-block directives (closes #23063)
Patch by Marc Abramowitz.
2015-01-14 23:56:35 -05:00
Victor Stinner 29ad0111bd asyncio: sync with Tulip
* PipeHandle now uses None instead of -1 for a closed handle
* Sort imports in windows_utils.
* Fix test_events on Python older than 3.5. Skip SSL tests on the
  ProactorEventLoop if ssl.MemoryIO is missing
* Fix BaseEventLoop._create_connection_transport(). Close the transport if the
  creation of the transport (if the waiter) gets an exception.
* _ProactorBasePipeTransport now sets _sock to None when the transport is
  closed.
* Fix BaseSubprocessTransport.close(). Ignore pipes for which the protocol is
  not set yet (still equal to None).
* TestLoop.close() now calls the close() method of the parent class
  (BaseEventLoop).
* Cleanup BaseSelectorEventLoop: create the protocol on a separated line for
  readability and ease debugging.
* Fix BaseSubprocessTransport._kill_wait(). Set the _returncode attribute, so
  close() doesn't try to terminate the process.
* Tests: explicitly close event loops and transports
* UNIX pipe transports: add closed/closing in repr(). Add "closed" or "closing"
  state in the __repr__() method of _UnixReadPipeTransport and
  _UnixWritePipeTransport classes.
2015-01-15 00:04:21 +01:00
Victor Stinner b92626df5c Issue #23197: On SSL handshake failure on matching hostname, check if the
waiter is cancelled before setting its exception.
2015-01-14 17:13:28 +01:00
Victor Stinner 177e9f0855 Issue #23197, asyncio: On SSL handshake failure, check if the waiter is
cancelled before setting its exception.

* Add unit tests for this case.
* Cleanup also sslproto.py
2015-01-14 16:56:20 +01:00
Victor Stinner f651a60407 Python issue #23173: sync with Tulip
* If an exception is raised during the creation of a subprocess, kill the
  subprocess (close pipes, kill and read the return status). Log an error in
  such case.
* Fix SubprocessStreamProtocol.connection_made() to handle cancelled waiter.
  Add unit test cancelling subprocess methods.
2015-01-14 02:10:33 +01:00
Victor Stinner c2c12e433a Issue #23198: Reactor asyncio.StreamReader
- Add a new _wakeup_waiter() method
- Replace _create_waiter() method with a _wait_for_data() coroutine function
- Use the value None instead of True or False to wake up the waiter
2015-01-14 00:53:37 +01:00
Victor Stinner 231b404cb0 Issue #22560: New SSL implementation based on ssl.MemoryBIO
The new SSL implementation is based on the new ssl.MemoryBIO which is only
available on Python 3.5. On Python 3.4 and older, the legacy SSL implementation
(using SSL_write, SSL_read, etc.) is used. The proactor event loop only
supports the new implementation.

The new asyncio.sslproto module adds _SSLPipe, SSLProtocol and
_SSLProtocolTransport classes. _SSLPipe allows to "wrap" or "unwrap" a socket
(switch between cleartext and SSL/TLS).

Patch written by Antoine Pitrou. sslproto.py is based on gruvi/ssl.py of the
gruvi project written by Geert Jansen.

This change adds SSL support to ProactorEventLoop on Python 3.5 and newer!

It becomes also possible to implement STARTTTLS: switch a cleartext socket to
SSL.
2015-01-14 00:19:09 +01:00
Victor Stinner 9036e49ba1 Tulip issue 184: Fix test_pipe() on Windows
Pass explicitly the event loop to StreamReaderProtocol.
2015-01-13 16:13:06 +01:00
Victor Stinner b86a96802b Issue #22922: Fix ProactorEventLoop.close()
Close the IocpProactor before closing the event loop. IocpProactor.close() can
call loop.call_soon(), which is forbidden when the event loop is closed.
2015-01-13 16:11:19 +01:00
Benjamin Peterson 82f34ada45 fix instances of consecutive articles (closes #23221)
Patch by Karan Goel.
2015-01-13 09:17:24 -05:00
Victor Stinner 38dc250521 Issue #23209, #23225: selectors.BaseSelector.close() now clears its internal
reference to the selector mapping to break a reference cycle. Initial patch
written by Martin Richard.
2015-01-13 09:58:33 +01:00
Donald Stufft d9fe22ce9b Update setuptools to 11.3.1 2015-01-11 15:51:11 -05:00
Victor Stinner 686fe6e156 Issue #23209: Revert change on selectors, test_selectors failed. 2015-01-09 21:56:28 +01:00
Victor Stinner 587feb19e8 Issue #23209: Break some reference cycles in asyncio. Patch written by Martin
Richard.
2015-01-09 21:34:27 +01:00
Victor Stinner 70db9e428a asyncio: sync with Tulip
* Tulip issue 184: FlowControlMixin constructor now get the event loop if the
  loop parameter is not set. Add unit tests to ensure that constructor of
  StreamReader and StreamReaderProtocol classes get the event loop.
* Remove outdated TODO/XXX
2015-01-09 21:32:05 +01:00
Victor Stinner 3531d9044d asyncio: sync with Tulip
* Document why set_result() calls are safe
* Cleanup gather(). Use public methods instead of hacks to consume the
  exception of a future.
* sock_connect(): pass directly the fd to _sock_connect_done instead of the
  socket.
2015-01-09 01:42:52 +01:00
Victor Stinner 15cc678d89 asyncio: Truncate to 80 columns 2015-01-09 00:09:10 +01:00
Victor Stinner 8d9c145f61 asyncio: _make_ssl_transport: make the waiter parameter optional 2015-01-08 12:06:36 +01:00
Terry Jan Reedy 84d64a9c02 Issue #23184: delete unused idlelib file. 2015-01-07 23:48:46 -05:00
Nick Coghlan b9fdb7a452 Issue 19548: update codecs module documentation
- clarified the distinction between text encodings and other codecs
- clarified relationship with builtin open and the io module
- consolidated documentation of error handlers into one section
- clarified type constraints of some behaviours
- added tests for some of the new statements in the docs
2015-01-07 00:22:00 +10:00
Victor Stinner fcfed19913 Issue #21356: Make ssl.RAND_egd() optional to support LibreSSL. The
availability of the function is checked during the compilation. Patch written
by Bernard Spil.
2015-01-06 13:54:58 +01:00