Currently, asyncio.wait_for(fut), upon reaching the timeout deadline,
cancels the future and returns immediately. This is problematic for
when *fut* is a Task, because it will be left running for an arbitrary
amount of time. This behavior is iself surprising and may lead to
related bugs such as the one described in bpo-33638:
condition = asyncio.Condition()
async with condition:
await asyncio.wait_for(condition.wait(), timeout=0.5)
Currently, instead of raising a TimeoutError, the above code will fail
with `RuntimeError: cannot wait on un-acquired lock`, because
`__aexit__` is reached _before_ `condition.wait()` finishes its
cancellation and re-acquires the condition lock.
To resolve this, make `wait_for` await for the task cancellation.
The tradeoff here is that the `timeout` promise may be broken if the
task decides to handle its cancellation in a slow way. This represents
a behavior change and should probably not be back-patched to 3.6 and
earlier.
(cherry picked from commit e2b340ab41)
Co-authored-by: Elvis Pranskevichus <elvis@magic.io>
Use transport.set_write_buffer_limits() in sendfile tests of
test_asyncio to make sure that the protocol is paused after sending
4 KiB. Previously,
test_sendfile_fallback_close_peer_in_the_middle_of_receiving() failed
on FreeBSD if the DATA was smaller than the default limit of 64 KiB.
(cherry picked from commit 9551f77192)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
Remove the docstring attribute of AST types and restore docstring
expression as a first stmt in their body.
Co-authored-by: INADA Naoki <methane@users.noreply.github.com>
Fix a race condition in SSLProtocol.connection_made() of
asyncio.sslproto: start immediately the handshake instead of using
call_soon(). Previously, data_received() could be called before the
handshake started, causing the handshake to hang or fail.
(cherry picked from commit be00a5583a)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
In this commit:
* Support BufferedProtocol in set_protocol() and start_tls()
* Fix proactor to cancel readers reliably
* Update tests to be compatible with OpenSSL 1.1.1
* Clarify BufferedProtocol docs
* Bump TLS tests timeouts to 60 seconds; eliminate possible race from start_serving
* Rewrite test_start_tls_server_1
(cherry picked from commit dbf102271f)
Co-authored-by: Yury Selivanov <yury@magic.io>
Fixed bug where calling write_eof() on a _SelectorSocketTransport after
it's already closed raises AttributeError.
(cherry picked from commit 23f587e395)
Co-authored-by: twisteroid ambassador <twisteroidambassador@users.noreply.github.com>
The failure may be due to the use oF ZFS, a case we already ignore
for Solaris-based systems where ZFS is frequently used.
(cherry picked from commit 09c4a7dee2)
Co-authored-by: Ned Deily <nad@python.org>
Pickles of type variables and subscripted generics are now future-proof
and compatible with older Python versions.
(cherry picked from commit 09f3221fbb)
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
bpo-32374, bpo-33629: Use support.SuppressCrashReport() in
test_bad_traverse() of MultiPhaseExtensionModuleTests to prevent
leaking a core dump file.
(cherry picked from commit 483000e164)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
bpo-26510 in 3.7.0a2 changed the behavior of argparse to make
subparsers required by default, returning to the behavior of 2.7
and 3.2. The behavior was changed in 3.3 to be no longer required.
While it might make more sense to have the default to required,
compatibility with 3.3 through 3.6 is probably less disruptive
than trying to reintroduce compatibility with 2.7 at this point.
This change restores the 3.6 behavior.
(cherry picked from commit 8ebf5ceb0f)
Co-authored-by: Ned Deily <nad@python.org>
Add a new block_on_close class attribute to ForkingMixIn and
ThreadingMixIn classes of socketserver to opt-in for pre-3.7 behaviour.
(cherry picked from commit 453bd0bc65)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
* bpo-33353: test_asyncio uses smaller sendfile data (#7083)
bpo-32622, bpo-33353: sendfile() tests of test_asyncio use socket
buffers of 1 kB "to test on relative small data sets". Send only
160 KiB rather 10 MB to make the test much faster.
Shrink also SendfileBase.DATA from 1600 KiB to 160 KiB.
