Compare commits

..

2 Commits

Author SHA1 Message Date
Gud Boi e617b498ec Merge the supervise error handlers into one
Step B2 of the `._ria_nursery` removal (issue #477; see
`ai/conc-anal/ria_nursery_removal_plan.md`). With the 2ndary
nursery gone (step B), the two nested error handlers in
`_open_and_supervise_one_cancels_all_nursery` collapse to one,

- the outer `except (Exception, BaseExceptionGroup,
  trio.Cancelled)` existed to catch errors bubbling from the
  old `._ria_nursery.__aexit__` reaper-group; that nursery no
  longer exists.
- trace shows the outer handler's `raise` was already DEAD: the
  inner handler records `errors[uid]` as its first action, so
  `errors` is always non-empty by the time anything could reach
  the outer handler, and the `finally`'s raise-from-`errors`
  always superseded the outer `raise`.
- so fold both into a single `except BaseException as
  _scope_err` guarding the lone daemon nursery; the `finally`
  (unchanged) still raises the collected `errors` as a single
  exc or `BaseExceptionGroup`.
- drop the now-unused `outer_err`/`inner_err` locals.

Behaviour-preserving (net ~30 lines lighter); the big diff is
the one-level de-indent of the handler body. The two remaining
`maybe_wait_for_debugger()` guards collapse to the single
pre-teardown wait.

Prompt-IO: ai/prompt-io/claude/20260702T222544Z_9201a2ed_prompt_io.md

(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
2026-07-02 18:33:17 -04:00
Gud Boi b14bc8e273 Doc step-B2 handler-merge + prompt-io
Split from the step-B2 code commit to keep the runtime diff
free of `ai/` meta noise,

- `ai/conc-anal/ria_nursery_removal_plan.md`: add a "Step-B2
  outcome" section — the dead-outer-`raise` trace, why the
  merge is behavior-preserving, and the gate results.
- `ai/prompt-io/claude/20260702T222544Z_9201a2ed_*`: NLNet
  provenance (log + unedited raw) for the step-B2 work.

(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
2026-07-02 18:33:05 -04:00
4 changed files with 285 additions and 145 deletions

View File

@ -229,6 +229,53 @@ Gate (trio backend, all 0-failure):
completing within an 800s bound — but split across the
above three runs EVERY module passed under step B.
## Step-B2 outcome (2026-07-02, done in tree)
Step B committed as `9201a2ed` (code) + `d2e812fb` (docs), then
branched to `drop_ria_nursery`. Step B2 (the deferred
handler-merge) implemented on top (uncommitted):
- the two nested handlers in
`_open_and_supervise_one_cancels_all_nursery` collapse to
ONE `except BaseException as _scope_err` + the existing
`finally`. The `outer_err`/`inner_err` locals go away.
Why it's safe (trace, not hope): the OLD inner handler records
`errors[actor.aid.uid]` as its FIRST statement (before any
await). So whenever an error path runs, `errors` is non-empty.
The OLD outer handler was only reachable via leakage from the
inner handler (it catches `BaseException`, so nothing from the
`yield` scope bypasses it) — and by then `errors` is already
populated, so the `finally`'s `raise` from `errors` ALWAYS
superseded the outer handler's own `raise`. i.e. the outer
`raise` was dead. The outer handler's other effects
(`_scope_error`, a 2nd debugger-wait, child-cancel) are
redundant with the merged handler + `finally`. So one handler
+ `finally` is observably equivalent.
Residual nuance (accepted): in the rare "`trio.Cancelled`
delivered during the non-shielded `maybe_wait_for_debugger`"
path, the merged form may leave `_cancel_called` False (cancel
happens after the wait), so `open_nursery`'s tb-hiding guard
(`not cancel_called and _scope_error`) can show a tb it
previously hid. More informative, not less; no test asserts on
it.
Gate (box ran ~2.7x slow this session, load-induced
`TooSlowError` flakiness on timing tests — NOT code; see
[[env_cpu_throttle_masquerades_as_regression]]):
- baseline (pre-B2 tip `9201a2ed`) full suite
(`-k 'not dynamic_pub_sub'`) = 300 passed + 1
`test_ext_types_over_ipc` `TooSlowError` that passes 6/6 in
isolation (4.89s).
- B2 error/cancel gate (`test_cancellation remote_exc_relay
inter_peer_cancellation advanced_faults oob_cancellation
to_actor spawning local rpc`) = 71 passed, 1 xfailed
(125s).
- B2 full-suite run: see `b2_full.log` (result appended on
completion). RECOMMEND a clean full-suite run on a
normal-speed box before this merges.
## Verification gate
- `tests/test_cancellation.py test_spawning.py test_local.py

