## Summary
Phase #2 of the `.swb` step-cache size-reduction roadmap (after #2465, v5 uniform elision). Stores the 16 base directional Z arrays as masked residuals against each cell's own SourceZ and omits any array that matches its prediction. Lossless, byte-identical reconstruction.
Trammel: 231.9 MB → 124.7 MB (−46%).
## Details
- Predictor: `predict = mask bit ? SourceZ : 0` (matches the baker's 0 on unwalkable directions); residual `Z − predict` via unchecked two's-complement (byte-exact for all inputs); reconstruct `Z = predict + residual`.
- A `u16 ZArrayMask` flags which of the 16 base arrays differ from prediction; matching arrays are omitted and synthesized from mask + SourceZ at read.
- Serializer-layer only: StepChunk, the cache, the algorithm, and the baker are unchanged.
- Format v6; v5 files rejected and re-baked once.
## Tests
21 v6 unit tests; full pathfinding suite green; Release build clean.
## Issue
Fixes#2452. A player with 30 Ninjitsu reported that the Animal Form menu showed **every** form; selecting one above their skill (e.g. Dog, req 40) **consumed mana** and returned "you need at least 40 skill", and afterwards the **gump never reopened** — every recast silently re-attempted the unusable form and drained more mana.
## Root cause
Three linked bugs, all reproduced from the code:
1. **Gump not gated by skill.** `AnimalFormGump.BuildLayout` compared `Skill.Fixed` (which is `Value * 10`, so 30 skill → `300`) against the raw 0–100 `ReqSkill` (Dog = `40`). `300 >= 40` is always true, so all forms were shown. `Morph` itself correctly uses `.Value`.
2. **Mana charged on a no-skill cast.** `Morph` returns `MorphResult.NoSkill` for an under-skilled form, but both call sites (`OnCast`, `OnResponse`) only special-cased `MorphResult.Fail`; `NoSkill` fell through to the branch that deducts mana.
3. **Menu never reopened.** Per OSI ([uo.com](https://uo.com/wiki/ultima-online-wiki/skills/ninjitsu/), [uoguide](https://www.uoguide.com/Animal_Form)), casting while **standing still always opens the selection menu**, and casting while **moving** quick-transforms into the last selected form. ModernUO only opened the menu when `lastAnimalForm == -1`, so once any form was selected a stationary recast skipped the menu.
## Fix
- Add `AnimalForm.CanSelectEntry` (compares `Skill.Value` to `ReqSkill`, plus the talisman check) and use it for the gump's per-entry enable check.
- `OnCast`: standing still always opens the menu; moving quick-transforms into the last form. `NoSkill` no longer costs mana.
- `OnResponse`: handle `Success` / `Fail` / `NoSkill` explicitly so `NoSkill` costs no mana.
## Tests
Adds `AnimalFormTests`:
- `CanSelectEntry` rejects forms above skill, accepts forms at/below skill, and requires a talisman for talisman-gated forms.
- `Morph` returns `NoSkill` (without transforming) when under-skilled, and `Success` when sufficiently skilled.
Verified the gating test catches the regression (reintroducing `.Fixed` fails it). Full solution build is clean; the 5 new tests plus 284 other UOContent tests pass (the pathfinding/AI sequential tests were excluded only because they deadlock under concurrent local runs — they are unrelated to this change).
## Summary
Sub-project #1 of the `.swb` step-cache size-reduction roadmap (`dev-docs/pathfinding.md` § Future work). Adds **uniform-chunk elision** to the `StepCacheFile` format, bumping it **v4 → v5**.
A fully-uniform 16×16 chunk — no strata, **no swim layer**, all 19 base arrays constant (open ocean, Green Acres, void) — serializes to a **~28-byte record** (`KindUniform`) instead of ~5,393, and reconstructs **byte-identically** via `Array.Fill`. Non-uniform chunks use the existing v4 body (`KindFull`) with the swim-layer and strata trailers **fully preserved** — the Kind byte is just prepended.
