forked from goodboy/tractor
1
0
Fork 0
Commit Graph

1699 Commits (5c1401bf813097dc210f75b0b530318dd70ba2f9)

Author SHA1 Message Date
Tyler Goodlet 121a8cc891 Drop `Optional` usage from root mod 2023-01-26 16:00:08 -05:00
Tyler Goodlet c54b8ca4ba Begin deprecation of `arbiter_addr` -> `registry_addr` 2023-01-26 16:00:08 -05:00
goodboy de93c8257c
Merge pull request #349 from goodboy/prompt_on_ctrlc
Re-draw `pdbpp` prompt on `SIGINT`
2023-01-26 15:56:37 -05:00
Tyler Goodlet 5b8a87d0f6 Slightly better `xonsh` check hack, fix typing 2023-01-26 15:48:15 -05:00
Tyler Goodlet 9e5c8ce6f6 Add nooz file 2023-01-26 15:39:03 -05:00
Tyler Goodlet 965cd406a2 Use std `pdbpp` release 2023-01-26 15:27:55 -05:00
Tyler Goodlet 2e278ceb74 Add a super hacky check for `xonsh`, smh.. 2023-01-26 15:26:43 -05:00
Tyler Goodlet 6d124db7c9 Never run ctlc-with-intermediary-actor cases locally either 2023-01-26 12:44:13 -05:00
Tyler Goodlet dba8118553 Always attempt prompt redraw on ctl-c in REPL
The stdlib has all sorts of muckery with ignoring SIGINT in the
`Pdb._cmdloop()` but here we just override all that since we don't trust
their decisions about cancellation handling whatsoever. Adds
a `Lock.repl: MultiActorPdb` attr which is set by any task which
acquires root TTY lock indicating (via actor global state) that the
current actor is using the debugger REPL and can be expected to re-draw
the prompt on SIGINT. Further we mask out log messages from any actor
who also has the `shield_sigint_handler()` enabled to avoid logging
noise when debugging.
2023-01-26 12:44:13 -05:00
Tyler Goodlet fca2e7c10e Simplify closed abruptly log msg 2023-01-26 12:44:13 -05:00
Tyler Goodlet 5ed62c5c54 Add note about intermediary-actor in debug issue 2023-01-26 12:44:13 -05:00
goodboy 588b7ca7bf
Merge pull request #344 from goodboy/harden_cluster_tests
Harden cluster tests
2022-12-12 15:02:23 -05:00
Tyler Goodlet d8214735b9 Add bugfix nooz 2022-12-12 14:53:59 -05:00
Tyler Goodlet 48f6d514ef Handle earlier name error crash in debug test 2022-12-12 14:05:32 -05:00
Tyler Goodlet 6c8cacc9d1 Adjust all default is `None` annots (per new `mypy`) 2022-12-12 13:18:22 -05:00
Tyler Goodlet 38326e8c15 Avoid error on context double pops 2022-12-11 23:46:33 -05:00
Tyler Goodlet b5192cca8e Always greedily `list`-cast`mngrs` input sequence 2022-12-11 23:20:58 -05:00
Tyler Goodlet c606be8c64 Passthrough runtime kwargs from `open_actor_cluster()` 2022-12-11 19:56:08 -05:00
Tyler Goodlet d8e48e29ba Add `mngrs=(<gen_comprehension>)` test 2022-12-11 19:56:01 -05:00
goodboy a0f6668ce8
Merge pull request #333 from goodboy/exceptiongroups
`ExceptiongGroup`s and `trio>=0.22`
2022-10-14 20:11:26 -04:00
Tyler Goodlet 274c66cf9d Add nooz 2022-10-14 19:42:23 -04:00
Tyler Goodlet f2641c8964 Avoid "task never called `.started()`" runtime erros when cancelling 2022-10-14 19:42:23 -04:00
Tyler Goodlet c47575997a Expand nested case to include error prop and breakpointing 2022-10-14 19:42:23 -04:00
Tyler Goodlet f39414ce12 Drop error-repacking for `.run_in_actor()`s block
If we pack the nursery parent task's error into the `errors` table
directly in the handler, we don't need to specially handle packing that
same error into any exception group raised while handling sub-actor
cancellation; drops some ugly indentation ;)
2022-10-14 19:42:23 -04:00
Tyler Goodlet 0a1bf8e57d Tolerate eg in runtime test teardown 2022-10-14 19:42:23 -04:00
Tyler Goodlet e298b70edf Drop added `.pdp()` level msgs used duringn dev 2022-10-14 19:42:23 -04:00
Tyler Goodlet c0dd5d7ffc Adjust multi-daemon test to be more deterministic 2022-10-14 19:42:23 -04:00
Tyler Goodlet 347591c348 Expect egs in tests which retreive portal results 2022-10-14 19:42:23 -04:00
Tyler Goodlet 38f9d35dee Fix errors table type annot 2022-10-14 19:42:23 -04:00
Tyler Goodlet 88448f7281 Fix handler type annot 2022-10-14 19:42:23 -04:00
Tyler Goodlet 0956d5f461 Restore the `trio` SIGINT handler, cancel root lock tasks on no-peers
Pretty sure this is the final touch to alleviate all our debug lock
headaches! Instead of trying to revert to the "last" handler (as `pdb`
does internally in the stdlib) we always just revert to the handler
`trio` registers during startup. Further this seems to allow cancelling
the root-side locking task if it's detected as stale IFF we only do this
when the root actor is in a "no more IPC peers" state.