View File

@ -0,0 +1,68 @@
---
model: claude-opus-4-8
service: claude
session: c7860302-6a33-4d92-821c-0807d68be464
timestamp: 2026-07-02T22:25:44Z
git_ref: 9201a2ed
scope: code
substantive: true
raw_file: 20260702T222544Z_9201a2ed_prompt_io.raw.md
---
## Prompt
Continuing on branch `drop_ria_nursery` after step B
(`9201a2ed`). User green-lit the deferred error-handler merge:
> yes, i checked us out on a new branch for the ria-tn removal;
> work off the new wkt branch drop_ria_nursery.
and (re the migration) "nope, later we'll do this", plus:
> do ensure the test suite is green in the current state and
> those going forward ya?
and the standing reminder: no commits without a human ack.
## Response summary
Step B2 of the `._ria_nursery` removal (issue #477): merge the
two nested error handlers in
`_open_and_supervise_one_cancels_all_nursery` into one, now
that the 2ndary nursery (whose `__aexit__` the outer handler
originally guarded) is gone.
Trace-driven justification (see
`ai/conc-anal/ria_nursery_removal_plan.md` Step-B2 section):
the old inner handler records `errors[uid]` as its FIRST
action, so `errors` is always non-empty by the time control
could reach the outer handler — making the outer handler's
`raise` DEAD (always superseded by the `finally`'s
raise-from-`errors`). Collapsing to a single
`except BaseException as _scope_err` + the existing `finally`
is therefore behavior-preserving.
- removed the outer `try` + the
`except (Exception, BaseExceptionGroup, trio.Cancelled)`
handler; de-indented the inner handler one level.
- dropped the now-unused `outer_err`/`inner_err` locals
(single `_scope_err`).
- net ~30 lines lighter.
Gated hard: baseline (pre-B2 tip) = 300 passed + 1
load-induced `TooSlowError` flake (passes 6/6 in isolation);
B2 error/cancel gate (`test_cancellation remote_exc_relay
inter_peer_cancellation advanced_faults oob_cancellation
to_actor spawning local rpc`) = 71 passed, 1 xfailed; B2
full-suite run following.
## Files changed
- `tractor/runtime/_supervise.py` — collapse the two handlers
into one; drop `outer_err`/`inner_err`
## Human edits
None yet — committed via the drafted
`.claude/git_commit_msg_ria_b2.md` (user-driven
`git commit --edit`).

View File

@ -0,0 +1,55 @@
---
model: claude-opus-4-8
service: claude
timestamp: 2026-07-02T22:25:44Z
git_ref: 9201a2ed
diff_cmd: git diff 9201a2ed..drop_ria_nursery
---
# Raw AI output (diff-ref mode)
Step-B2 code lives on `drop_ria_nursery` after `9201a2ed`; per
diff-ref mode the verbatim content is reachable via the pointer
below.
## Generated files
> `git diff 9201a2ed..drop_ria_nursery -- tractor/runtime/_supervise.py`
`_open_and_supervise_one_cancels_all_nursery`:
- removed the outer `try:` wrapper and the
`except (Exception, BaseExceptionGroup, trio.Cancelled) as
_outer_err:` safety-net handler.
- the former inner `except BaseException` is now THE handler,
renamed local `_inner_err` -> `_scope_err`, de-indented one
level; it sets `an._scope_error`, records `errors[uid]`,
waits on the debugger, `_join_procs.set()`, then a shielded
classify/log + snapshot-ria + `an.cancel()` + 0.5s-bounded
`_reap_ria_portals()`. No re-raise (the `finally` raises
from `errors`).
- `finally` block unchanged.
- dropped the `outer_err`/`inner_err` local decls at fn top.
(The diff is large — ~119+/149- — because de-indenting the
handler body one level rewrites every line in the block; the
logic delta is just "two handlers -> one".)
## Test runs (verbatim)
```
baseline (pre-B2, step-B tip 9201a2ed), full suite
(dynamic_pub_sub deselected):
1 failed, 300 passed, 9 skipped, 2 deselected, 1 xfailed,
2 xpassed in 1499.49s
-> the 1 failure = test_ext_types_over_ipc[...] trio.TooSlowError
(load-induced; passes 6/6 in isolation in 4.89s)
B2 error/cancel gate:
tests/test_cancellation test_remote_exc_relay
test_inter_peer_cancellation test_advanced_faults
test_oob_cancellation test_to_actor test_spawning test_local
test_rpc
-> 71 passed, 1 xfailed in 125.26s
B2 full-suite run: see b2_full.log
```

