Files
shard-memory/projects/SOR.md
T

13 KiB
Raw Blame History

SOR Project Working Memory

Mission

Help Jason ship SOR with low-friction, safe autonomy.

Operating Rhythm

  • Keep an updated short backlog.
  • Surface blockers quickly.
  • Prioritize next 1-3 concrete actions.
  • Log decisions and rationale.

Current Focus

  • Build momentum on SOR implementation and coordination.

Next Review Questions

  • What is the highest-impact next task?
  • What is blocked and by what dependency?
  • What can be delegated to sub-agents safely?

TODO (tracking)

  • Evaluate and optionally pilot obra/superpowers for Prime-only coding workflows; keep Shard focused on infrastructure/deploy operations.

  • Add duration semantics flags for status/lock effects: choose whether ForNTurns counts total turns or owner-only turns. Current gameplay expectation: “cannot X for 2 turns” should cover the target’s next two turns (not just one full round).

  • Improve client-facing internal error payload behavior by environment: expose actionable diagnostics in dev/test (exception message/type + traceId), keep production responses sanitized.

  • 2026-03-11 16:59 UTC: Repo clean at commit da6899b; no blockers.

  • 2026-03-11 21:42Z: No priority change. Repo clean; continue current track.

Heartbeat Notes

  • 2026-03-11 23:12 UTC: Repo clean; no blockers; latest da6899b. Priorities unchanged.

  • [2026-03-12 03:27 UTC] Heartbeat: repo clean, no immediate blockers; continue current roadmap.

  • Heartbeat 2026-03-12 03:57 UTC: repo clean at da6899b; no immediate blockers.

  • 2026-03-18 17:27 UTC: Repo clean at 6645231; no blockers. Priority remains execution momentum on current implementation track.

  • 2026-03-19 15:57 UTC: Repo clean at cf05bac (Phase 6: GameGetLobbyHandler + CustomLobbyDetailDto); no blockers. Priority unchanged: continue implementation momentum.

  • 2026-03-19 19:57 UTC: Repo clean at e9476a3 (Fix ReturnToHandEffect: revoke passives + skip target conditions at activation); no blockers. Priority unchanged: maintain implementation momentum.

  • 2026-03-19 23:57 UTC: Repo clean at 7d572ae (Standardize active ability execution between unit and card paths); no blockers. Priority unchanged: maintain implementation momentum.

  • 2026-03-19 20:27 UTC: Local WIP patch detected in SOR (AbilitySystem.cs) for Repentance/Instinct effect execution compatibility. Awaiting Jason decision to commit/deploy.

  • 2026-03-19 21:04 UTC: Landed schema-unification commit 9f23e56 (canonical AbilitiesJson + EffectsJson compatibility + backfill migration). Priority: rollout with migration.

  • 2026-03-19 21:34 UTC: Canonical AbilitiesJson rollout landed (7a90e2b) with compatibility fallback + backfill migration applied. Next: redeploy SOR.Game and SOR.Admin and monitor Repentance/Instinct effect execution in live playtests.

  • 2026-03-19 23:04 UTC: New server fix commit 7ca86ce addresses Repentance lock enforcement gap (movement + ability suppression); next high-value step is push and redeploy sor-game.

  • 2026-03-20 00:27 UTC: sor-game includes standardized active ability execution semantics (7d572ae); next high-value step is redeploy and validate Repentance lock card behavior in live match.

  • 2026-03-20 14:48 UTC: Repo head advanced to 2e3a5e4 (Enable decks to have multiple Tier 1 ArchSpirits). No blocker detected; priority remains gameplay/deck-flow validation in live playtests.

  • 2026-03-20 15:48 UTC: Repo clean at 6c6a66c (Deck editor client cleanup). No blockers detected; priority unchanged.

  • 2026-03-25 18:57 UTC: Repo clean at 85e9b63 (Account mailbox UX now wired to notifications and thread actions). No blockers; priority remains implementation momentum with live validation.

  • 2026-03-26 14:57 UTC: Repo clean at 84284be (fix active ability target gating for server-resolved ownership, including ExactSelf). No blockers; next high-value step is live-play validation of self-target activated abilities to confirm no regressions for explicit-target abilities.

  • 2026-03-26 15:57 UTC: Repo clean at 84284be (Fix active ability target gating for server-resolved ownership scopes). No blockers; next high-value step remains live-play validation for self-target activated abilities + regression check for explicit-target abilities.