Deatz:
- (always) set `._debug.Lock._trio_handler` as the `trio` version, not
  some last used handler to make sure we're getting the ctrl-c handling
  we want when not in debug mode.
- assign the trio handler in `open_root_actor()`
  `._runtime._async_main()` to be sure it's applied in subactors as well
  as the root.
- only do debug lock blocking and root-side-locking-task cancels when
  a "no peers" condition is detected in the root actor: i.e. no IPC
  channels are detected by the root meaning it's impossible any actor
  has a sane lock-state ongoing for debug mode.
2022-10-14 18:18:01 -04:00
Tyler Goodlet c646c79a82 Adjust root-errors debug tests for blocking and egs 2022-10-14 18:18:01 -04:00
Tyler Goodlet 33f2234baf Hide some stack layers the user doesn't really need to see 2022-10-14 18:18:01 -04:00
Tyler Goodlet 7521bded3d Pack error from the parent task into the actor nursery 2022-10-14 18:16:51 -04:00
Tyler Goodlet 0f523b65fb Change cancel test over the exception group 2022-10-14 18:16:51 -04:00
Tyler Goodlet 50fe098e06 First pass, swap `MultiError` for `BaseExceptionGroup` 2022-10-14 18:16:51 -04:00
Tyler Goodlet d87d6af7e1 Add `exceptiongroup` (3.11 backport lib) as dep 2022-10-14 18:16:51 -04:00
Tyler Goodlet df69aedcd5 Pin to latest `trio` version 2022-10-14 18:16:51 -04:00
Tyler Goodlet b15e4ed9ce Adjust "no arbiter" test for new runtime defaults
Turns out this test was being silently ignored due to incorrect usage of
sync opening of our `.open_nursery()` block (with a `with` not `async
with`) and thus was an noop XD

Instead this fixes the test to call a `tractor` discovery built-in
without starting the runtime (which is now done implicitly when a user
opens a nursery) which should result in the prior expected outcome,
a `RuntimeError`.
2022-10-12 12:46:20 -04:00
Tyler Goodlet 98056f6ed7 Move logging context map into `log.py` module 2022-10-12 12:46:20 -04:00
goodboy 247d3448ae
Merge pull request #337 from goodboy/debug_lock_blocking
Debug lock blocking
2022-10-12 12:41:14 -04:00
Tyler Goodlet fc17f6790e Bump `towncrier` alpha version 2022-10-12 12:36:09 -04:00
Tyler Goodlet b81b6be98a Drop extra log msgs, some old commented code 2022-10-12 12:35:35 -04:00
Tyler Goodlet 72fbda4cef Add nooz file 2022-10-12 12:35:11 -04:00
Tyler Goodlet fb721f36ef Support debug-lock blocking, use on no-more IPC
This is a lingering debugger locking race case we needed to handle:

- child crashes acquires TTY lock in root and attaches to `pdb`
- child IPC goes down such that all channels to the root are broken
  / non-functional.
- root is stuck thinking the child is still in debug even though it
  can't be contacted and the child actor machinery hasn't been
  cancelled by its parent.
- root get's stuck in deadlock with child since it won't send a cancel
  request until the child is finished debugging, but the child can't
  unlock the debugger bc IPC is down.

To avoid this scenario add debug lock blocking list via
`._debug.Lock._blocked: set[tuple]` which holds actor uids for any actor
that is detected by the root as having no transport channel connections
with said root (of which at least one should exist if this sub-actor at
some point acquired the debug lock). The root consequently checks this
list for any actor that tries to (re)acquire the lock and blocks with
a `ContextCancelled`. When a debug condition is tested in
`._runtime._invoke` the context's `._enter_debugger_on_cancel` which
is set to `False` if the actor is on the block list in which case the
post-mortem entry is skipped.

Further this adds a root-locking-task side cancel scope to
`Lock._root_local_task_cs_in_debug` which can be cancelled by the root
runtime when a stale lock is detected after all IPC channels for the
actor have been torn down. NOTE: right now we're NOT doing this since it
seems to cause test failures likely due because it may cause pre-mature
cancellation and maybe needs a bit more experimenting?
2022-10-11 20:00:05 -04:00
Tyler Goodlet 734d8dd663 Move `trio` scope outside first inter-task-chan receive 2022-10-11 20:00:05 -04:00
Tyler Goodlet 30ea7a06b0 Avoid inf nursery hang by reversing `async with` ordering 2022-10-11 20:00:05 -04:00
Tyler Goodlet 3398153c52 Add timeout around `trio`-callee-task 2022-10-11 20:00:05 -04:00
Tyler Goodlet 1c480e6c92 Add `Context` cancel message and debug toggle flag
In the case of a callee-side context cancelling itself it can be handy
to let the caller-side task know (even if through logging) that the
cancel was due to some known reason. Make `.cancel()` accept such
a message on the callee side and have it included in the
`._runtime._invoke()` raised `ContextCancelled` emission.

Also add a `Context._trigger_debugger_on_cancel: bool` flag which can be
set to `False` to avoid the debugger post-mortem crash mode from
engaging on cross-context tasks which cancel themselves for a known
reason (as is needed for blocked tasks in the debug TTY-lock machinery).
2022-10-11 20:00:05 -04:00
goodboy dfdad4d1fa
Merge pull request #336 from goodboy/callable_key_maybe_open_context
Callable key input to maybe open context
2022-10-10 00:32:27 -04:00