rom24-quickmud-python 2.7.0__tar.gz → 2.8.21__tar.gz

{rom24_quickmud_python-2.7.0 → rom24_quickmud_python-2.8.21}/CHANGELOG.md

@@ -7,6 +7,291 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.8.21]
+
+### Changed
+- **README status refresh**: updated the public project status, audit coverage, test counts, and active-focus guidance so the README now matches the audited trackers and current session status.
+- **Release wrap-up**: recertified the published repo state around the current fully green parity program (`4560 passed, 4 skipped`) and the completed 40/40 audit-bound ROM C file coverage.
+
+## [2.8.20]
+
+### Changed
+- **Stability investigation recertified clean**: executed the queued descriptor/session reproduction subset and a full-suite rerun with no reproducing leak (`36 passed, 36 deselected` in the networking subset; full suite still `4560 passed, 4 skipped`).
+- **Session guidance tightened**: `docs/sessions/SESSION_STATUS.md` now records that descriptor/session cleanup work should not proceed without a deterministic reproduction, and points the next agent at the standing investigation plan only if a leak reappears.
+
+## [2.8.19]
+
+### Added
+- **Stability / invariant hardening handoff plan**: added a concrete next-agent execution plan for descriptor/session global-state isolation, targeting `descriptor_list`, `character_registry`, telnet/session teardown, and any newly discovered cross-file invariant.
+
+### Changed
+- **Session handoff updated**: `docs/sessions/SESSION_STATUS.md` now points at the stability / invariant hardening queue, with a dedicated handoff summary and explicit first verification commands for the next agent.
+
+## [2.8.18]
+
+### Added
+- **Descriptor-backed `wiznet` regression coverage**: `tests/test_wiznet.py` now registers ROM-style playing descriptors for listeners and includes a stale-descriptor regression, so the suite exercises the production descriptor path instead of relying on the character-registry fallback.
+
+### Fixed
+- **`wiznet` descriptor-path immortal gate parity**: `mud/wiznet.py` now mirrors ROM `src/act_wiz.c:171-194` on the real descriptor path by rejecting NPC and non-immortal recipients before checking wiznet flags.
+
+### Changed
+- **Full-suite recertification**: the suite now reruns clean at `4560 passed, 4 skipped`.
+
+## [2.8.17]
+
+### Added
+- **`BOARD-014` AFK parity coverage**: added ROM-backed integration tests for note-editor AFK ownership, including `note write`, `note send`, and the new `note forget` cancel path.
+
+### Fixed
+- **`BOARD-014` note-editor AFK parity**: `mud/commands/notes.py` and `mud/models/board.py` now mirror ROM `src/board.c` by setting AFK when note composition begins, clearing only note-owned AFK when posting or forgetting a note, and preserving manually enabled AFK.
+
+### Changed
+- **`board.c` audit closure**: `docs/parity/BOARD_C_AUDIT.md` and `docs/parity/ROM_C_SUBSYSTEM_AUDIT_TRACKER.md` now record `board.c` as fully closed with `BOARD-014` no longer deferred.
+- **Full-suite recertification**: the suite now reruns clean at `4559 passed, 4 skipped`.
+
+## [2.8.16]
+
+### Fixed
+- **`MUSIC-005` global-channel recipient parity**: `mud/music/__init__.py` now mirrors ROM `src/music.c:88-97` by iterating the ROM-style descriptor list, requiring `CON_PLAYING`, checking `COMM_NOMUSIC` / `COMM_QUIET` on `descriptor.original` when switched, and delivering the line to the active descriptor character.
+- **`MUSIC-006` jukebox visibility parity**: jukebox room output now resolves `$p` per viewer, so recipients who cannot see the jukebox get ROM-style `"something"` fallback instead of a leaked short description. Implemented in `mud/music/__init__.py` with object visibility fallback in `mud/utils/act.py`.
+
+### Changed
+- **`music.c` audit closure**: `docs/parity/MUSIC_C_AUDIT.md` and `docs/parity/ROM_C_SUBSYSTEM_AUDIT_TRACKER.md` now record `music.c` at 100% with all six gaps closed.
+
+## [2.8.15]
+
+### Changed
+- **Parity audit reconciliation**: refreshed stale `nanny.c` audit headers and the historical `save.c` tracker narrative so the docs match the verified codebase and enforced invariants (`INV-003`, `INV-008`).
+
+## [2.8.14]
+
+### Added
+- **Deterministic ROM skill-improvement coverage**: `tests/integration/test_skills_integration.py` now enforces `check_improve()` on the live combat path, including the ROM rule that combat learning can continue above class adept up to 100.
+- **Executable money and spell-affect integration coverage**: replaced stale skipped placeholders with live tests for drop-money consolidation, cursed no-remove equipment, poison damage-over-time, and plague contagion spread.
+
+### Fixed
+- **`check_improve()` adept-cap parity**: `mud/skills/registry.py` now mirrors ROM `src/skills.c` by allowing post-use learning through 100 instead of stopping at the class adept cap.
+- **Plague contagion parity**: `mud/game_loop.py` now applies a full `AffectData` record to newly infected victims during `char_update()`, matching ROM `src/update.c:839-840`.
+- **Kick ROM parity test stability**: `tests/test_skill_combat_rom_parity.py` now isolates `check_improve()` from the kick damage-roll assertions so the test verifies the ROM damage call instead of incidental follow-up RNG.
+
+### Changed
+- **Integration coverage tracker reconciliation**: the skills and spell-affects slices now record their real enforced status, and the stale info-command / equipment / group-combat historical narratives were collapsed to canonical current-state notes.
+- **Full-suite recertification**: the suite now reruns clean at `4553 passed, 4 skipped`.
+
+## [2.8.13]
+
+### Added
+- **ROM wear-off suppression enforcement**: added targeted regression coverage for `char_update()` and `obj_update()` so consecutive same-type affect expirations now stay locked to ROM `src/update.c` semantics, including single-message suppression for object wear-off.
+- **Deterministic `mobile_update()` wander-gate enforcement**: added isolated integration coverage for `ACT_STAY_AREA`, `ACT_OUTDOORS`, and `ACT_INDOORS` movement restrictions using synthetic room topology instead of boot-world assumptions.
+
+### Fixed
+- **`GL-010` character affect wear-off parity**: `mud/affects/engine.py` now preserves merged spell effects while any same-type `AffectData` remains active, matching ROM `src/update.c:762-786`.
+- **`GL-017` object affect wear-off parity**: `mud/game_loop.py` now suppresses duplicate same-type object wear-off messages exactly once per tick, matching ROM `src/update.c:939-957`.
+
+### Changed