On Linux, 3 test_sock_sendfile_mix_with_regular_send() runs now take
less than 1 second, instead of 18 seconds.
On FreeBSD, the 3 tests didn't hang, but took 3 minutes. Now
the 3 tests pass in less than 1 seconds.
(cherry picked from commit 2932755cc1)
* bpo-33353: test_asyncio set SO_SNDBUF after connect (GH-7086)
bpo-32622, bpo-33353: On macOS, sock.connect() changes the
SO_SNDBUF value. Only set SO_SNDBUF and SO_RCVBUF buffer sizes
once a socket is connected or binded, not before.
(cherry picked from commit b97de3dd86)
TLS 1.3 behaves slightly different than TLS 1.2. Session tickets and TLS
client cert auth are now handled after the initialy handshake. Tests now
either send/recv data to trigger session and client certs. Or tests
ignore ConnectionResetError / BrokenPipeError on the server side to
handle clients that force-close the socket fd.
To test TLS 1.3, OpenSSL 1.1.1-pre7-dev (git master + OpenSSL PR
https://github.com/openssl/openssl/pull/6340) is required.
Signed-off-by: Christian Heimes <christian@python.org>
(cherry picked from commit 529525fb5a)
bpo-33604: Bump removal notice from 3.6 to 3.8 and change PendingDeprecationWarning to DeprecationWarning as we had intended to do earlier...
(cherry picked from commit 8bb0b5b03c)
Co-authored-by: Matthias Bussonnier <bussonniermatthias@gmail.com>
Change TLS 1.3 cipher suite settings for compatibility with OpenSSL
1.1.1-pre6 and newer. OpenSSL 1.1.1 will have TLS 1.3 cipers enabled by
default.
Also update multissltests and Travis config to test with latest OpenSSL.
Signed-off-by: Christian Heimes <christian@python.org>
(cherry picked from commit e8eb6cb792)
Co-authored-by: Christian Heimes <christian@python.org>
* bpo-33263 Fix FD leak in _SelectorSocketTransport. (GH-6450)
Under particular circumstances _SelectorSocketTransport can try to add a reader
even the transport is already being closed. This can lead to FD leak and
invalid stated of the following connections. Fixed the SelectorSocketTransport
to add the reader only if the trasport is still active.
(cherry picked from commit a84d0b361a)
Co-authored-by: Vlad Starostin <drtyrsa@yandex.ru>
uuid._ipconfig_getnode did not validate the maximum length of the value,
so long as the value had the same type of formatting as a MAC address.
This let it select DUIDs as MAC addresses. It now requires an exact
length match.
(cherry picked from commit c66c342cb4)
Co-authored-by: CtrlZvi <viz+github@flippedperspective.com>
Fix the following bugs in the C implementation:
* get_future_loop() silenced all exceptions raised when look up the get_loop
attribute, not just an AttributeError.
* enter_task() silenced all exceptions raised when look up the current task,
not just a KeyError.
* repr() was called for a borrowed link in enter_task() and task_step_impl().
* str() was used instead of repr() in formatting one error message (in
Python implementation too).
* There where few reference leaks in error cases.
(cherry picked from commit 6655354afc)
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
The proactor event loop has a race condition when reading with
pausing/resuming. `resume_reading()` unconditionally schedules the read
function to read from the current future. If `resume_reading()` was
called before the previously scheduled done callback fires, this results
in two attempts to get the data from the most recent read and an
assertion failure. This commit tracks whether or not `resume_reading`
needs to reschedule the callback to restart the loop, preventing a
second attempt to read the data.
(cherry picked from commit 4151061855)
Co-authored-by: CtrlZvi <viz+github@flippedperspective.com>
Miss Islington (bot)
Serhiy Storchaka
Victor Stinner