summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* Future.set_exception(exc) now instanciate exc if it is a classVictor Stinner2014-01-302-0/+7
| | | | For example, Future.set_exception(RuntimeError) is now allowed.
* Fix granularity of test_utils.TestLoopVictor Stinner2014-01-301-0/+1
|
* Fix _UnixWritePipeTransport: raise BrokenPipeError when the pipe is closedVictor Stinner2014-01-301-1/+4
|
* subprocess_shell() and subprocess_exec() methods of BaseEventLoop now raises aVictor Stinner2014-01-292-5/+46
| | | | | ValueError instead of raising an AssertionError. Moreover, bufsize different than 0 is now considered as an error.
* Fix _make_subprocess_transport(): pass extra value to the constructorVictor Stinner2014-01-292-2/+2
|
* Move async() call back to its original position. Issue 117.Guido van Rossum2014-01-281-1/+1
|
* wait_for() now accepts None as timeoutVictor Stinner2014-01-282-1/+15
|
* Fix ResourceWarning in tcp and udp echo examplesVictor Stinner2014-01-282-5/+9
|
* examples: close the event loopVictor Stinner2014-01-285-10/+25
|
* pass through pause/resume from subprocess pipe proto to subprocess proto.Guido van Rossum2014-01-281-2/+5
|
* Refactor drain logic in streams.py to be reusable.Guido van Rossum2014-01-281-36/+61
|
* Refactoring: get rid of _try_connected().Victor Stinner2014-01-281-18/+14
|
* Add write flow control to unix pipes.Guido van Rossum2014-01-281-3/+11
|
* Refactoring: move write flow control to a subclass/mixin.Guido van Rossum2014-01-271-37/+61
|
* _fatal_error() of _UnixWritePipeTransport and _ProactorBasePipeTransport don'tVictor Stinner2014-01-282-2/+4
| | | | | | log BrokenPipeError nor ConnectionResetError Same behaviour than _SelectorTransport._fatal_error()
* Close loop in crawl.py example (mostly for IOCP).Guido van Rossum2014-01-281-0/+1
|
* Remove temporary aliases Full/Empty.Guido van Rossum2014-01-271-6/+0
|
* The standard readexactly() now raises on a short read, so kill the custom ↵Guido van Rossum2014-01-261-22/+2
| | | | wrapper.
* Code cleanup: remove unused functionAndrew Svetlov2014-01-263-7/+0
|
* Rename {Empty,Full} to {QueueEmpty,QueueFull} and no longer get them from ↵Guido van Rossum2014-01-252-11/+21
| | | | queue.py.
* Fix whitespace.Guido van Rossum2014-01-251-1/+1
|
* Locks refactor: use a separate context manager; remove Semaphore._locked.Guido van Rossum2014-01-252-22/+95
|
* Fix race in FastChildWatcher (by its original author, Anthony Baire).Guido van Rossum2014-01-251-20/+16
|
* Make the new granularity attribute privateVictor Stinner2014-01-263-5/+5
|
* Simplify BaseEventLoop._run_once(): avoid math.ceil(), use simple arithmeticVictor Stinner2014-01-251-6/+1
| | | | instead
* Don't export BaseEventLoop, BaseSelectorEventLoop nor BaseProactorEventLoopVictor Stinner2014-01-255-15/+13
| | | | Import them from submodules if you really need them.
* Strip trailing spaceVictor Stinner2014-01-251-1/+1
|
* Export more symbols: BaseEventLoop, BaseProactorEventLoop,Victor Stinner2014-01-258-53/+60
| | | | BaseSelectorEventLoop, Queue and Queue sublasses, Empty, Full
* Add a granularity attribute to BaseEventLoop: maximum between the resolution ofVictor Stinner2014-01-256-1/+41
| | | | | the BaseEventLoop.time() method and the resolution of the selector. The granuarility is used in the scheduler to round time and deadline.
* selectors: add a resolution attribute to BaseSelectorVictor Stinner2014-01-251-1/+22
|
* Revert changes in selectors for PollSelector and EpollSelector: round again theVictor Stinner2014-01-251-8/+2
| | | | timeout towards zero, instead of rounding away from zero
* AUTHORS: remove Authors/Contributors titles, only mention that Guido is theVictor Stinner2014-01-241-31/+26
| | | | author of the project
* Issue #111: StreamReader.readexactly() now raises an IncompleteReadError if theVictor Stinner2014-01-242-8/+22
| | | | | end of stream is reached before we received enough bytes, instead of returning less bytes than requested.
* Only skip PTY tests with the kqueue selectorVictor Stinner2014-01-241-6/+12
| | | | PTY tests pass with select and poll selectors on Mac OS 10.9.
* _UnixWritePipeTransport now also supports character devices, asVictor Stinner2014-01-243-4/+56
| | | | _UnixReadPipeTransport. Patch written by Jonathan Slenders.
* Unit tests: pick symbols from the asyncio moduleVictor Stinner2014-01-241-5/+4
|
* Unit tests use the main asyncio module instead of submodules like eventsVictor Stinner2014-01-2411-812/+778
|
* Skip test_events.test_read_pty_output() on Mac OS X older than 10.9 (Maverick)Victor Stinner2014-01-231-0/+3
|
* Issue #110: StreamReader.read() and StreamReader.readline() now raise aVictor Stinner2014-01-231-6/+13
| | | | | RuntimeError, instead of using an assertion, if another coroutine is already waiting for incoming data
* Fix open_connection() docstring, writer is a StreamWriterVictor Stinner2014-01-231-1/+1
|
* wait_for() now cancels the future on timeout. Patch written by GustavoVictor Stinner2014-01-223-17/+19
| | | | Carneiro.
* Create a list of authors and contributors based on the Mercurial historyVictor Stinner2014-01-221-0/+31
|
* selectors: don't mention the Python bugVictor Stinner2014-01-221-3/+1
| | | | Copy selectors.py from CPython to have exactly the same file in Tulip.
* Cleanup logging in BaseEventLoop._run_once()Victor Stinner2014-01-221-2/+5
| | | | | | logger.log() is now responsible to format the timeout. It might be faster if the log is disabled for DEBUG level, but it's also more readable and fix an issue with Python 2.6 in the Trollius project.
* Restore asyncio/__init__.py to have the same file in Python 3.4 and TulipVictor Stinner2014-01-211-3/+7
|
* Fix timeout rounding issues in selectorsVictor Stinner2014-01-212-9/+22
| | | | | | | | Round timeouts away from zero to wait *at least* 'timeout' seconds in PollSelector and EpollSelector. The change in EpollSelector works around a Python bug in select.epoll.poll(): http://bugs.python.org/issue20311
* Optimize BaseEventLoop._run_once()Victor Stinner2014-01-211-8/+11
| | | | | | | | Logger.log() is "slow", logger.isEnabledFor() is faster and the logger is disabled in most cases. A microbenchmark executing 100,000 dummy tasks is 22% faster with this change. See the CPython issue: http://bugs.python.org/issue20275
* Reincarnate CoroWrapper's docstring as a comment.Guido van Rossum2014-01-161-0/+2
|
* Fix CoroWrapper (fix my previous commit)Victor Stinner2014-01-161-3/+1
| | | | Add __name__ and __doc__ to __slots__
* Issue #104: Fix a typo in CoroWrapperVictor Stinner2014-01-161-1/+1
| | | | __slot__ => __slots__