View File

@ -623,9 +623,6 @@ async def _open_and_supervise_one_cancels_all_nursery(
# normally don't need to show user by default
__tracebackhide__: bool = hide_tb
outer_err: BaseException|None = None
inner_err: BaseException|None = None
# the collection of errors retreived from spawned sub-actors
errors: dict[tuple[str, str], BaseException] = {}
@ -645,156 +642,129 @@ async def _open_and_supervise_one_cancels_all_nursery(
errors
)
try:
try:
# spawning of actors happens in the caller's scope
# after we yield upwards
yield an
# spawning of actors happens in the caller's scope
# after we yield upwards
yield an
# When we didn't error in the caller's scope,
# signal all process-monitor-tasks to conduct
# the "hard join phase".
log.runtime(
'Waiting on subactors to complete:\n'
f'>}} {len(an._children)}\n'
)
an._join_procs.set()
# When we didn't error in the caller's scope,
# signal all process-monitor-tasks to conduct
# the "hard join phase".
log.runtime(
'Waiting on subactors to complete:\n'
f'>}} {len(an._children)}\n'
)
an._join_procs.set()
# collect results (and errors) from all
# `.run_in_actor()` children then cancel
# each, one reaper task per child.
await _reap_ria_portals(an, errors)
# collect results (and errors) from all
# `.run_in_actor()` children then cancel
# each, one reaper task per child.
await _reap_ria_portals(an, errors)
except BaseException as _inner_err:
inner_err = _inner_err
errors[actor.aid.uid] = inner_err
# Single one-cancels-all handler for the (now single)
# daemon nursery. Pre-#477 a 2ndary `._ria_nursery`
# required a separate *outer* handler to catch errors
# bubbling from its task-reaping `__aexit__`; with that
# nursery gone this lone handler covers every scope
# error. NB: we deliberately do NOT re-raise here — the
# `finally` below raises the collected `errors` (as a
# single exc or `BaseExceptionGroup`), which already
# superseded the old outer handler's `raise` anyway
# since `errors` is populated (below) before any await.
except BaseException as _scope_err:
an._scope_error = _scope_err
errors[actor.aid.uid] = _scope_err
# If we error in the root but the debugger is
# engaged we don't want to prematurely kill (and
# thus clobber access to) the local tty since it
# will make the pdb repl unusable.
# Instead try to wait for pdb to be released before
# tearing down.
await debug.maybe_wait_for_debugger(
child_in_debug=an._at_least_one_child_in_debug
)
# if the caller's scope errored then we activate our
# one-cancels-all supervisor strategy (don't
# worry more are coming).
an._join_procs.set()
# XXX NOTE XXX: hypothetically an error could
# be raised and then a cancel signal shows up
# slightly after in which case the `else:`
# block here might not complete? For now,
# shield both.
with trio.CancelScope(shield=True):
etype: type = type(inner_err)
if etype in (
trio.Cancelled,
KeyboardInterrupt,
) or (
is_multi_cancelled(inner_err)
):
log.cancel(
f'Actor-nursery cancelled by {etype}\n\n'
f'{current_actor().aid.uid}\n'
f' |_{an}\n\n'
# TODO: show tb str?
# f'{tb_str}'
)
elif etype in {
ContextCancelled,
}:
log.cancel(
'Actor-nursery caught remote cancellation\n'
'\n'
f'{inner_err.tb_str}'
)
else:
log.exception(
'Nursery errored with:\n'
# TODO: same thing as in
# `._invoke()` to compute how to
# place this div-line in the
# middle of the above msg
# content..
# -[ ] prolly helper-func it too
# in our `.log` module..
# '------ - ------'
)
# snapshot `.run_in_actor()` children
# BEFORE cancelling: each backend
# spawn-task pops its `._children`
# entry as the proc gets reaped.
ria_children: list = [
(portal, subactor)
for subactor, _, portal
in an._children.values()
if portal in
an._cancel_after_result_on_exit
]
# cancel all subactors
await an.cancel()