+- **Parity tracker reconciliation**: cross-file invariant and integration-coverage trackers now record the enforced wander-gate and wear-off parity slices, and the full suite is recertified at `4547 passed, 10 skipped`.
+
+## [2.8.12]
+
+### Added
+- **Canonical ROM `weapon_table`**: added `/mud/models/weapon_table.py` as the shared source for weapon class name, school weapon vnum, weapon type, and linked proficiency mapping, then wired combat, character-load skill seeding, and creation-time weapon handling to it.
+- **`NANNY-010` reconnect descriptor sweep**: `mud/net/connection.py` now mirrors ROM `src/nanny.c:307-352` by sweeping duplicate descriptors on break-connect reconnect, including the switched-immortal `original->name` branch.
+
+### Fixed
+- **`COMM-005` duplicate-newbie parity**: new-character name validation now closes matching non-playing duplicate descriptors and emits ROM's `Double newbie alert` wiznet notice.
+- **`FLAG-002` settable-bit preservation**: `do_flag` now preserves ROM `settable=FALSE` bits across the `=` operator for `act`, `plr`, and `comm` fields.
+- **Test determinism and suite stability**: stabilized wiznet and advancement integration tests by isolating shared descriptor-list state and replacing wall-clock ROM RNG seeding with a fixed Mitchell-Moore seed.
+
+### Changed
+- **Parity tracker reconciliation**: refreshed stale audit/tracker rows for `act_obj.c`, `act_wiz.c`, `scan.c`, `special.c`, `healer.c`, `const/tables/lookup`, `fight.c`, and `comm.c` so docs match the verified codebase.
+- **Full-suite recertification**: the suite now reruns clean at `4535 passed, 11 skipped`.
+
+## [2.8.11]
+
+### Fixed
+- **`log` command trust parity**: restored the dispatcher registration for `log` to ROM `src/interp.c` level `L1` (`MAX_LEVEL - 1`) and updated the admin-logging tests to use a ROM-valid immortal trust level instead of `LEVEL_HERO`.
+
+## [2.8.10]
+
+### Fixed
+- **Legacy character-row compatibility**: `mud/db/migrations.py` now backfills the DB-canonical `characters` columns that `load_character()` queries, so older SQLite databases with only the pre-2.8.0 schema can still load characters after migration. `mud/models/character.py:from_orm` and `mud/account/account_service.py:create_character` also keep persistent `points` aligned with ROM creation-point state.
+
+## [2.8.9]
+
+### Fixed
+- **Command parser test parity**: leading punctuation commands now follow ROM `src/interp.c` single-character tokenization, restoring apostrophe-as-`say` handling. Updated scripted-session/command tests to match current ROM-faithful transcript ordering and scan output, and replaced a stale non-takeable school-sword pickup assumption with an explicit takeable test object.
+
+## [2.8.8]
+
+### Fixed
+- **Admin `log` command parity**: dispatcher trust for `log` now matches ROM `src/interp.c`/`src/act_wiz.c`, `cmd_log` keeps the persisted `PLR_LOG` bit and `log_commands` boolean in sync, and admin player lookup now resolves through the live runtime character registry so `log <player>` works reliably.
+
+## [2.8.7]
+
+### Fixed
+- **hedit RecursionError**: `_interpret_hedit` fallthrough on unknown commands now returns `"Unknown help editor command: <cmd>\n\r"` so `_should_fallback_from_olc` routes to the normal command table instead of re-entering hedit via `process_command`. Mirrors ROM `src/hedit.c:258` `interpret(ch, arg)` without the re-entry loop.
+
+### Changed
+- **tests/test_builder_hedit.py rewrite**: All 19 stale tests replaced with 23 ROM-parity tests that verify actual `cmd_hedit`/`cmd_hesave` behavior (ROM-exact strings: `"HEdit: There is no default help to edit.\n\r"`, `"Ok.\n\r"`, `"Keyword : [...]"`, `"Level   : [...]"`, etc.). Covers entry-point guard, `new` subcommand, session open/show/keyword/level/text/done, session-lost recovery, unknown-command non-recursion, and full `hesave` workflow.
+
+## [2.8.6]
+
+### Fixed
+- **FIGHT-002 safe-room parity**: `mud/combat/engine.py:apply_damage` now mirrors ROM `src/fight.c:725-733` by re-checking `is_safe(ch, victim)` before any fighting-state or HP mutation. This stops melee damage, weapon procs, and other `apply_damage()`-backed attack paths from landing after combatants are moved into a safe room mid-fight. Regression coverage: `tests/integration/test_fight_c_safe_room_damage_gate.py`.
+
+## [2.8.5]
+
+### Fixed
+- **FIGHT-001 combat-entry parity**: `mud/commands/combat.py:do_kill` now mirrors ROM `src/fight.c:2815-2817` by entering combat through `multi_hit(ch, victim, TYPE_UNDEFINED)` instead of a single `attack_round()`. This restores haste / second-attack / third-attack follow-up swings for `kill` command combat starts. Regression coverage: `tests/integration/test_fight_c_do_kill_parity.py`.
+
+## [2.8.4]
+
+### Changed
+- **Pyright cleanup across test suite.** Eliminated ~30 pre-existing `reportOptionalMemberAccess` / `reportArgumentType` / `reportAttributeAccessIssue` warnings across `tests/test_boards.py`, `tests/test_logging_admin.py`, `tests/test_commands.py`, `tests/test_wiznet.py`, `tests/test_affects.py`, and `tests/test_account_auth.py`. Approach: per-call-site `assert x is not None` guards on `Optional` returns from `find_board(...)`, `from_orm(db_char)`, `load_character(...)`, `Character.pcdata`, etc.; `cast(StreamReader, …)` / `cast(TelnetStream, …)` / `cast(Room, …)` for test doubles; widened `DummyWriter.write` overrides to `bytes | bytearray | memoryview`; initialized possibly-unbound `menu_task: asyncio.Task | None = None`; `# type: ignore[arg-type]` on the two intentional `dam_type=None` cases that exercise edge-case handling. No behavior change; tests still pass at the prior pre-existing-failure baseline.
+
+## [2.8.3]
+
+### Changed
+- **Extract `time_info` persistence to `mud/world/time_persistence.py`.** `save_time_info`, `load_time_info`, `TimeSave`, and `TIME_FILE` now live in the new module. `tests/test_time_persistence.py` updated to import from the new location.
+- **Extract serialization helpers to `mud/db/serializers.py`.** All live helpers imported by `mud/account/account_manager.py` (`_normalize_int_list`, `_serialize_colour_table`, `_serialize_object`, `_serialize_pet`, `_serialize_skill_map`, `_serialize_groups`) and `mud/models/character.py:from_orm` (`_apply_colour_table`, `_normalize_int_list`, `_deserialize_object`, `ObjectSave`, `_deserialize_pet`) now live in `mud/db/serializers.py`. Both importers updated.
+
+### Removed
+- **`mud/persistence.py` deleted.** The deprecated stub (holding only the two extracted surfaces above, plus dead code from the now-gone JSON-pfile path) is gone. No behavior change; `mud/persistence` was already a non-functional deprecation banner since 2.8.1.
+
+## [2.8.2]
+
+### Changed
