jc211
/
tractor
forked from goodboy/tractor
1
0
Fork 0
Code Pull Requests Activity
1,629 Commits
211 Branches
0 Tags
9.3 MiB
Commit Graph

1629 Commits (c6b4da5788b4d7604d8bcd5921d5a516eaa40098)

Author SHA1 Message Date
overclockworked64 63ddf119fd
Drop 3.8 support 2021-10-23 18:18:36 +02:00
goodboy 828754dbb5
Merge pull request #245 from goodboy/immediate_remote_cancels 
Immediate remote cancels
 2021-10-17 08:16:50 -04:00
Tyler Goodlet b3c4851ffb Grab lock if cancelled during spawn before hard kill 2021-10-15 18:26:46 -04:00
Tyler Goodlet 5cfac58873 Don't pop a child entry that was never inserted 2021-10-15 18:16:58 -04:00
Tyler Goodlet 5d827f78e2 Fix pluggy readme link and typo 2021-10-15 11:42:57 -04:00
Tyler Goodlet 4f222a5f9c Use type match of expected error 2021-10-15 10:25:50 -04:00
Tyler Goodlet e4ed0fd2b3 Right, only worry about pdb lock when in debug mode 2021-10-15 09:29:25 -04:00
Tyler Goodlet a42ec1f571 Add nooz 2021-10-15 09:28:45 -04:00
Tyler Goodlet 533457c64d Handle nested multierror case on windows 2021-10-15 09:16:51 -04:00
Tyler Goodlet 51259c4809 Pass uid not actor object 2021-10-14 13:46:27 -04:00
Tyler Goodlet 7ee121aeaf Try to handle variable windows errors 2021-10-14 13:39:46 -04:00
Tyler Goodlet 9d83ef82b2 Remove union type for root getter 2021-10-14 13:39:46 -04:00
Tyler Goodlet fa317d1600 Change lock helper to take an actor uid tuple 2021-10-14 13:39:46 -04:00
Tyler Goodlet 6f5c35dd1b Fix missing task status type 2021-10-14 13:39:46 -04:00
Tyler Goodlet b14699d40b Adjust debugger tests to expect depth > 1 crashes 
With the new fixes to the trio spawner we can expect that both root
*and* depth > 1 nursery owning actors will now not clobber any children
that are in debug (either via breakpoint or through crashing). The tests
changed now include more checks which ensure the 2nd level parent-ish
actors also bubble up through into `pdb` and don't kill any of their
(crashed) children before they're done themselves debugging.
 2021-10-14 13:39:46 -04:00
Tyler Goodlet daa28ea0e9 Handle depth > 1 nursery owners which use debug mode 2021-10-14 13:39:46 -04:00
Tyler Goodlet 4b2710b8a5 Add tty lock acquire ctx mngr 2021-10-14 13:39:46 -04:00
Tyler Goodlet d30ce96740 Breakout `wait_for_parent_stdin_hijack()`, increase root pdb checker poll time 2021-10-14 13:39:46 -04:00
Tyler Goodlet f3a6ab62af Use debugger helper in nursery and spawn tasks 2021-10-14 13:39:46 -04:00
Tyler Goodlet 62035078ce Reduce some loglevels, stick in comment about blocking till next tick 2021-10-14 13:39:46 -04:00
Tyler Goodlet 893bad72d5 Add a maybe-open-debugger helper 2021-10-14 13:39:46 -04:00
Tyler Goodlet 77ec29008d Simplify to soft and hard reap sequences 
This is actually surprisingly easy to grok having gone through a lot of
pain understanding edge cases in the zombie lord dev branch. Basically
we just need to make sure actors are managed in a 2 step reap sequence.
In the "soft" reap phase we wait for the process to terminate on its own
concurrently with (maybe) waiting for its portal's final result (if it's
a `.run_in_actor()`). If this path is cancelled or errors, then we do
a "hard" reap where we timeout and send a signal to the proc to
terminate immediately. The only last remaining trick is to tie in the
root-is-debugger-aware logic to yet again avoid tty clobbers.
 2021-10-14 13:39:46 -04:00
Tyler Goodlet 2df16c1557 Lol, fix sub-actor case 2021-10-14 13:39:46 -04:00
Tyler Goodlet 46ff558556 Unwind process opening and shield hard reap 2021-10-14 13:39:46 -04:00
Tyler Goodlet bb9d9c74b1 Do immediate remote task cancels 
As for `Actor.cancel()` requests, do the same for
`Actor._cancel_task()` but use `_invoke()` to ensure
correct msg transactions with caller. Don't cancel task
cancels on a cancel-all-tasks operation in attempt at
more determinism.