# then collect any already-relayed
# results/errors from ria children.
# Tightly bounded: anything
# collectable is already queued in
# the local ctx (relayed BEFORE the
# cancel above); a child hard-killed
# without relaying just parks its
# reaper which then self-cleans (a
# `trio.Cancelled` result is never
# stashed), mirroring the old
# backend-side reaper-vs-`soft_kill`
# cancel race.
with trio.move_on_after(0.5):
await _reap_ria_portals(
an,
errors,
ria_children=ria_children,
)
# Safety-net handler for anything escaping the inner
# `except BaseException` above (e.g. a `trio.Cancelled`
# during its non-shielded debugger-wait, or a failure
# inside the shielded teardown itself). Post-#477 the
# 2ndary `._ria_nursery` (and its dedicated handler) is
# gone; this now guards the single daemon nursery scope.
# TODO: can this be merged into the inner handler now that
# there's only one nursery? (higher-risk, own PR)
except (
Exception,
BaseExceptionGroup,
trio.Cancelled
) as _outer_err:
outer_err = _outer_err
an._scope_error = outer_err or inner_err
# XXX: yet another guard before allowing the cancel
# sequence in case a (single) child is in debug.
# If we error in the root but the debugger is
# engaged we don't want to prematurely kill (and
# thus clobber access to) the local tty since it
# will make the pdb repl unusable.
# Instead try to wait for pdb to be released before
# tearing down.
await debug.maybe_wait_for_debugger(
child_in_debug=an._at_least_one_child_in_debug
)
# If actor-local error was raised while waiting on
# ".run_in_actor()" actors then we also want to cancel all
# remaining sub-actors (due to our lone strategy:
# one-cancels-all).
if an._children:
log.cancel(
'Actor-nursery cancelling due error type:\n'
f'{outer_err}\n'
)
with trio.CancelScope(shield=True):
await an.cancel()
raise
# if the caller's scope errored then we activate our
# one-cancels-all supervisor strategy (don't
# worry more are coming).
an._join_procs.set()
# XXX NOTE XXX: hypothetically an error could
# be raised and then a cancel signal shows up
# slightly after in which case the `else:`
# block here might not complete? For now,
# shield both.
with trio.CancelScope(shield=True):
etype: type = type(_scope_err)
if etype in (
trio.Cancelled,
KeyboardInterrupt,
) or (
is_multi_cancelled(_scope_err)
):
log.cancel(
f'Actor-nursery cancelled by {etype}\n\n'
f'{current_actor().aid.uid}\n'
f' |_{an}\n\n'
# TODO: show tb str?
# f'{tb_str}'
)
elif etype in {
ContextCancelled,
}:
log.cancel(
'Actor-nursery caught remote cancellation\n'
'\n'
f'{_scope_err.tb_str}'
)
else:
log.exception(
'Nursery errored with:\n'
# TODO: same thing as in
# `._invoke()` to compute how to
# place this div-line in the
# middle of the above msg
# content..
# -[ ] prolly helper-func it too
# in our `.log` module..
# '------ - ------'
)
# snapshot `.run_in_actor()` children
# BEFORE cancelling: each backend
# spawn-task pops its `._children`
# entry as the proc gets reaped.
ria_children: list = [
(portal, subactor)
for subactor, _, portal
in an._children.values()
if portal in
an._cancel_after_result_on_exit
]
# cancel all subactors
await an.cancel()
# then collect any already-relayed
# results/errors from ria children.
# Tightly bounded: anything
# collectable is already queued in
# the local ctx (relayed BEFORE the
# cancel above); a child hard-killed
# without relaying just parks its
# reaper which then self-cleans (a
# `trio.Cancelled` result is never
# stashed), mirroring the old
# backend-side reaper-vs-`soft_kill`
# cancel race.
with trio.move_on_after(0.5):
await _reap_ria_portals(
an,
errors,
ria_children=ria_children,
)
finally:
# No errors were raised while awaiting ".run_in_actor()"