+- **Type cleanup on the new DB-canonical persistence surface.** `save_character_to_db` now types its `session` parameter as `sqlalchemy.orm.Session` (TYPE_CHECKING import) instead of `object`, dropping the `# type: ignore[union-attr]` on the `session.query` call. `mud/models/character.py:from_orm` casts `_deserialize_object(...)` to `Object | None` at the equipment-restore site so the `restored_equipment: dict[str, Object]` annotation is honored. No behavior change; pyright cleanup only. 25/25 critical persistence tests still green.
+
+## [2.8.1]
+
+### Changed
+- **INV-008 reversal — Phase 2: DB-canonical persistence is live.** `mud/account/account_manager.save_character` and `load_character` now hit the SQL `Character` row directly (via Phase 1's `save_character_to_db` and `Character.from_orm`). The JSON-pfile delegation is removed; `mud/persistence.py` keeps only `time_info` save/load and a few ROM-only helpers (deprecated banner at module top). `tests/integration/test_inv008_persistence_coherence.py` now asserts the DB-canonical round-trip including a new `test_inv008_db_canonical_is_sole_path` case. INV-008 is ✅ ENFORCED again under the new architecture.
+
+### Removed
+- **JSON-pfile test files** — deleted because the surface they covered no longer exists: `tests/test_persistence.py`, `tests/test_persistence_password_hash.py`, `tests/test_player_save_format.py`, `tests/test_inventory_persistence.py`, `tests/integration/test_pet_persistence.py`, `tests/integration/test_save_load_parity.py`, `tests/integration/test_tables_001_affect_migration.py`. Persistence coverage is now provided by `tests/integration/test_db_canonical_round_trip.py` (7 tests) + `tests/integration/test_inv008_persistence_coherence.py` (5 tests). The 3 pre-existing `test_persistence.py` / `test_inventory_persistence.py` failures (broken-area-JSON / vnum 3022) are no longer relevant — those tests went away with the surface they tested.
+
+## [2.8.0]
+
+### Added
+- **INV-008 reversal — Phase 1 of 2: DB-canonical persistence schema.** Following the discovery that `mud/account/account_service.create_character` and `mud/models/character.py:from_orm` still write/read the DB row at first-login (the JSON-pfile path was load-bearing only after first save), the project is reversing course on INV-008: the DB row is being made canonical for ALL player state, the JSON pfile path will be deleted in Phase 2. This commit extends the `Character` SQLAlchemy model with 39 new columns to hold every field `PlayerSave` round-trips (scalars: `max_hit`, `mana`, `move`, `gold`, `silver`, `exp`, `trust`, `invis_level`, `incog_level`, `saving_throw`, `hitroll`, `damroll`, `wimpy`, `position`, `played`, `logon`, `lines`, `prompt`, `prefix`, `title`, `bamfin`, `bamfout`, `security`, `points`, `last_level`, `affected_by`, `comm`, `wiznet`, `log_commands`, `pfile_version`, `board`; JSON columns: `mod_stat`, `armor`, `conditions`, `aliases`, `skills`, `groups`, `last_notes`, `colours`, `pet_state`, `inventory_state`, `equipment_state`). Extended `Character.from_orm` to hydrate every new column. New `save_character_to_db(session, char)` in `mud/account/account_manager.py` writes the full state via UPDATE. Round-trip proven by 7 new tests in `tests/integration/test_db_canonical_round_trip.py` (all green). Public `save_character` / `load_character` surface unchanged in this phase — Phase 2 swaps the implementations and deletes `mud/persistence.py`. INV-008 reopened in the cross-file invariants tracker.
+- **`docs/parity/INV008_REVERSAL_AUDIT.md`** — 71-field map (PlayerSave → Character columns), `from_orm` audit, new code surface, caller surface, INV-008 test rewrite plan, and risk register; produced as the spec for both phases of the migration.
+
+### Notes
+- Pre-existing test slowness (full suite ~12 min vs. AGENTS.md's "~16s") and ~30 pre-existing test failures verified at the pre-Phase-1 baseline (git stash). Not introduced by this commit.
+
+## [2.7.6]
+
+### Changed
+- **INV-008 DUAL-LOAD-CHARACTER-COHERENCE consolidated (hybrid path).** `mud/account/account_manager.py` is now a thin shim delegating `load_character` and `save_character` to `mud.persistence` (the JSON pfile path). The DB row (`mud/db/models.py:Character`) keeps `name` + `password_hash` for auth only; gameplay columns are vestigial and will be dropped in a later schema-migration session. Field-level audit at `docs/parity/INV008_DIVERGENCE_AUDIT.md`. No data migration was required — pre-launch.
+- **`mud/persistence.py:PlayerSave` now persists `password_hash`** so the JSON pfile is the single ROM-faithful source of truth for all PC state, including auth credentials. `save_character` writes `pcdata.pwd`; `load_character` restores it. The shim's `save_character` also syncs the hash to the DB row so the auth path (`account_service.login_character`) stays consistent after `do_password`.
+
+### Fixed
+- **30+ PC fields previously dropped on every WS logout** (because the DB-backed `account_manager.save_character` only persisted ~18 columns). Now restored on next login: current mana / current move, gold, silver, exp, hitroll, damroll, saving throw, AC array, wimpy, position, mod_stat array, comm bitfield, affected_by, wiznet flags, conditions (drunk/full/thirst/hunger), full skill learned-percent map, aliases, colour table, prompt, title, bamfin/bamfout, played time, logon timestamp, pets, item timer/value/condition/enchanted/affects on inventory and equipment, container contents.
+
+### Added
+- **INV-008 enforcement test** (`tests/integration/test_inv008_persistence_coherence.py`): four cases asserting the `account_manager` shim round-trips full gameplay state, registers loaded characters in `character_registry` (INV-003), preserves `pcdata.pwd`, and restores room placement. INV-008 flipped from ⚠️ KNOWN DIVERGENCE → ✅ ENFORCED. **All 8 cross-file invariants are now ✅ ENFORCED.**
+- `docs/parity/INV008_DIVERGENCE_AUDIT.md`: field-level provenance audit (51 vs 29 fields, caller surface, recommendation).
+
+## [2.7.5]
+
+### Added
+- **INV-007 RNG-DETERMINISM enforcement test** (`tests/test_rng_determinism.py`): scans every `.py` file under `mud/combat/`, `mud/skills/`, and `mud/spells/` for `import random`, `from random`, or `random.` and fails with path:line detail if any match. Runs in <1s. INV-007 flipped from ⚠️ ENFORCED BY CONVENTION → ✅ ENFORCED. 7 of 8 cross-file invariants are now ✅ ENFORCED; INV-008 (DUAL-LOAD-CHARACTER-COHERENCE) remains a known divergence.
+
+### Removed
