test: extract site explorer power shelf tests#2250
Open
poroh wants to merge 1 commit into
Open
Conversation
Contributor
|
Important Review skippedAuto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
WalkthroughThis PR consolidates test infrastructure by establishing a centralized ChangesAPI Core Test Support Consolidation
New Test-Harness Crate Infrastructure
Power-Shelf Test Suite Migration
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes This PR exhibits substantial scope: a new crate with non-trivial test infrastructure (harness builder with background task management, state controller iteration, RPC integration), significant test suite migration (removal of 13 tests from api-core, addition of 12 comprehensive integration tests in site-explorer), and API surface expansion. The test-harness builder requires careful review of lifecycle management (cancellation tokens, JoinSet cleanup, drop guards), transaction handling, and metric wiring. The power-shelf test suite is dense with multiple independent test scenarios, database assertions, metric validation, and state machine behavior. Cross-module refactoring (endpoint-explorer re-exports, logging consolidation) adds moderate coupling complexity. Heterogeneous changes across configuration, types, infrastructure, and integration tests require separate reasoning per layer. Possibly related PRs
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 6
🧹 Nitpick comments (10)
crates/site-explorer/tests/power_shelf.rs (8)
1305-1314: 💤 Low valueRemove commented-out assertion code.
This large block of commented-out state version verification logic should be deleted if it's no longer needed, or re-enabled if it's required for the test.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/site-explorer/tests/power_shelf.rs` around lines 1305 - 1314, Remove the large commented-out assertion block that builds state_versions from final_state (the commented lines referencing state_versions, final_state, and the assert! about state_versions.len() > 1); either delete these comment lines entirely if the verification is no longer needed, or re-enable them by uncommenting and ensuring the test compiles and uses state_versions/state_version properly inside the test function where final_state is available.
1087-1094: 💤 Low valueRemove commented-out code and second early return.
Lines 1089-1091 contain commented-out debug
println!statements that should be deleted. Lines 1092-1094 contain anotherif 1 == 1 { return Ok(()); }block that's redundant given the earlier return at line 1067-1069. Clean up this section entirely.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/site-explorer/tests/power_shelf.rs` around lines 1087 - 1094, Remove the commented-out debug loop and the redundant early-return by deleting the lines containing the commented println!/env.run_power_shelf_controller_iteration() block and the `if 1 == 1 { return Ok(()); }` statement so the test no longer contains dead commented code or a second early return; search for the `println!`/`env.run_power_shelf_controller_iteration().await` snippet and the `if 1 == 1` block in the test and remove them, leaving the function flow as originally intended with only the earlier return.
254-366: 💤 Low valueRemove commented-out code.
Line 350 contains a commented-out call to
explorer.run_single_iteration().await.unwrap();. Either remove this line entirely or document why it's disabled with a clear TODO/FIXME comment.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/site-explorer/tests/power_shelf.rs` around lines 254 - 366, In test_site_explorer_power_shelf_discovery_with_static_ip remove the commented-out call to explorer.run_single_iteration().await.unwrap(); (or replace it with a one-line TODO/FIXME explaining why it's disabled) so there is no stray commented code left in the test; locate the commented line referencing explorer.run_single_iteration() and either delete it or add a concise TODO/FIXME comment directly above the line documenting the reason for disabling the extra iteration.
1481-1491: ⚡ Quick winAvoid unused variable pattern.
Line 1481 uses
let _history_by_ids = ...followed by commented-out assertions at lines 1489-1491. If this query result is genuinely unused, uselet _ = ...(without a variable name) to make the intent explicit. Otherwise, re-enable the assertions or remove the query entirely.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/site-explorer/tests/power_shelf.rs` around lines 1481 - 1491, The local binding `_history_by_ids` is unused after calling `db::state_history::find_by_object_ids` in the test; either replace `let _history_by_ids = ...` with `let _ = db::state_history::find_by_object_ids(...).await?;` to explicitly discard the result, or re-enable the commented assertions that reference `history_by_ids` (and rename `_history_by_ids` to `history_by_ids`) so the query result is actually asserted against; update the call site around `db::state_history::find_by_object_ids`, the `_history_by_ids` binding, and the commented lines that assert `contains_key(&power_shelf1_id)`/`contains_key(&power_shelf2_id)` accordingly.
1700-1720: ⚖️ Poor tradeoffFragile JSON comparison using string manipulation.
