e118963b30
issue #254 : fork: take care not to rely on FD numbers.
6 years ago
fc59f57ba2
issue #213 : core: split out import_module() for use in services.py.
6 years ago
49fb25ee1c
issue #213 : core: fix shutdown crash due to member variable rename
6 years ago
40c6c6426f
issue #213 : core: fix test breakage due to log message change
6 years ago
2310497d55
issue #213 : core: have Message.reply() log msg for zero reply_to
...
It's easy to call msg.reply() by accident on a message that never had
reply_to set, resulting in a "invalid handle" error message coming from
router. Instead log a more accurate message on the stack that actualy
caused the problem.
6 years ago
633585524f
tests: don't need separate module for id_allocation_test any more
...
This used to be because everything would explode while importing
__main__ under py.test, but that was fixed months ago.
6 years ago
dd48c41332
Ignore another annoying flake8 message.
6 years ago
d2714752ee
docs: tidy ups
6 years ago
61365236ad
docs/select: fix up more references, fix headings.
6 years ago
4bf3d01104
select: add missing get(block=..) parameter.
6 years ago
ddf28987a0
master: split Select() into new module to reduce wire size.
...
service.py currently imports master.py(+parent.py) just to get Select().
6 years ago
7a592d1c34
core: better Poller.__repr__
6 years ago
b0ce6eecd7
fork: support on_start= argument.
6 years ago
00edf0d66d
core: have ExternalContext accept a config dict rather than kwargs.
...
The parameter lists had gotten out of control.
6 years ago
eac4cc7afe
tests: speed things up
6 years ago
863a95e860
docs: update contributors
6 years ago
b0aa413173
tests: import benchmark reproduction setup playbook
...
Incomplete, also needs to mess around with Gcloud routing and handle box
setup/teardown, because that's another "well engineered" waste of time.
6 years ago
a99c7a12f9
tests: output split groups in gcloud.py
6 years ago
1a0e630a80
tests: add debops to requirements
6 years ago
dbcee4041a
tests: refactor gcloud.py to be dynamic inventory.
6 years ago
a3995f8e5f
ansible: remove hard-coded dw username.
6 years ago
1eb5c20f57
ansible: add dummy init.pys so setup.py find_packages() DTRT.
6 years ago
2c58591129
Bump Ansible version in dev_requirements.txt
6 years ago
7d0209d8de
issue #249 : have upgrade_router() upgrade the poller too.
...
Now when a child becomes a parent, it gets a new poller suitable for
many more children than was possible using select().
6 years ago
55fff54774
core: make try/catch logic a little clearer in Latch.get()
6 years ago
05a5f2b6e5
core: if Poller.poll() fails, TimeoutError would be raised.
...
We must check whether poller threw an exception both in the case that we
weren't woken and the case that we were.
6 years ago
5bdc1719c5
issue #249 : epoll() raises IOError for EINTR, not select.error.
6 years ago
70376d861a
issue #217 : docs: remove limitation
6 years ago
6d18ce81d8
issue #249 : restore duplex behaviour for epoll
...
With epoll() there is only one kernel-side object per file descriptor,
which is why _control() is such a pain. Since we merge receive/transmit
watching into that single object, we must always test the mask for both
conditions when reading results.
Kqueue isn't/doesn't appear to be like this. The identity of a Kqueue
event is keyed on (fd, filter), and we register a separate event for
both transmit and receive, so the 'elif' in KqueuePoller.poll() does not
appear to need to change.
Previously, a FD marked for read+write would not indicate writeability
until it was no longer readable.
6 years ago
6b98818046
issue #249 : epoll distinguishes between hangup and disconnect
...
..typical Linux, for certain descriptor types only. So our receive mask
must match both, and normalize it into a read event like every other
poller.
6 years ago
07056b0dd1
issue #249 : fix ordering bug masked by previous implementation
6 years ago
4cd9e09130
issue #249 : docs: remove limitation
6 years ago
36a1024861
issue #249 : port Latch to poller too.
...
This is probably going to suck for perf :/
6 years ago
1070dfae72
issue #249 : fix iter_read() regression.
6 years ago
dcf0aa351e
issue #249 : whoops, fix new poller timeouts.
6 years ago
aa8f786413
issue #249 : fix Poller.from_existing() for v2 API
6 years ago
5645629e5d
issue #249 : the new pollers must handle syscall restarts too.
6 years ago
4df020827d
issue #249 : explicitly close pollers when done.
6 years ago
9905f6d8b4
issue #249 : make write_all() and iter_read() use PREFERRED_POLLER.
6 years ago
b6124f8396
issue #249 : EpollPoller v2.
6 years ago
9abcf63155
issue #249 : Poller API v2 (BSD only).
...
Now it's BasicStream/Side-agnostic, so it can be reused for Latch and
iter_read().
6 years ago
11c2e4ab3e
core: set _v and _vv to True in enable_debug_logging().
...
router.enable_debug() has been broken for ages.
6 years ago
874c0fdc01
Merge pull request #248 from badeadan/master
...
fix host_keys= exception message
6 years ago
7320c542df
issue #249 : EpollPoller() for Linux.
6 years ago
bc7be1879d
issue #249 : initial poller implementation (BSD only)
6 years ago
020c8b5013
fix host_keys= exception message
6 years ago
b7ab473343
issue #186 : split handle list up so it makes sense
6 years ago
d1a22cb5d4
issue #186 : parent: implement FORWARD_MODULE.
...
To support detach, we must be able to preload the target with every
module it will need prior to detachment. This implements the
intermediary part of the process (i.e. the Ansible fork parent) --
receiving LOAD_MODULE/FORWARD_MODULE pairs and ensuring they reach the
child.
6 years ago
8a089e975d
docs: Document Router.unidirectional.
6 years ago
bb61745a1a
issue #217 : pass through non-custom module utils to regular importer.
...
This may come back to bite later, but in the meantime it avoids shipping
up to 12KiB of junk metadata for every single task invocation.
For detachment (aka. async), we must ensure the target has two types of
preloads completed (modules and module_utils files) before detaching.
6 years ago
David Wilson