+- **Vestigial `stdlib Random` from `SkillRegistry`** (`mud/skills/registry.py`): `__init__` accepted `rng: Random | None` and stored `self.rng`, but `self.rng` was never read — all rolls in registry.py already went through `rng_mm`. Removed the dead field, the dead parameter, and the `from random import Random` import. `tests/test_skills.py:load_registry` updated accordingly. Required before INV-007's grep test could be written cleanly.
+
+## [2.7.4]
+
+### Added
+- **INV-005 SAME-ROOM-COMBAT-ONLY enforcement test** (`tests/integration/test_inv005_same_room_combat.py`): proves `mud/game_loop.py:violence_tick` skips `multi_hit` and stops fighting when attacker and victim are in different rooms (ROM `src/fight.c:violence_update`). Plus a same-room sanity case so a flipped equality wouldn't silently pass.
+- **INV-006 FIGHTING-POINTER-COHERENCE enforcement test** (`tests/integration/test_inv006_fighting_pointer_coherence.py`): proves `mud/combat/engine.py:stop_fighting(victim, both=True)` sweeps `character_registry` and clears every attacker pointing at the victim (ROM `src/fight.c:stop_fighting(victim, TRUE)`); inverse case confirms `both=False` does not sweep.
+- Both invariants flipped from ⚠️ VERIFIED MANUALLY → ✅ ENFORCED in `docs/parity/CROSS_FILE_INVARIANTS_TRACKER.md`.
+- **Cross-File Invariants tracker** (`docs/parity/CROSS_FILE_INVARIANTS_TRACKER.md`): new top-level parity doc enumerating contracts the per-file audit methodology can't enforce on its own (single-delivery, registry membership, prompt-render-after-raw_kill, etc.). Eight INV-NNN entries seeded from this year's bug history; each has a stable ID, ROM mechanism, Python enforcement point, and regression test (or an action-item placeholder when the test is still missing).
+- **AGENTS.md "Cross-File Invariants" section**: methodology note explaining what per-file audits miss and how to use the new tracker. Tracker added to the "Trackers" table.
+- **rom-parity-audit skill update**: Phase 2 now requires following the call chain across module boundaries and consulting the cross-file invariants tracker; Phase 5 requires citing relevant INV-NNN statuses in each tracker row's Notes column ("Audited (per-file)" replaces bare "Audited").
+
+### Changed
+- `docs/parity/ROM_C_SUBSYSTEM_AUDIT_TRACKER.md`: `comm.c`, `fight.c`, and `save.c` rows annotated with the cross-file invariants they touch and the bugs the per-file audit missed (single-delivery, prompt clamp, registry membership). Per-file ratings preserved; the new annotations make the audit's actual coverage explicit.
+
+
+## [2.7.3]
+
+### Fixed
+- **Combat parity (death path)**: Combat messages now reach connected PCs exactly once. `mud/combat/engine.py:_push_message` previously appended every message to `char.messages` AND scheduled an async send, and `mud/net/connection.py:1756` unconditionally drained `char.messages` after every command — so each combat line was delivered TWICE to WebSocket clients. Live repro: PC dies, types `look`, sees the entire combat sequence (including "You have been KILLED!!") replayed milliseconds later, making it appear they died twice. ROM `src/comm.c:write_to_buffer` is one-shot. Per `docs/divergences/MESSAGE_DELIVERY.md`, the mailbox is fallback-only for tests/disconnected characters; fix returns immediately after the async dispatch when a connection exists.
+- **Combat parity (prompt display)**: `bust_a_prompt` now clamps displayed hit to >= 0. The death path could leave `char.hit` negative between `update_pos` setting `Position.DEAD` (at hit <= -11) and `raw_kill` clamping to `max(1, hit)` — Python's async dispatch can interleave a prompt render in that window, exposing the negative transient. ROM never shows negative hp because `raw_kill` always finishes first in its single-threaded loop. Display-only clamp at the two render sites (default prompt + `%h` token).
+- **Combat parity (death-path comments)**: `_handle_death` documents the bidirectional fighting-pointer clear and its relationship with `raw_kill → _stop_fighting`'s `char_list` sweep. ROM ref: `src/fight.c:damage` death branch.
+
+### Added
+- `tests/test_prompt_clamps_hp.py` — 4 cases guarding prompt clamp (default + custom `%h`).
+- `tests/integration/test_message_delivery_no_duplicate.py` — connected PC gets one async send, no mailbox queue; disconnected falls back to mailbox.
+- `tests/integration/test_pc_death_no_message_replay.py` — end-to-end: pushes the actual death-message sequence, drives the read-loop drain manually, asserts "You have been KILLED!!" appears exactly once across both passes.
+- `tests/integration/test_pc_death_keeps_connection.py` — regression guard that `raw_kill` keeps the PC in `character_registry`, in the death room, with hit/mana/move >= 1, position `RESTING`, connection intact.
+- `mud/net/connection.py` — diagnostic logging upgrade: read-loop's outer `except` now prints `type(exc).__name__: exc!r` plus traceback so future disconnect causes are visible in server stdout.
+
+## [2.7.2]
+
+### Fixed
+- Restore `version` field in `pyproject.toml` (accidentally dropped in `cdcd0cc`).
+- Combat one-way bug: `mud/account/account_manager.load_character` now appends loaded PCs to `character_registry` so `violence_tick`/`char_update`/idle pump iterate them. Mirrors ROM `src/save.c:fread_char` `char_list` membership.
+- World mob spawning: regenerated `resets` for 46 of 54 JSON area files via `scripts/patch_json_resets.py` so the school arena and other populated areas spawn ROM mobs on boot.
+- BOARD-010: `note read again` is now a no-op (ROM `src/board.c:569-572`).
+
+### Changed
+- Reconciled `BOARD_C_AUDIT.md`, `OLC_C_AUDIT.md`, and `ROM_C_SUBSYSTEM_AUDIT_TRACKER.md` against current implementation; folded several stale-test fixes (`test_save_load_parity`, `test_olc_alist`, `test_spell_affects_persistence`, `test_tables_001_affect_migration`, `test_nanny_login_parity`, `test_fighting_state`, `test_olc_save`) and added `tests/test_obj_loader.py`.
+- Session summaries added under `docs/sessions/` for the 2026-05-02 work (combat triage, board audit, OLC audit, asave cleanup, broad-suite triage / JSON itemtype fix).
+
+
+### Fixed
+- **OLC-004/005**: Active OLC editors now support ROM-style `commands` listings with five fixed-width columns, mirroring `src/olc.c:153-209`.
+- **OLC-010/015**: `do_olc` / `editor_table[]` ported — `olc <area|room|object|mobile|mpcode|hedit>` now dispatches to the real per-editor entry points via prefix matching (`str_prefix` parity), NPC guard, and remainder-arg forwarding, mirroring ROM `src/olc.c:646-690`. Both `olc` and `edit` command aliases are live. Integration tests: `tests/integration/test_olc_010_015_do_olc_router.py` (14 cases).
+- **OLC_SAVE-014**: `cmd_asave <vnum>` now silent on success (ROM `src/olc_save.c:982-995`)
+- **OLC_SAVE-015**: `cmd_asave world` returns `"You saved the world.\n\r"` (ROM exact)
