Test/5329 mana echoes shares type regression

[andriypolanski]

Summary

Mana Echoes ("Whenever a creature enters, you may add an amount of {C} equal to the number of creatures you control that share a creature type with it.") was reported as producing zero mana. Investigation on current main shows the full stack path already works when:

1. Parser binds SharesQuality to TargetFilter::TriggeringSource (not ParentTarget).
2. Optional resume stashes pending_optional_trigger_event and restores it on DecideOptionalEffect { accept: true }.
3. all_creature_types is populated (deck load / test seed) so CreatureType sharing resolves.
   No engine logic changes were required. This PR adds regression coverage so a future regression in parser binding, optional trigger-event plumbing, or shares-type counting is caught immediately.

Root cause of the original report (triage)

Layer	Finding
Parser AST	SharesQuality { reference: TriggeringSource } on ObjectCount filter
Stack + optional	Trigger event round-trips through optional pause; mana lands on accept
Quantity / filter	object_count_matching_ids returns correct count with current_trigger_event set
Integration harness foot-gun	Test helper cloned base_card_types before subtypes were pushed, which broke subtype-based matching in some paths

Changes

Integration tests (crates/engine/tests/integration/issue_5329_mana_echoes.rs)

• Parse assertion: shares-type reference is TriggeringSource.
• Runtime positive: three Goblins on board + Goblin ETB → 3 colorless after accepting optional trigger.
• Runtime zero case: subtypeless entering creature + Goblin on board → 0 colorless (no shared creature type).
• Helper fixes: priority loop passes until OptionalEffectChoice (no pass_both_players fallback); enter_creature_from_hand syncs base_card_types after subtypes.

Unit tests

• filter.rs: game_scenario_mana_echoes_shares_type_count_with_triggering_source, mana_echoes_optional_stack_pause_preserves_shares_type_count
• Existing: object_count_shares_creature_type_with_triggering_source_for_mana_echoes, colorless_production_counts_creatures_sharing_type_with_triggering_source

Registration

• mod issue_5329_mana_echoes; in crates/engine/tests/integration/main.rs

Verification

cargo fmt --all
cargo test -p engine mana_echoes
cargo test -p engine game_scenario_mana_echoes mana_echoes_optional_stack object_count_shares_creature colorless_production_counts

Test plan

• cargo test -p engine mana_echoes — 3 integration + lib unit tests green
• Manual: Mana Echoes on battlefield, accept optional on creature ETB, mana pool increases by matching creature count
• Manual: decline optional — no mana added

Closes #5329

[gemini-code-assist]

Warning

You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again!

[github-actions]

Parse changes introduced by this PR

✓ No card-parse changes detected.

[matthewevans]

Thanks for the #5329 regression — the shape (engine fix in mana.rs/filter.rs + a registered integration test) is the right one. Two mechanical blockers before I can review the logic:

1. One-line compile error. crates/engine/src/game/effects/mana.rs:2351 imports use crate::types::identifiers::{CardId, ObjectId}; but ObjectId is unused. Under the workspace's -D warnings, an unused import is a hard compile error — that's why all three Rust gates (lint, both test shards) fail with could not compile engine (lib test) due to 1 previous error, not the test itself. Drop ObjectId from that use (keep CardId) and the build clears.

2. Branch is BEHIND main. Please rebase onto origin/main (or use the update-branch button) so CI runs against current main.

Once it's green I'll do the full architecture + rules pass on the shares-a-creature-type counting in filter.rs (CR 202.2 / the Mana Echoes trigger). Holding until then.

[matthewevans]

Re-reviewed on head 480994ae: both mechanical blockers from the prior head are resolved — the -D warnings compile abort is gone (mana.rs import is CardId-only) and mod issue_5329_mana_echoes; is registered in integration/main.rs (not inert). This is clean test-only regression coverage (534 additions, 0 removals; all engine-file hunks under mod tests; parse_diff reports no card-parse changes), locking in the #5329 fix that already works on main. Coverage is discriminating: positive (3 Goblins → 3 colorless), negative (subtypeless entering creature → 0), the optional-stack-pause preservation case, and SharesQuality bound to TriggeringSource (the actual regression axis, not ParentTarget). Full CI rollup is settled green. Approving and enqueuing. Thanks for the thorough regression net.