: system: audit tests #2005
Closed
: system: audit tests #2005
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
meta-cla
bot
added
the
CLA Signed
|
@shayne-fletcher has exported this pull request. If you are a Meta employee, you can view the originating Diff in D87929235. |
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Differential Revision: D87929235
shayne-fletcher
force-pushed
the
export-D87929235
branch
from
0b62c05 to
264b395
Compare
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Differential Revision: D87929235
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Differential Revision: D87929235
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Differential Revision: D87929235
shayne-fletcher
force-pushed
the
export-D87929235
branch
from
264b395 to
54fc97a
Compare
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Differential Revision: D87929235
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Differential Revision: D87929235
Summary: this change applies the same V0->V1 test audit to 'system_actor.rs' that we did for 'proc_actor.rs'. i went through all 7 tests in this file and tagged each one as V0-specific, with no V1 equivalent: they all exercise `SystemActor`-centric behavior that simply doesn’t exist in the V1 design (world-level supervision state, world orchestration via hosts joining and `UpsertWorld`, snapshot filtering, and `ReportingRouter`/dynamic address updates). the new comments spell out what each test is validating in V0 and why there is no direct V1 analog, and where relevant they point at the closest V1 mechanism (for example, undeliverable handling now covered by `actor_mesh::test_undeliverable_message_return`). there are no behavior changes here; this is purely test documentation and migration context for the remaining V0 multiprocess tests in 'system_actor.rs'. Reviewed By: pzhan9 Differential Revision: D87890512
Summary: port stop-timeout / abort coverage from hyperactor_multiprocess into the v1 mesh layer. this adds a `SleepActor` helper and a new `actor_mesh::test_actor_mesh_stop_timeout` that overrides `ACTOR_SPAWN_MAX_IDLE`, spawns sleeping actors across a proc mesh, calls `stop()`, and asserts we get a timeout error within the expected window, proving hung actors are aborted via `Proc::destroy_and_wait` rather than waited out. the V0 test is left in place but re-documented to point at the new v1 test and to note the slight API difference (V0 reports abort counts, V1 surfaces the timeout as an error). Reviewed By: dulinriley Differential Revision: D87904499
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Reviewed By: dulinriley Differential Revision: D87929235
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Reviewed By: dulinriley Differential Revision: D87929235
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Reviewed By: dulinriley Differential Revision: D87929235
shayne-fletcher
force-pushed
the
export-D87929235
branch
2 times, most recently
from
6ccb747 to
e6ed997
Compare
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Reviewed By: dulinriley Differential Revision: D87929235
shayne-fletcher
added a commit
to shayne-fletcher/monarch-1
that referenced
this pull request
Nov 26, 2025
Summary: this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized `SystemActor` model (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator. the only behavioral change in this diff is the removal of `test_channel_dial_count`. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward. as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until `hyperactor_multiprocess` is retired, but they are now explicitly marked as V0-only. Reviewed By: dulinriley Differential Revision: D87929235
|
This pull request has been merged in 60c5129. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
this completes the V0→V1 audit for system.rs. all remaining tests here are V0-specific and do not have V1 equivalents because the centralized
SystemActormodel (join, snapshot, system-wide stop, selective world stop, address-update routing, etc.) does not exist in the mesh-based architecture. each test is now annotated accordingly so it's clear which parts of the file are legacy behavior tied to the old system-wide coordinator.the only behavioral change in this diff is the removal of
test_channel_dial_count. that test was originally written as a debugging aid to understand when dials occur in the V0 forwarding path; it has never been a correctness test and has no meaningful counterpart in V1 (which uses static addressing and structured proc/host meshes). dropping it reduces noise in the V0 test corpus and clarifies which tests matter going forward.as with the previous files, the intent of this stack is audit and classification, not behavior change: V0 tests stay in place until
hyperactor_multiprocessis retired, but they are now explicitly marked as V0-only.Differential Revision: D87929235