+- **OLC_SAVE-016**: `cmd_asave changed` returns `"Saved zones:\n\r"` + per-area rows + `"None.\n\r"` fallback (ROM exact)
+- **OLC_SAVE-017**: `cmd_asave area` returns `"Area saved.\n\r"` (ROM exact)
+- **BOARD-010**: `note read again` is now a no-op (returns `""`) — mirrors ROM `src/board.c:569-572` empty `if`-body
+- **MPEDIT-003**: Added `MprogCode` dataclass + `mprog_code_registry` dict + `get_mprog_index()` — mirrors ROM `MPROG_CODE` struct and `mprog_list` linked list (`src/olc_mpcode.c`)
+- **MPEDIT-002**: `do_mpedit` now looks up `mprog_code_registry` (not `mob_prototypes`); opens `mpedit` session silently on success; exact Spanish error strings; `create` branch delegates to `_mpedit_create`
+- **MPEDIT-001**: `_interpret_mpedit` session loop — smash_tilde, empty→show, `done` silent, security re-check, area=NULL→edit_done, dispatch table, fallback to `interpret()`
+- **MPEDIT-004**: `_mpedit_show` — ROM exact format `"Vnum:       [%d]\n\rCode:\n\r%s\n\r"`
+- **MPEDIT-005**: `_mpedit_code` — no-arg enters string_edit mode; arg → `"Syntax: code\n\r"`
+- **MPEDIT-006**: `_mpedit_list` — `[%3d] (%c) %5d\n\r` format; `*/space/?` builder indicator; `all`/area filter; exact empty messages
+- **HEDIT-002**: `hedit level` accepts -1..MAX_LEVEL; exact ROM error message
+- **HEDIT-003/004**: `hedit keyword`/`hedit level` return `"Ok.\n\r"` / exact ROM syntax strings
+- **HEDIT-005**: `hedit text` no-arg is valid (triggers `string_append`); arg present → `"Syntax: text\n\r"`
+- **HEDIT-006**: security check returns `"HEdit: Insufficient security to edit helps.\n\r"` (with `\n\r`)
+- **HEDIT-007**: empty input in hedit session → `hedit_show` (not syntax string)
+- **HEDIT-008**: `done` is silent (returns `""`, no verbose save message)
+- **HEDIT-009**: unknown hedit command falls back to normal command table (`interpret`)
+- **HEDIT-010**: `hedit delete` implemented — removes from `help_entries` + all keyword buckets
+- **HEDIT-011**: `hedit list all` / `list area` implemented with ROM 4-column `%3d. %-14.14s` format
+- **HEDIT-012**: `do_hedit new <topic>` path correctly creates entry + enters editor
+- **HEDIT-013**: `do_hedit <keyword>` uses `is_name` word-match (not exact key lookup)
+- **HEDIT-014**: `hedit_level`/`hedit_keyword` return `"Ok.\n\r"` (triggers `had->changed = TRUE` equivalent) (`src/olc_act.c:770-790`) — sets `pArea->age`, validates numeric arg, does not set `changed` (ROM parity). 6 integration tests added.
+- **OLC-012/013/014**: `redit`/`oedit`/`medit` fallback to `interpret()` verified — `_should_fallback_from_olc()` + `_process_descriptor_input()` returning `None` correctly re-dispatches unknown OLC input through the normal command table (mirrors ROM `src/olc.c:519-521`, `575-577`, `631-633`). 14 integration tests added in `tests/integration/test_olc_012_014_editor_fallback.py`.
+- **OLC-009**: `medit` missing subcommands ported — `act`, `affect`, `armor`, `form`, `part`, `imm`, `res`, `vuln`, `off`, `size`, `hitdice`, `manadice`, `damdice`, `position`, `addmprog`, `delmprog` now dispatched from `_interpret_medit`. Helpers mirror ROM `src/olc_act.c:4142-4969`. Flag toggles use XOR pattern (act/affect/form/part/imm/res/vuln/off); `act` always re-sets `IS_NPC` (ROM:4153); dice stored as `tuple[int,int,int]`; `addmprog` validates vnum against mob_registry; `delmprog` splices list and clears mprog_flags bit. Integration tests: `tests/integration/test_olc_009_medit_missing_cmds.py` (30 cases).
+- **OLC-008**: `oedit` missing subcommands ported — `addaffect`, `addapply`, `delaffect`, `extra`, `wear` now dispatched from `_interpret_oedit`. Helpers mirror ROM `src/olc_act.c:2818-3450`: flag-value toggle for `extra`/`wear` (ExtraFlag/WearFlag XOR), affect list append/splice for `addaffect`/`addapply`/`delaffect` with `TO_OBJECT=1`, `type=-1`, `duration=-1` ROM defaults. Integration tests: `tests/integration/test_olc_008_oedit_missing_cmds.py` (16 cases).
+- **OLC-007**: `redit` missing subcommands ported — `rlist`, `mlist`, `olist`, `mshow`, `oshow` are now dispatched from `_interpret_redit`. Helpers `_redit_rlist/mlist/olist/mshow/oshow` mirror ROM `src/olc_act.c:329-570` (3-column vnum/name listing, `is_name` filtering, item-type prefix matching, mob/obj show via `_medit_show`/`_oedit_show`). Integration tests: `tests/integration/test_olc_007_redit_list_show.py` (16 cases).
+- **OLC-011**: `aedit` flag-toggle prefix ported — typing a bare `area_flags` name (e.g. `loading`, `added`) inside an active aedit session now toggles the flag via `flag_value(AreaFlag, command)` and sends `"Flag toggled."`, mirroring ROM `src/olc.c:443-449`. Integration tests: `tests/integration/test_olc_011_aedit_flag_toggle.py` (7 cases).
+
+## [2.7.1] — update.c Parity Gap Closures (GL-004/005/009/011/012/013/014/015/018)
+
+### Fixed
+- **GL-004**: `mana_gain` now uses `room.mana_rate` instead of `room.heal_rate`; rooms with custom mana rates now regenerate mana at the correct rate (ROM `update.c:300`).
+- **GL-005**: `mana_gain` furniture bonus now reads `value[4]` (mana bonus) instead of `value[3]` (hit bonus) (ROM `update.c:218,300`).
+- **GL-009**: NPC `char_update` wanders-home: out-of-zone NPCs that are not fighting and not charmed now have a 5% per-tick chance of being extracted (despawned) — mirrors ROM `update.c:688-696`. Previously missing entirely.
+- **GL-011**: Plague tick implemented: spreads to room occupants (5% per-tick per target, saves vs disease), drains mana and move, and deals HP damage per tick (ROM `update.c:794-846`). Previously missing.
+- **GL-012**: Poison tick implemented: sends "You shiver and suffer" message and deals `level // 10 + 1` HP damage per tick (ROM `update.c:848-862`). Previously missing.
+- **GL-013**: INCAP position tick damage: 50% chance of 1 HP damage per tick (ROM `update.c:864-867`). Previously missing.
+- **GL-014**: MORTAL position tick damage: 1 HP damage every tick unconditionally (ROM `update.c:868-871`). Previously missing.
+- **GL-015**: `_idle_to_limbo` now calls `stop_fighting(ch, both=True)` instead of `ch.fighting = None`, properly cleaning both sides of any fight (ROM `update.c:741-744`).
+- **GL-018**: Decay messages for objects inside an untakeable pit (vnum 3010) are now suppressed (ROM `update.c:1017-1018`). Previously objects inside a pit always broadcast their decay message to the room.
+
+### Added
+- `_char_update_tick_effects()` helper in `mud/game_loop.py` encapsulating all per-tick damage effects (plague, poison, INCAP, MORTAL).
+- Integration tests: `tests/integration/test_update_c_parity.py` — 11 tests covering all closed gaps.
+
 ## [2.7.0] — ROM Character-First Login
 
