[milindsrivastava1997]

Why

Part of #1096 (Stage 1 layering). Pass 1 runs before Pass 2, but Pass 1's production code imported crate::pass2::memo: the code that fills Stage 0's memo groups with Pass 1 candidates lived under pass2/.

What

• pass2/memo/groups.rs → pass1/memo_population.rs (try_populate_pass1, NodeToStructuralIdCache).
• CandidatePopulationError moves into the same file; pass2/memo/candidates.rs is deleted.
• New tests/pass_layering.rs: fails if Pass 1's production code references pass2. Test modules may still use Pass 2 for fixtures. No exceptions: Remove unnecessary deadlines and caps in the group selection MILP #1185 part 2 (refactor(planner): plan without deadlines over every tumbling-window width (#1185, part 2) #1198) deleted check_generation_deadline, the last shared item.
• Two stale paths in planner-layering.md.

How

A pure move plus import-path updates. Public paths are unchanged: CandidatePopulationError (crate root) and test_support::try_populate_pass1 keep their names.

Pass 2 → Pass 1 is still allowed: PerComputationCandidateError is effectively Stage 1's shared error, and asap_aware_cse.rs uses Pass 1 helpers. Removing that is #1191.

Key code interfaces added / modified

None (crate-private module moves only).

Major algorithm changes

None.

Before this PR

pass1/logical_memo_candidates.rs:  use crate::pass2::memo::candidates::CandidatePopulationError;
pass1/logical_memo_candidates.rs:  use crate::pass2::memo::groups::try_populate_pass1;
pass1/per_computation_candidates.rs:  use crate::pass2::memo::NodeToStructuralIdCache;

After this PR

No pass2 references in Pass 1's production code, and cargo test -p asap-logical-asap-optimization --test pass_layering fails on any new one.

Snapshot changes

None.

Verification

• cargo test -p asap-logical-asap-optimization --all-targets: 207 passed, 1 ignored; pass_layering 1 passed. After rebasing onto refactor(planner): plan without deadlines over every tumbling-window width (#1185, part 2) #1198: lib 201 passed, pass_layering 1 passed.
• pass_layering fails when a use crate::pass2::… line is planted in pass1/realization.rs (checked, then reverted).
• cargo check -p asap-physical-asap-optimization -p asap-planner -p asap-integration-tests --all-targets (after the rebase: -p asap-planner -p asap-devtools --tests), cargo clippy -p asap-logical-asap-optimization --all-targets, cargo fmt --all --check: clean.
• Pre-commit hooks were skipped locally (memory-heavy on this machine); relying on CI.

🤖 Generated with Claude Code

[milindsrivastava1997]

Automated code review (Claude Code); findings are unverified.

• crates/logical-asap-optimization/tests/pass_layering.rs:290: The guard only matches the substring pass2, so Pass 1 can still import Pass 2 types through crate-root re-exports (crate::Pass2RuleError, crate::TumblingWindowWidthDomain, crate::Pass2RoundState). E.g. use crate::TumblingWindowWidthDomain; in pass1/ passes the test; this PR's own tests in pass1/logical_memo_candidates.rs already use that route.
• crates/logical-asap-optimization/tests/pass_layering.rs:288: Splitting production from test code on the literal "#[cfg(test)]\nmod tests" misses CRLF checkouts (core.autocrlf=true) and any other test-module layout. Then test code counts as production, and a_summary_pass2_may_reuse_for_another_aggregate_is_kept in per_computation_candidates.rs trips a false violation.