Lines 1701-1702 and 1717-1719 compare JSON by removing spaces with
.replace(" ", ""). This approach is brittle and fails if the serialization format changes in other ways (key ordering, escaping, etc.). Consider either:
- Deserializing both strings back to the state enum and comparing structurally
- Using a proper JSON equality library
- Comparing the typed
PowerShelfControllerStatevalues directly without round-tripping through JSON🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/site-explorer/tests/power_shelf.rs` around lines 1700 - 1720, The test currently compares JSON strings by stripping spaces (history_entry.state.replace(" ", "") vs serde_json::to_string(&state)?), which is brittle; instead parse the stored JSON and compare structurally — either deserialize history_entry.state into the concrete type (e.g., PowerShelfControllerState) and assert equality with state, or deserialize both into serde_json::Value via serde_json::from_str and compare the Value to serde_json::to_value(&state). Update the assertions around history_entry.state and found_entry.unwrap().state to perform these structural deserializations/comparisons rather than string replacement.
1279-1286: 💤 Low valueRemove commented-out code and redundant early return.
Lines 1281-1283 contain commented-out iteration code. Lines 1284-1286 contain another
if 1 == 1 { return Ok(()); }block that's redundant. Remove this entire section.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/site-explorer/tests/power_shelf.rs` around lines 1279 - 1286, Remove the commented-out loop and the redundant early-return block: delete the three commented lines calling env.run_power_shelf_controller_iteration() and the subsequent if 1 == 1 { return Ok(()); } so the test continues normally; look for the env.run_power_shelf_controller_iteration() commented loop and the literal if 1 == 1 return Ok(()) and remove those lines together.
127-1755: 🏗️ Heavy liftConsider extracting common test setup patterns.
The test suite contains significant code duplication:
EndpointExplorationReportconstruction (appears ~11 times with similar fields)- DHCP discovery + power shelf creation flow (repeated in multiple tests)
- Expected power shelf database seeding
Extract these into helper methods on
Envor standalone functions to improve maintainability. For example:impl Env { fn create_basic_exploration_report(&self, serial: &str, model: &str) -> EndpointExplorationReport { EndpointExplorationReport { endpoint_type: EndpointType::Bmc, vendor: Some(bmc_vendor::BMCVendor::Nvidia), systems: vec![ComputerSystem { serial_number: Some(serial.to_string()), ..Default::default() }], chassis: vec![Chassis { model: Some(model.to_string()), ..Default::default() }], model: Some(model.to_string()), ..Default::default() } } async fn setup_power_shelf_with_dhcp(&self, power_shelf: &mut FakePowerShelf) -> Result<(), Box<dyn std::error::Error>> { let response = self.api() .discover_dhcp( DhcpDiscovery::builder( power_shelf.bmc_mac_address.to_string(), power_shelf.relay_address.to_string(), ) .tonic_request(), ) .await? .into_inner(); power_shelf.ip = response.address; Ok(()) } }🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/site-explorer/tests/power_shelf.rs` around lines 127 - 1755, Several tests duplicate constructing EndpointExplorationReport, running DHCP discovery to set FakePowerShelf.ip, and seeding expected_power_shelf rows; extract helpers to reduce duplication: add Env::create_basic_exploration_report(&self, serial: &str, model: &str) returning EndpointExplorationReport and Env::setup_power_shelf_with_dhcp(&self, power_shelf: &mut FakePowerShelf) -> Result<(), Box<dyn std::error::Error>> to perform the DhcpDiscovery call and set power_shelf.ip, plus a helper Env::seed_expected_power_shelf(&self, power_shelf: &FakePowerShelf) that wraps db::expected_power_shelf::create; update tests like test_site_explorer_power_shelf_discovery, test_site_explorer_power_shelf_with_static_ip, test_site_explorer_power_shelf_creation_limit, etc. to call these helpers and replace the repeated EndpointExplorationReport literals with create_basic_exploration_report calls.
1513-1520: 💤 Low valueRemove commented-out iteration code.
Lines 1515-1517 contain commented-out controller iteration calls that should be deleted.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/site-explorer/tests/power_shelf.rs` around lines 1513 - 1520, Delete the commented-out controller iteration block (the three lines calling env.run_power_shelf_controller_iteration() that are prefixed with //) so the test no longer contains dead commented code; keep the surrounding TODO comment if desired and leave the remaining early return logic unchanged—look for the env.run_power_shelf_controller_iteration() calls to locate the commented lines to remove.crates/test-harness/src/network/controller.rs (1)
68-110: ⚡ Quick winDocument the state iteration count requirement.