 ### Changed

{rom24_quickmud_python-2.7.0/rom24_quickmud_python.egg-info → rom24_quickmud_python-2.8.21}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: rom24-quickmud-python
-Version: 2.7.0
+Version: 2.8.21
 Summary: A modern Python port of the ROM 2.4b6 MUD engine with full telnet server and JSON world loading
 Author-email: Mark Jedrzejczyk <mark.jedrzejczyk@gmail.com>
 Maintainer-email: Mark Jedrzejczyk <mark.jedrzejczyk@gmail.com>
@@ -39,6 +39,7 @@ Provides-Extra: dev
 Requires-Dist: pytest>=8.0; extra == "dev"
 Requires-Dist: pytest-cov>=6.0; extra == "dev"
 Requires-Dist: pytest-timeout>=2.1.0; extra == "dev"
+Requires-Dist: httpx>=0.27; extra == "dev"
 Requires-Dist: mypy>=1.0; extra == "dev"
 Requires-Dist: ruff>=0.1.0; extra == "dev"
 Requires-Dist: black>=22.0; extra == "dev"
@@ -49,16 +50,17 @@ Provides-Extra: test
 Requires-Dist: pytest>=8.0; extra == "test"
 Requires-Dist: pytest-cov>=6.0; extra == "test"
 Requires-Dist: pytest-timeout>=2.1.0; extra == "test"
+Requires-Dist: httpx>=0.27; extra == "test"
 Dynamic: license-file
 
 # QuickMUD - A Modern ROM 2.4 Python Port
 
-[![Version](https://img.shields.io/badge/version-2.6.3-blue.svg)](https://github.com/Nostoi/rom24-quickmud-python)
+[![Version](https://img.shields.io/badge/version-2.8.21-blue.svg)](https://github.com/Nostoi/rom24-quickmud-python)
 [![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
-[![Tests](https://img.shields.io/badge/tests-3508%2F3521%20passing-brightgreen.svg)](https://github.com/Nostoi/rom24-quickmud-python)
+[![Tests](https://img.shields.io/badge/tests-4560%20passing-brightgreen.svg)](https://github.com/Nostoi/rom24-quickmud-python)
 [![ROM 2.4b Parity](https://img.shields.io/badge/ROM%202.4b%20Parity-gameplay%20100%25-success.svg)](ROM_2.4B6_PARITY_CERTIFICATION.md)
-[![Function Coverage](https://img.shields.io/badge/ROM%20C%20Functions-96.1%25-blue.svg)](FUNCTION_MAPPING.md)
+[![ROM C Audit](https://img.shields.io/badge/ROM%20C%20Audit-40%2F40%20audited-success.svg)](docs/parity/ROM_C_SUBSYSTEM_AUDIT_TRACKER.md)
 [![Integration Tests](https://img.shields.io/badge/integration%20tests-1000%2B-brightgreen.svg)](tests/integration/)
 
 **QuickMUD is a modern Python port of the legendary ROM 2.4b6 MUD engine**, derived from ROM 2.4b6, Merc 2.1 and DikuMUD. This is a complete rewrite that brings the classic text-based MMORPG experience to modern Python with async networking, JSON world data, and **100% ROM 2.4b behavioral parity**.
@@ -77,7 +79,7 @@ A "[Multi-User Dungeon](https://en.wikipedia.org/wiki/MUD)" (MUD) is a text-base
 - **⚔️ ROM Combat System**: Classic ROM combat mechanics and skill system
 - **👥 Social Features**: Say, tell, shout, and 100+ social interactions
 - **🛠️ Admin Commands**: Teleport, spawn, ban management, and OLC building
-- **📊 Comprehensive Testing**: 3,508 / 3,521 tests passing (99.6%) across unit, integration (1,000+), and command-registry suites
+- **📊 Comprehensive Testing**: 4,560 passing tests across unit, integration, and command-registry suites
 - **🔧 ROM C-Compatible API**: Public API wrappers for external tools and scripts (27 functions)
 
 ## 📦 Installation
@@ -191,8 +193,6 @@ parity logic in this backend repo.
 
 ## 🏗️ For Developers
 
-## 🏗️ For Developers
-
 ### Development Installation
 
 ```bash
@@ -206,8 +206,8 @@ pip install -e .[dev]
 ### Running Tests
 