  • 2026-03-26 19:57 UTC: Repo clean at cf473d8 (Fix board shapes, spectator-stuck, and display name on reconnect). No blockers; next high-value step is end-to-end reconnect validation across player/spectator paths including board shape rendering and display-name correctness.

  • 2026-03-26 22:57 UTC: Repo clean at 3c8614c (sparse-board adjacency fix for non-rectangular shapes). No blockers observed; priority remains reconnect/board-shape validation in live flows.

  • 2026-03-26 23:57 UTC: Repo clean at 4737d33 (Fix triangular board compile error (nested Math.Max)). No blockers; next high-value step is reconnect + board-shape regression validation with explicit triangular layout coverage.

  • 2026-03-27 00:27 UTC: Repo clean at 4737d33; no blockers. Priority update: investigate/fix 3+ player custom-lobby state propagation so additional participants receive live lobby updates and ready-state sync regardless of join/ready order.

  • 2026-03-27 00:57 UTC: Landed 32aa816 (3+ player custom lobby state/ready sync). No blockers. Priority: regression test multiplayer lobby participant/ready propagation across out-of-order joins/readies; then polish stale-ready UX for SESSION_NOT_INITIALIZING.

  • 2026-03-27 01:57 UTC: Landed a84313e (3+ lobby ready UX clarity). No blockers. Priority: validate full 3+ participant ordering matrix in live flow and confirm no stale-ready regressions.

  • 2026-03-27 02:27 UTC: Landed c2e19a3 to undo game.ready state regression; repo clean. Priority: validate join/start pipeline end-to-end with 3+ lobbies.

  • 2026-03-27 03:27 UTC: Landed 775b30d to make lobby leave idempotent on stale IDs; priority is post-restart create/join validation and log watch for silent join failures.

  • 2026-03-27 03:57 UTC: Repo clean at 775b30d (Make custom lobby leave idempotent for stale lobby IDs). No blockers; next high-value step is validating leave/rejoin behavior under stale lobby IDs and checking for spectator/reconnect regressions.

  • 2026-03-27 03:57 UTC: Landed 775b30d to harden stale lobby leave handling (server+client). Priority: verify join path reliability immediately after stale-state scenarios.

  • 2026-03-27 11:57 UTC: Repo clean at d257583 ("missing comma"); no blockers. Priority update: perform quick live validation of the updated response/payload path to ensure parser/consumer compatibility and no regressions.

  • 2026-03-27 12:57 UTC: Repo clean at cb024c7 (game.state_delta cutover to account-scoped multiplayer schema; client/server delta apply path updated). No blockers detected. Priority: live-validate 3+ player + spectator state convergence, especially ArchSpirit visibility timing and delta-version continuity.

  • 2026-03-27 13:10 UTC: Repo clean at bf9e059 (Fix delta arch-spawn propagation + viewer private lookup robustness). Remaining validation priority: 3+ player turn-by-turn state convergence from creator perspective, plus spectator continuity. Known local env limitation: client build not runnable here (docker/godot unavailable).

  • 2026-03-27 13:33 UTC: Repo clean at de3c62e (client typo fix). Build validation context improved: local Godot 4.5+templates now available for HTML5 export checks in this environment; Docker daemon remains unavailable due to no-elevation runtime.

  • 2026-03-27 13:45 UTC: Repo clean at 6d20af8 (active-session rejoin detection now includes additional players). Validation status: both local build checks passed this cycle (server dotnet build + client Godot HTML5 export). Next priority: live-validate reconnect prompt behavior for seats 3+ and ensure concede path semantics remain correct.

  • 2026-03-27 14:07 UTC: Repo clean at 8736217 (rejoin gating for eliminated players, force-concede guard, board presence cleanup on elimination/concede). Both local build checks passed this cycle. Next validation target: multiplayer concede/elimination reconnect matrix + idle-timeout abuse regression.

  • 2026-03-27 14:17 UTC: Repo clean at e5a9bde (client nil-guard for archSpirit after concede/elimination). Both local build checks passed this cycle. Next validation target: concede→relaunch→spectate path in multiplayer to confirm crash is gone.

  • 2026-03-27 14:25 UTC: Repo clean at a656c70 (spectator null-array getter hardening). Both local build checks passed this cycle. Next validation target: conceded-player spectate path and any remaining null-typed getter crashes in live reconnect flows.

  • 2026-03-27 14:55 UTC: Repo clean at a656c70 (nullable spectator/player-private array handling in client Game getters). Priority unchanged: validate concede→relaunch→spectate in multiplayer and monitor for further nil-hand/field regressions.

