Team 5 Sprint 2 verification and delivery status

Snapshot: 17 September 2026. This page records evidence for named revisions, not a claim that every current branch or manual scenario has passed.

Delivery baselines

Baseline	State
main 9625e56	Checked main version, including main's room/boss/weapon work
remote items 9eb2186	Pushed Team 5 integration: direct 7/8/9/0 use, top-right HUD, Shield countdown, consumable effects and production drops
local items 4b5e7c2	Additional Speed/Strength countdowns; not pushed at this snapshot
demo/local-health-lock	Local-only test aid; excluded from production items and all normal-death validation claims

PRs #162 and #169 were closed for consolidation; their original input and HUD histories are retained in items. The final items-to-main PR and its review/merge remain separate delivery steps.

Recorded automated evidence

• The integration preparing 9eb2186 ran ./gradlew --offline spotlessCheck core:test: 946 tests, 0 failures, 0 errors, 0 skipped; formatting passed.
• Earlier main-sync implementation a110647 passed spotlessCheck core:test core:javadoc, with 946 tests. JavaDoc had existing warnings; it was not a warning-free run.
• No new test run was performed just to edit this Wiki. The 946 result must not be relabelled as validation of the later local countdown commit.
• Versioned evidence record contains the earlier run history and coverage. Statements saying “local only” there are historical: 9eb2186 has since been pushed.

Reproduce

From the checked-out branch's source directory:

./gradlew spotlessCheck core:test core:javadoc
./gradlew desktop:run

Use JDK 21 through the repository's configured Gradle toolchain. Add --offline only when the required dependencies are cached. Unit XML results are under core/build/test-results/test; HTML results under core/build/reports/tests/test; coverage is under core/build/reports/jacoco/test.

What the regression suite covers

Area	Evidence in remote items
Configurable Gold/probability, all four choices, seeded decisions	EnemyDropPolicyTest / ItemDropSpecTest
Drop registration, captured position, repeated events, room disposal	EnemyManagerComponentTest
Real enemy reward → pickup → HUD → keyboard → health/count	enemyRewardFlowsThroughPickupHudKeyboardAndEffect
Stock and exactly-once use, empty/full-health/invalid/dead requests	InventoryComponentTest / ConsumableUseIntegrationTest
Shield actual damage blocking and expiry	ConsumableUseIntegrationTest
Strength refresh, Charm coexistence; Speed/mist independent lifetime	ConsumableUseIntegrationTest
HUD quantity refresh, replacement cleanup, Shield progress	Team5CombatHudIntegrationTest
Team 4 bottom hotbar and Team 5 top-right HUD separation	HotbarDisplayTest at 1920x1080, 1280x720 and 906x600

Browse versioned tests.

Manual evidence still to record

Yuezhou reported that effects were visible and working during a local health-lock playtest. That is feedback, not a recorded final-main demonstration or a normal-death test. No screenshot/recording is attached here as evidence of those missing checks.

• Record the exact production items/main commit used for the final demonstration.
• Show enemy rewards, E pickup, Gold/stock updates and each of the four consumables.
• Show quantity decrement once, repeated-use refresh and effect expiry.
• Check room transitions preserve inventory and avoid duplicate/misplaced HUD actors.
• Check normal damage/death on a build without the health-lock demo.
• Attach the final gameplay screenshot or recording and note any remaining limitations.

Known limits

Full inventory-book data/management is not integrated in this baseline; the current shared HUD does not display Charm count. Save-across-sessions, Shop and special boss-phase rewards are not part of this Team 5 delivery. Historical Sprint 1 screenshots and test totals remain historical.

Ownership and AI assistance

The original contributions of Sumith, Aarash, Devendera and Jeremy remain in Git history. Yuezhou leads the consolidated integration and delivery. Each member retains their original contribution; integration changes and verification additions do not reassign authorship. Codex assisted Yuezhou with code comparison and this documentation update. No new test or manual recording is claimed by the documentation authoring step.