 ```bash
-pytest  # Run all 3,500+ tests
-pytest tests/integration/ -v  # Run integration suite (1,000+ tests)
+pytest  # Run the full suite
+pytest tests/integration/ -v  # Run the integration suite
 ```
 
 ### Development Server
@@ -218,26 +218,19 @@ python -m mud  # Start development server
 
 ## 🎯 Project Status
 
-- **Version**: 2.6.73 (Production Ready — gameplay parity shipped, OLC cluster in progress)
+- **Version**: 2.8.21
 - **ROM 2.4b Gameplay Parity**: ✅ **100%** ([official certification](ROM_2.4B6_PARITY_CERTIFICATION.md)) —
   combat, skills, spells, movement, communication, world/db, save/load, mob programs,
   and all 255 ROM commands are implemented and audited.
-- **ROM C Source Audit**: 34 of 43 ROM C source files audited (24 at strict 100%);
-  remaining work is the OLC editor cluster (`olc.c`, `olc_act.c`, `olc_save.c`,
-  `olc_mpcode.c`, `hedit.c`) plus a small partial-audit tail (`board.c` 95%).
-  Several audited files sit at 85–98% with deferred-by-design or sibling-gated
-  gaps (`act_move.c`, `comm.c`, `nanny.c`, `const.c`, `music.c`, `board.c`).
-  Gameplay is not blocked on any of them. See
+- **ROM C Source Audit**: ✅ **100% audit-bound coverage** — 40 of 40 applicable ROM C files are audited, with
+  3 additional ROM files intentionally N/A (`recycle.c`, `mem.c`, `imc.c`). See
   [`docs/parity/ROM_C_SUBSYSTEM_AUDIT_TRACKER.md`](docs/parity/ROM_C_SUBSYSTEM_AUDIT_TRACKER.md).
-- **ROM C Function Coverage**: 96.1% (716/745 functions mapped)
-- **Test Suite**: 3,508 / 3,521 passing (99.6%), 11 skipped, 2 known
-  pre-existing failures, full run ~8 min. Three layers — unit (`tests/test_*.py`),
-  integration (`tests/integration/`, 1,000+ tests), and command-registry
-  (`test_all_commands.py`). Pre-existing skips/failures are catalogued in
-  [`docs/parity/PRE_EXISTING_FAILURES_TRIAGE.md`](docs/parity/PRE_EXISTING_FAILURES_TRIAGE.md).
-- **Active focus**: `olc.c` cluster — OLC-020/022/023 + STRING-001..012 closed;
-  remaining CRITICAL gaps (OLC-016/017/018/019, the `*edit create` subcommands)
-  gated on the held `olc_act.c` sibling audit.
+- **Cross-file Invariants**: ✅ **8/8 enforced** — message delivery, prompt clamping, registry membership,
+  same-room combat, death/connection behavior, RNG determinism, and persistence coherence are locked by dedicated tests.
+- **Test Suite**: ✅ **4560 passed, 4 skipped**. Three layers — unit (`tests/test_*.py`),
+  integration (`tests/integration/`), and command-registry (`test_all_commands.py`).
+- **Active focus**: parity work is effectively complete; current work is maintenance, frontend coordination,
+  and investigation only when a deterministic regression is reproduced.
 - **Compatibility**: Python 3.10+, cross-platform
 
 ## 🏛️ Architecture
@@ -323,7 +316,7 @@ Game systems are implemented in Python modules:
 - `mud/game_loop.py` drives the tick-based update loop.
 - `mud/commands` contains the command dispatcher and handlers.
 - `mud/combat` and `mud/skills` implement combat and abilities.
-- `mud/persistence.py` handles saving characters and world state.
+- `mud/account/` and `mud/db/` handle character persistence and account state.
 
 Start the server with:
 
@@ -377,17 +370,12 @@ QuickMUD is a **production-ready ROM 2.4b MUD** with ✅ **100% behavioral parit
 
 ### 📈 Quality Metrics
 
-- **Test Suite**: 3,508 / 3,521 passing (99.6%), 11 skipped, 2 known
-  pre-existing failures (full run ~8 min); pre-existing skips/failures are
-  catalogued in
-  [`docs/parity/PRE_EXISTING_FAILURES_TRIAGE.md`](docs/parity/PRE_EXISTING_FAILURES_TRIAGE.md).
+- **Test Suite**: 4,560 passing, 4 skipped on the last full recertification.
 - **Behavioral Parity**: 100% of ROM 2.4b6 gameplay subsystems audited (combat,
   skills, spells, movement, communication, world/db, save/load, mob programs,
   255/255 commands).
-- **Function Coverage**: 716/745 ROM C functions mapped (96.1%)
-- **ROM C Source Audit**: 34 of 43 files audited (24 at strict 100%); remainder
-  is the OLC editor cluster (`olc.c`/`olc_act.c`/`olc_save.c`/`olc_mpcode.c`/`hedit.c`)
-  and a small partial tail (`board.c` 95%).
+- **ROM C Source Audit**: 40 of 40 applicable ROM files audited, plus 3 intentional N/A files.
+- **Cross-file Invariants**: 8 of 8 enforced by dedicated regression tests.
 
 ### 🔧 Advanced Features
 

{rom24_quickmud_python-2.7.0 → rom24_quickmud_python-2.8.21}/README.md

@@ -1,11 +1,11 @@
 # QuickMUD - A Modern ROM 2.4 Python Port
 
-[![Version](https://img.shields.io/badge/version-2.6.3-blue.svg)](https://github.com/Nostoi/rom24-quickmud-python)
+[![Version](https://img.shields.io/badge/version-2.8.21-blue.svg)](https://github.com/Nostoi/rom24-quickmud-python)
 [![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
-[![Tests](https://img.shields.io/badge/tests-3508%2F3521%20passing-brightgreen.svg)](https://github.com/Nostoi/rom24-quickmud-python)
+[![Tests](https://img.shields.io/badge/tests-4560%20passing-brightgreen.svg)](https://github.com/Nostoi/rom24-quickmud-python)
 [![ROM 2.4b Parity](https://img.shields.io/badge/ROM%202.4b%20Parity-gameplay%20100%25-success.svg)](ROM_2.4B6_PARITY_CERTIFICATION.md)
-[![Function Coverage](https://img.shields.io/badge/ROM%20C%20Functions-96.1%25-blue.svg)](FUNCTION_MAPPING.md)
+[![ROM C Audit](https://img.shields.io/badge/ROM%20C%20Audit-40%2F40%20audited-success.svg)](docs/parity/ROM_C_SUBSYSTEM_AUDIT_TRACKER.md)
 [![Integration Tests](https://img.shields.io/badge/integration%20tests-1000%2B-brightgreen.svg)](tests/integration/)
 
 **QuickMUD is a modern Python port of the legendary ROM 2.4b6 MUD engine**, derived from ROM 2.4b6, Merc 2.1 and DikuMUD. This is a complete rewrite that brings the classic text-based MMORPG experience to modern Python with async networking, JSON world data, and **100% ROM 2.4b behavioral parity**.
@@ -24,7 +24,7 @@ A "[Multi-User Dungeon](https://en.wikipedia.org/wiki/MUD)" (MUD) is a text-base
 - **⚔️ ROM Combat System**: Classic ROM combat mechanics and skill system
 - **👥 Social Features**: Say, tell, shout, and 100+ social interactions
 - **🛠️ Admin Commands**: Teleport, spawn, ban management, and OLC building
-- **📊 Comprehensive Testing**: 3,508 / 3,521 tests passing (99.6%) across unit, integration (1,000+), and command-registry suites
+- **📊 Comprehensive Testing**: 4,560 passing tests across unit, integration, and command-registry suites
 - **🔧 ROM C-Compatible API**: Public API wrappers for external tools and scripts (27 functions)
 
 ## 📦 Installation
@@ -138,8 +138,6 @@ parity logic in this backend repo.
 
 ## 🏗️ For Developers
 
-## 🏗️ For Developers
-
 ### Development Installation
 