 2021-10-14 13:39:46 -04:00
Tyler Goodlet 41f0992445 Don't whine about ; it ain't rpc 2021-10-14 13:39:46 -04:00
Tyler Goodlet 7643bbf183 Make actor runtime cancellation immediate 2021-10-14 13:39:46 -04:00
goodboy dfeebd6382
Merge pull request #243 from goodboy/less_logging 
Less logging, add a `CANCEL` log level
 2021-10-14 13:37:28 -04:00
Tyler Goodlet 6cda17436a Add nooz 2021-10-14 11:47:06 -04:00
Tyler Goodlet 1f0cc15675 Just set flag for use-after-closed service nursery calls 2021-10-06 17:02:13 -04:00
Tyler Goodlet 10f66e5141 De-noise warnings, add a 'cancel' log level 
Now that we're on our way to a (somewhat) serious beta release I think
it's about time to start de-noising the logging emissions. Since we're
trying out this approach of "stack layer oriented" log levels, I figured
this is a good time to move most of the "warnings" to what they should
be: cancellation monitoring status messages. The level is set to 16
which is just above our "runtime" level but just below the traditional
"info" level. I think this will be a decent approach since usually if
you're confused about why your `tractor` app is behaving unlike you
expect, it's 90% of the time going to be to do with cancellation or
error propagation. This this setup a user can specify the 'cancel' level
and see all the msgs pertaining to both actor and task-in-actor
cancellation mechanics.
 2021-10-06 17:02:13 -04:00
Tyler Goodlet 4d5a5c147a Move core actor runtime logging to, well, "runtime" 2021-10-06 17:02:13 -04:00
Tyler Goodlet d2f0843041 Make custom log levels report the right stack frame 
The stdlib's `logging.LoggingAdapter` doesn't currently pass through
`stacklevel: int` down to its wrapped logger instance. Hack it here
and get our msgs looking like they would if using a built-in level.
 2021-10-06 17:02:13 -04:00
Tyler Goodlet 3f6d4d6af4 Don't log.error if it was intentional 2021-10-06 17:02:13 -04:00
goodboy 21e60554cc
Merge pull request #214 from goodboy/optional_msgspec_support 
Optional msgspec support
 2021-10-06 17:01:44 -04:00
Tyler Goodlet b496e790fe Use from `.from_stream()` in TCP handler 2021-10-06 15:54:27 -04:00
Tyler Goodlet c6dc96b08c Add "message transport" structured sub-typing 
In an effort to have some kind of more formal interface around the
transport layer, add a `MsgTransport` protocol type and use with
the channel composition of message streams. Start a little "key map"
of `(<codec>, <protocol>)` to `MsgTransport` types which can be
dynamically loaded. Add a `Channel.from_stream()` constructor thus
cleaning up the mangled logic that was in the constructor based on
inputs. Drop all the "auto reconnect" channel logic for now since
nothing is using it (internally) and it's likely it will need rework
once we bring in a protocol besides TCP.
 2021-10-06 15:54:27 -04:00
Tyler Goodlet 135459ca25 Tolerate one decode error; may have been a registry ping 2021-10-05 13:37:17 -04:00
Tyler Goodlet ef75883b62 Add fragment 2021-10-05 13:37:17 -04:00
Tyler Goodlet f7fc464ce8 Add `msgspec` mentions to readme 2021-10-05 13:37:17 -04:00
Tyler Goodlet 07e8821cd5 Add a stream type factory 2021-10-05 13:37:17 -04:00
Tyler Goodlet 5b23a3bc35 Don't expect list value from registry 2021-10-05 13:37:17 -04:00
Tyler Goodlet 1382ad653d Ugh, appease mypy yet again 2021-10-05 13:37:17 -04:00
Tyler Goodlet 076f37c589 Attempt to gracefully handle channel breakage? 2021-10-05 13:37:17 -04:00
Tyler Goodlet 19d6885243 Ensure tuple for passed in arbiter addr 2021-10-05 13:37:17 -04:00
Tyler Goodlet dbc4e3dd46 Pin to latest and greatest `msgspec` 2021-10-05 13:37:17 -04:00
Tyler Goodlet 93726f1392 Call registry getter method in test 2021-10-05 13:37:17 -04:00
Tyler Goodlet 486e983964 Cast `defaultdict` to `dict` for registry get 2021-10-05 13:37:17 -04:00
Tyler Goodlet 1ab495a64d Map broken stream errs to transport closed; msgspec seems to be racy 2021-10-05 13:37:17 -04:00
Tyler Goodlet 562419c907 Convert actor UIDs to hashable tuples 
`msgspec` sends python lists over the wire
(https://github.com/jcrist/msgspec/issues/30) which is fine and dandy
but we use them as lookup keys so we need to be sure we tuple-cast
first.
 2021-10-05 13:37:17 -04:00