Both
create_underlay_segment()andcreate_admin_segment()run exactly 2 state controller iterations (lines 101-102, 145-146) without explaining why this specific count is needed. This "magic number" should be documented to clarify the provisioning lifecycle requirements.📋 Suggested documentation
let segment = self .api .create_network_segment(tonic::Request::new(request)) .await .expect("Unable to create network segment") .into_inner(); + // Run two iterations to fully provision the segment: + // 1st iteration: allocate resources (VLAN, VNI) + // 2nd iteration: transition to active state self.run_single_iteration().await; self.run_single_iteration().await; TestNetworkSegment {🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/test-harness/src/network/controller.rs` around lines 68 - 110, Add a brief explanatory comment above the two run_single_iteration().await calls in both create_underlay_segment() and create_admin_segment() that documents why exactly two state controller iterations are required for provisioning (e.g., first iteration to enqueue/apply the network creation and a second to reconcile and persist resulting state/allocate IPs), mention any assumptions (such as async job scheduling or reconciliation semantics), and reference run_single_iteration() so future readers understand this “magic number” is intentional and tied to the controller’s two-step provisioning lifecycle.crates/test-harness/src/lib.rs (1)
57-72: 💤 Low valueConsider consolidating error handling in test setup.
The
test_domain()method chains multiple.unwrap()calls, including a particularly awkward.map().unwrap().unwrap()pattern at lines 68-70. While panicking on test setup failure is acceptable, the readability could be improved.♻️ Suggested refactor for clarity
pub async fn test_domain(&self) -> TestDomain { let name = "testharness.example.com"; - let id = self + let response = self .api .create_domain(Request::new(rpc::protos::dns::CreateDomainRequest { name: name.to_string(), })) .await - .unwrap() - .into_inner() - .id - .map(::carbide_uuid::domain::DomainId::try_from) - .unwrap() - .unwrap(); + .expect("create_domain RPC failed"); + + let proto_id = response + .into_inner() + .id + .expect("create_domain returned no id"); + + let id = ::carbide_uuid::domain::DomainId::try_from(proto_id) + .expect("invalid DomainId from create_domain"); + TestDomain { id, name } }🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/test-harness/src/lib.rs` around lines 57 - 72, The test_domain function uses multiple chained .unwrap() calls (notably the .map(::carbide_uuid::domain::DomainId::try_from).unwrap().unwrap()) which hurts readability; replace these with explicit, contextual failures by using expect() with clear messages on the RPC/result unwraps and convert the map/unwrap/unwrap into .map(|s| ::carbide_uuid::domain::DomainId::try_from(s).expect("failed to parse DomainId from response")) (or alternatively pattern-match the Option/Result to produce a clear expect message) so callers still panic on setup failure but the failure points (create_domain RPC, missing id, parse error) are explicit; update test_domain, create_domain call handling and the DomainId conversion accordingly while preserving the returned TestDomain { id, name }.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@crates/api-core/src/tests/site_explorer.rs`:
- Line 3239: Remove the redundant local import of the Forge trait by deleting
the line that says "use rpc::forge::forge_server::Forge;" in this test file; the
Forge trait is already imported at the module level, so remove the duplicate to
avoid unused/duplicate imports while leaving the module-level import intact.
- Around line 3292-3303: The test function
test_get_machine_position_info_no_endpoint is misnamed and miscommented because
create_managed_host actually creates an explored endpoint; rename the test
(e.g., test_get_machine_position_info_endpoint_without_position) and update the
leading comment to state that an explored endpoint exists but contains no
position data, and remove the redundant use rpc::forge::forge_server::Forge
import (duplicate of the previous test) so imports are not repeated; keep
existing variables like dpu_machine_id and the rest of the test logic unchanged.
In `@crates/site-explorer/tests/power_shelf.rs`:
- Around line 1319-1520: The test_power_shelf_state_history_multiple test is
prematurely short-circuited by the unconditional early return (if 1 == 1 {
return Ok(()); }), so the later state-history assertions never run; remove the
early-return block (the "if 1 == 1 { return Ok(()); }" stub) and either
implement the intended controller iterations and assertions (call
env.run_power_shelf_controller_iteration() the required number of times and then
run the assertions that follow) or, if not ready, mark the test with #[ignore]
and keep the rest of the test intact so the current assertions are either
executed or intentionally skipped.