 ```bash
@@ -153,8 +151,8 @@ pip install -e .[dev]
 ### Running Tests
 
 ```bash
-pytest  # Run all 3,500+ tests
-pytest tests/integration/ -v  # Run integration suite (1,000+ tests)
+pytest  # Run the full suite
+pytest tests/integration/ -v  # Run the integration suite
 ```
 
 ### Development Server
@@ -165,26 +163,19 @@ python -m mud  # Start development server
 
 ## 🎯 Project Status
 
-- **Version**: 2.6.73 (Production Ready — gameplay parity shipped, OLC cluster in progress)
+- **Version**: 2.8.21
 - **ROM 2.4b Gameplay Parity**: ✅ **100%** ([official certification](ROM_2.4B6_PARITY_CERTIFICATION.md)) —
   combat, skills, spells, movement, communication, world/db, save/load, mob programs,
   and all 255 ROM commands are implemented and audited.
-- **ROM C Source Audit**: 34 of 43 ROM C source files audited (24 at strict 100%);
-  remaining work is the OLC editor cluster (`olc.c`, `olc_act.c`, `olc_save.c`,
-  `olc_mpcode.c`, `hedit.c`) plus a small partial-audit tail (`board.c` 95%).
-  Several audited files sit at 85–98% with deferred-by-design or sibling-gated
-  gaps (`act_move.c`, `comm.c`, `nanny.c`, `const.c`, `music.c`, `board.c`).
-  Gameplay is not blocked on any of them. See
+- **ROM C Source Audit**: ✅ **100% audit-bound coverage** — 40 of 40 applicable ROM C files are audited, with
+  3 additional ROM files intentionally N/A (`recycle.c`, `mem.c`, `imc.c`). See
   [`docs/parity/ROM_C_SUBSYSTEM_AUDIT_TRACKER.md`](docs/parity/ROM_C_SUBSYSTEM_AUDIT_TRACKER.md).
-- **ROM C Function Coverage**: 96.1% (716/745 functions mapped)
-- **Test Suite**: 3,508 / 3,521 passing (99.6%), 11 skipped, 2 known
-  pre-existing failures, full run ~8 min. Three layers — unit (`tests/test_*.py`),
-  integration (`tests/integration/`, 1,000+ tests), and command-registry
-  (`test_all_commands.py`). Pre-existing skips/failures are catalogued in
-  [`docs/parity/PRE_EXISTING_FAILURES_TRIAGE.md`](docs/parity/PRE_EXISTING_FAILURES_TRIAGE.md).
-- **Active focus**: `olc.c` cluster — OLC-020/022/023 + STRING-001..012 closed;
-  remaining CRITICAL gaps (OLC-016/017/018/019, the `*edit create` subcommands)
-  gated on the held `olc_act.c` sibling audit.
+- **Cross-file Invariants**: ✅ **8/8 enforced** — message delivery, prompt clamping, registry membership,
+  same-room combat, death/connection behavior, RNG determinism, and persistence coherence are locked by dedicated tests.
+- **Test Suite**: ✅ **4560 passed, 4 skipped**. Three layers — unit (`tests/test_*.py`),
+  integration (`tests/integration/`), and command-registry (`test_all_commands.py`).
+- **Active focus**: parity work is effectively complete; current work is maintenance, frontend coordination,
+  and investigation only when a deterministic regression is reproduced.
 - **Compatibility**: Python 3.10+, cross-platform
 
 ## 🏛️ Architecture
@@ -270,7 +261,7 @@ Game systems are implemented in Python modules:
 - `mud/game_loop.py` drives the tick-based update loop.
 - `mud/commands` contains the command dispatcher and handlers.
 - `mud/combat` and `mud/skills` implement combat and abilities.
-- `mud/persistence.py` handles saving characters and world state.
+- `mud/account/` and `mud/db/` handle character persistence and account state.
 
 Start the server with:
 
@@ -324,17 +315,12 @@ QuickMUD is a **production-ready ROM 2.4b MUD** with ✅ **100% behavioral parit
 
 ### 📈 Quality Metrics
 
-- **Test Suite**: 3,508 / 3,521 passing (99.6%), 11 skipped, 2 known
-  pre-existing failures (full run ~8 min); pre-existing skips/failures are
-  catalogued in
-  [`docs/parity/PRE_EXISTING_FAILURES_TRIAGE.md`](docs/parity/PRE_EXISTING_FAILURES_TRIAGE.md).
+- **Test Suite**: 4,560 passing, 4 skipped on the last full recertification.
 - **Behavioral Parity**: 100% of ROM 2.4b6 gameplay subsystems audited (combat,
   skills, spells, movement, communication, world/db, save/load, mob programs,
   255/255 commands).
-- **Function Coverage**: 716/745 ROM C functions mapped (96.1%)
-- **ROM C Source Audit**: 34 of 43 files audited (24 at strict 100%); remainder
-  is the OLC editor cluster (`olc.c`/`olc_act.c`/`olc_save.c`/`olc_mpcode.c`/`hedit.c`)
-  and a small partial tail (`board.c` 95%).
+- **ROM C Source Audit**: 40 of 40 applicable ROM files audited, plus 3 intentional N/A files.
+- **Cross-file Invariants**: 8 of 8 enforced by dedicated regression tests.
 
 ### 🔧 Advanced Features