  • 2026-03-27 15:25 UTC: Current priority is completing/validating phase-1 ability additions across SOR.Game + SOR.Admin builder. Repo currently dirty with 7 files edited; next step is finish builder wiring, run server/client builds, then commit+push.

  • 2026-03-27 15:41 UTC: Repo clean at 9e4fbd1 with phase-1 ability primitives shipped and build-validated. High-value next step: live-author one or two cards using TriggerLimitPerTurn + RepositionUnit + move-affinity conditions to validate authoring ergonomics in admin builder.

  • 2026-03-27 15:57 UTC: Repo currently dirty with local WIP in AbilityBuilder + ability primitives (9e4fbd1 at HEAD). Priority update: convert current WIP into a coherent validated commit (or cleanly park partial changes) to avoid drift and reduce regression risk.

  • 2026-03-27 16:27 UTC: Phase 2 ability work remains in-progress; priority unchanged (finish implementation, run server/admin/client builds, commit coherent changeset).

  • 2026-03-27 16:57 UTC: Phase 2 implementation is still in-progress (prevention/replacement + sequence-scoped duration + builder support). Repo remains intentionally dirty pending full compile/export validation and coherent commit.

  • 2026-03-27 17:44 UTC: Repo clean at e887a8c with phase-2 shipped and validated. Suggested next focus: phase-3 implementation and a small authored-card validation pass covering BeforeDamageApplied/BeforeUnitDeath/PreventNextDamage semantics.

  • 2026-03-27 17:55 UTC: Phase 3 implementation progressing (relational adjacency filters/conditions, movement-state condition, source-qualified damage trigger, tile-control condition). Server game build green; pending admin build + client export before commit.

  • 2026-03-27 17:56 UTC: Phase 3 status improved: both server builds now passing after neighbour helper fix. Pending step before landing changes is client HTML5 export validation, then commit/push.

  • 2026-03-27 17:58 UTC: Repo clean at 8f0ed9d; Phases 1-3 complete and build-validated. Suggested next milestone: authored-card regression pack covering new trigger/condition/filter semantics and edge interactions (prevent-next-damage with sequence duration, source filter matching, tile-control comparators).

  • 2026-03-27 18:28 UTC: Monitoring check only; repo remains clean at 8f0ed9d. No new blockers/decisions since last update.

  • 2026-03-27 18:58 UTC: Steady state. Repo clean on phase-3 commit; no new deltas since last check.

  • 2026-03-27 19:28 UTC: Stable post-Phase-3 state, repo clean. Next useful activity remains authored-card regression matrix for newly added ability primitives.

  • 2026-03-27 19:57 UTC: Repo clean at 8f0ed9d (phase-3 adjacency/movement-state/tile-control primitives). No blockers. Priority update: run focused gameplay/regression validation around trigger ordering and control-state interactions before stacking additional primitives.

  • 2026-03-28 01:57 UTC: Repo not clean (OVERALL.md modified) due to newly captured post-Phase-3 ability backlog (tile markers/equality, ChooseOne runtime hardening, exact reposition picks, destination-affinity channel-cost mods). Suggest commit this doc-only delta for clean baseline before next implementation.

  • 2026-03-28 02:27 UTC: No new SOR changes since last heartbeat; pending OVERALL.md backlog update remains uncommitted.

  • 2026-03-28 03:57 UTC: Working tree not clean (OVERALL.md modified) while HEAD stays 8f0ed9d. Priority update: reconcile documentation WIP (commit or clean) so status checks clearly reflect code-state risk.

  • 2026-03-28 17:27 UTC: Heartbeat check — no state change. Repo still has local doc-only delta (OVERALL.md), HEAD 8f0ed9d; no blockers or priority changes.

  • 2026-03-28 20:27 UTC: Heartbeat check shows no priority change. SOR remains at 8f0ed9d with local doc-only OVERALL.md modification pending commit/cleanup.

  • 2026-03-28 20:57 UTC: Repo now at 2e3e284 (ability targeting gap fixes across multi-target/UpTo/ChooseOne/LinearOnly/LOS). Working tree remains dirty with local OVERALL.md doc delta. Priority update: run targeted gameplay regression around targeting legality/resolution paths, then reconcile doc-only drift for a clean baseline.

  • 2026-03-28 22:57 UTC: No code-state change from prior check (HEAD 2e3e284); local OVERALL.md remains modified for backlog reconciliation. Priority unchanged: run focused targeting/regression validation, then commit or clean doc drift.