- Around line 929-1069: The test_site_explorer_creates_power_shelf test is
short-circuited by the early return (if 1 == 1 { return Ok(()); }) which
prevents the final controller state assertions from running; either remove that
conditional return and call/run the controller iteration
(env.run_power_shelf_controller_iteration().await) and then re-enable the
subsequent assertions that verify state transitions, or mark the test with
#[ignore] and add a TODO explaining which state-machine wiring is required to
enable the controller assertions; update references in the test to
test_site_explorer_creates_power_shelf,
env.run_power_shelf_controller_iteration, and the post-return assertions
accordingly.
- Around line 1117-1247: The test_power_shelf_state_history test is being
short-circuited by the unconditional early return (if 1 == 1 { return Ok(());
}), so remove that guard and either (A) invoke the real controller iteration
(call env.run_power_shelf_controller_iteration().await; or the proper async
method that advances the power shelf state machine) and then run the remaining
state history assertions against db::state_history::for_object, or (B) if the
state machine isn't ready, mark the test with #[ignore] and add a brief comment
explaining why; ensure references to created_power_shelf.id / power_shelf_id and
the post-iteration assertions remain reachable.
In `@crates/test-harness/Cargo.toml`:
- Around line 18-23: The Cargo.toml for the crate named "carbide-test-harness"
is missing the required package description; open the [package] block in that
Cargo.toml and add a descriptive description field (e.g., description = "Short
one-line summary of the crate's purpose") so the package metadata includes a
clear description for the carbide-test-harness crate.
---
Nitpick comments:
In `@crates/site-explorer/tests/power_shelf.rs`:
- Around line 1305-1314: Remove the large commented-out assertion block that
builds state_versions from final_state (the commented lines referencing
state_versions, final_state, and the assert! about state_versions.len() > 1);
either delete these comment lines entirely if the verification is no longer
needed, or re-enable them by uncommenting and ensuring the test compiles and
uses state_versions/state_version properly inside the test function where
final_state is available.
- Around line 1087-1094: Remove the commented-out debug loop and the redundant
early-return by deleting the lines containing the commented
println!/env.run_power_shelf_controller_iteration() block and the `if 1 == 1 {
return Ok(()); }` statement so the test no longer contains dead commented code
or a second early return; search for the
`println!`/`env.run_power_shelf_controller_iteration().await` snippet and the
`if 1 == 1` block in the test and remove them, leaving the function flow as
originally intended with only the earlier return.
- Around line 254-366: In
test_site_explorer_power_shelf_discovery_with_static_ip remove the commented-out
call to explorer.run_single_iteration().await.unwrap(); (or replace it with a
one-line TODO/FIXME explaining why it's disabled) so there is no stray commented
code left in the test; locate the commented line referencing
explorer.run_single_iteration() and either delete it or add a concise TODO/FIXME
comment directly above the line documenting the reason for disabling the extra
iteration.
- Around line 1481-1491: The local binding `_history_by_ids` is unused after
calling `db::state_history::find_by_object_ids` in the test; either replace `let
_history_by_ids = ...` with `let _ =
db::state_history::find_by_object_ids(...).await?;` to explicitly discard the
result, or re-enable the commented assertions that reference `history_by_ids`
(and rename `_history_by_ids` to `history_by_ids`) so the query result is
actually asserted against; update the call site around
`db::state_history::find_by_object_ids`, the `_history_by_ids` binding, and the
commented lines that assert
`contains_key(&power_shelf1_id)`/`contains_key(&power_shelf2_id)` accordingly.
- Around line 1700-1720: The test currently compares JSON strings by stripping
spaces (history_entry.state.replace(" ", "") vs serde_json::to_string(&state)?),
which is brittle; instead parse the stored JSON and compare structurally —
either deserialize history_entry.state into the concrete type (e.g.,
PowerShelfControllerState) and assert equality with state, or deserialize both
into serde_json::Value via serde_json::from_str and compare the Value to
serde_json::to_value(&state). Update the assertions around history_entry.state
and found_entry.unwrap().state to perform these structural
deserializations/comparisons rather than string replacement.
- Around line 1279-1286: Remove the commented-out loop and the redundant
early-return block: delete the three commented lines calling
env.run_power_shelf_controller_iteration() and the subsequent if 1 == 1 { return
Ok(()); } so the test continues normally; look for the
env.run_power_shelf_controller_iteration() commented loop and the literal if 1
== 1 return Ok(()) and remove those lines together.
- Around line 127-1755: Several tests duplicate constructing
EndpointExplorationReport, running DHCP discovery to set FakePowerShelf.ip, and
seeding expected_power_shelf rows; extract helpers to reduce duplication: add
Env::create_basic_exploration_report(&self, serial: &str, model: &str) returning
EndpointExplorationReport and Env::setup_power_shelf_with_dhcp(&self,
power_shelf: &mut FakePowerShelf) -> Result<(), Box<dyn std::error::Error>> to
perform the DhcpDiscovery call and set power_shelf.ip, plus a helper
Env::seed_expected_power_shelf(&self, power_shelf: &FakePowerShelf) that wraps
db::expected_power_shelf::create; update tests like
test_site_explorer_power_shelf_discovery,
test_site_explorer_power_shelf_with_static_ip,
test_site_explorer_power_shelf_creation_limit, etc. to call these helpers and
replace the repeated EndpointExplorationReport literals with
create_basic_exploration_report calls.
- Around line 1513-1520: Delete the commented-out controller iteration block
(the three lines calling env.run_power_shelf_controller_iteration() that are
prefixed with //) so the test no longer contains dead commented code; keep the
surrounding TODO comment if desired and leave the remaining early return logic
unchanged—look for the env.run_power_shelf_controller_iteration() calls to
locate the commented lines to remove.
In `@crates/test-harness/src/lib.rs`:
- Around line 57-72: The test_domain function uses multiple chained .unwrap()
calls (notably the
.map(::carbide_uuid::domain::DomainId::try_from).unwrap().unwrap()) which hurts
readability; replace these with explicit, contextual failures by using expect()
with clear messages on the RPC/result unwraps and convert the map/unwrap/unwrap
into .map(|s| ::carbide_uuid::domain::DomainId::try_from(s).expect("failed to
parse DomainId from response")) (or alternatively pattern-match the
Option/Result to produce a clear expect message) so callers still panic on setup
failure but the failure points (create_domain RPC, missing id, parse error) are
explicit; update test_domain, create_domain call handling and the DomainId
conversion accordingly while preserving the returned TestDomain { id, name }.
In `@crates/test-harness/src/network/controller.rs`:
- Around line 68-110: Add a brief explanatory comment above the two
run_single_iteration().await calls in both create_underlay_segment() and
create_admin_segment() that documents why exactly two state controller
iterations are required for provisioning (e.g., first iteration to enqueue/apply
the network creation and a second to reconcile and persist resulting
state/allocate IPs), mention any assumptions (such as async job scheduling or
reconciliation semantics), and reference run_single_iteration() so future
readers understand this “magic number” is intentional and tied to the
controller’s two-step provisioning lifecycle.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 4790ace5-95d1-4ea4-a758-15de3d1d2bfe
⛔ Files ignored due to path filters (1)
Cargo.lockis excluded by!**/*.lock
📒 Files selected for processing (15)
crates/api-core/src/test_support/endpoint_explorer.rscrates/api-core/src/test_support/mod.rscrates/api-core/src/tests/common/api_fixtures/mod.rscrates/api-core/src/tests/mod.rscrates/api-core/src/tests/site_explorer.rscrates/site-explorer/Cargo.tomlcrates/site-explorer/tests/power_shelf.rscrates/test-harness/Cargo.tomlcrates/test-harness/src/builder.rscrates/test-harness/src/dns.rscrates/test-harness/src/lib.rscrates/test-harness/src/network/controller.rscrates/test-harness/src/network/mod.rscrates/test-harness/src/network/segment.rscrates/test-harness/src/prelude.rs
poroh
force-pushed
the
tests-refactoring-p3
branch
from
b5a448b to
fa2e40f
Compare
wminckler
approved these changes
Jun 5, 2026
poroh
force-pushed
the
tests-refactoring-p3
branch
from
fa2e40f to
c0cbea2
Compare
|
🌿 Preview your docs: https://nvidia-preview-pull-request-2250.docs.buildwithfern.com/infra-controller |
wminckler
approved these changes
Jun 5, 2026
Move power shelf integration coverage out of api-core site explorer tests into the site-explorer crate, and introduce a shared test harness foundation for cross-crate test setup. Signed-off-by: Dmitry Porokh <dporokh@nvidia.com>
poroh
force-pushed
the
tests-refactoring-p3
branch
from
c0cbea2 to
2dedd22
Compare
chet
approved these changes
Jun 5, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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.
Description
Move power shelf integration coverage out of api-core site explorer tests into the site-explorer crate, and introduce a shared test harness foundation for cross-crate test setup.
Type of Change
Related Issues (Optional)
#2001
Breaking Changes
Testing
Additional Notes