## Calibrated result (measured, not projected)
Baked Trammel via `SaveToFile`:
| | |
|---|---:|
| Chunks | 114,688 |
| Uniform (swim-aware) → elided | 62.7% |
| Swim-layer chunks (stay Full) | 8.9% |
| Strata chunks (stay Full) | 1.8% |
| Baseline (full records) | 592.2 MB |
| **Actual v5 `.swb`** | **231.9 MB (−61%)** |
The residual is ~150 MB of non-uniform land Z-blocks (targeted by #2 predictive-Z) + ~81 MB of swim-layer trailers (#2/#3). #2 and #3 are separate follow-up PRs.
## Implementation
- `StepChunk.IsUniform()` — false if it has strata **or a swim layer**, else true only when all 19 base arrays are constant (the "all-same" check uses the SIMD-accelerated `ContainsAnyExcept`).
- `StepCacheFile` v5 — `Kind` byte (`KindFull=0`/`KindUniform=2`, 1 reserved); uniform write/read; `FormatVersion`/`MinSupportedVersion` → 5 (v4 files rejected on open and re-baked). No `StepCache`/algorithm/index changes; fingerprint logic untouched.
## Tests
7 `StepCacheFileV5Tests` (uniform round-trip + `<200 B` compactness, varied-full, swim-layer-full, strata-full, swim+strata combined, v4 version-gate rejection) + the existing StepCache/pathfinding suite — **70 pass**, including the prior `SwimLayer_RoundTrips`. An independent review verified write/read symmetry, cast round-tripping, swim/strata preservation, and the version gate (READY TO MERGE).
Fixes#2462
## Summary
Removes the per-object `VirtualHairInfo` heap wrapper for mobile/corpse hair. Hair is now stored **inline** on `Mobile` and `Corpse` as `int _hairItemId` / `int _hairHue` plus a lazily-allocated, **non-serialized** ephemeral `Serial _hairSerial` (in the high virtual-serial range) — and likewise for facial hair. The `VirtualHairInfo` class is deleted, with a **lossless** save migration.
This delivers three things:
1. **Fixes a hair-removal bug.** `Delta(MobileDelta.Hair)` is deferred (it enqueues; `ProcessDeltaQueue` runs later in the tick). The old `HairItemID = 0` setter nulled `_hair` *immediately*, so by the time `ProcessDelta` built the remove packet the equipped virtual serial was already gone — the old `??=` code then re-materialized a **fresh** serial (≠ the equipped one), so clients never removed the right entity, and it left a phantom ItemId-0 object behind. The serial now lives on the entity and **persists across removal**, so remove packets carry the correct serial.
2. **Lightens the entity.** No heap hair object; bald mobiles allocate nothing (the serial is minted lazily only when hair is present). This was the original reason `HairItemID`/`HairHue` exist.
3. **Removes `VirtualHairInfo` entirely**, keeping the high-range virtual serial behavior.
## How
- **Mobile** (manual serialization): inline `_hairItemId/_hairHue/_hairSerial` (+facial); lazy `HairSerial`/`FacialHairSerial`; `ProcessDelta` reads those. Serialization **v36 → v37** — the v30-v37 deserialize is unified, reading the legacy per-hair `VirtualHairInfo` version int only when `version < 37`. Setting item id to 0 clears the hue (matching the old object-nulling) while retaining the serial.
- **Corpse** (codegen serialization): decomposed to `[SerializableField] int _hairItemId/_hairHue` (+facial) + ephemeral serial; **v16 → v17** with `MigrateFrom(V16Content)`.
- **Lossless migration:** the loader validates exact byte length, and the old corpse hair is a presence-bool-gated block, so a tiny **migration-only** `LegacyHairInfo` reader (no runtime role) consumes the legacy `[bool][int ver][int itemId][int hue]` bytes. Frozen `Corpse.v14/v15/v16.json` are retyped to it; `v17.json` describes the new int fields.
- All consumers updated to discrete accessors: `OutgoingMobilePackets`, `CorpsePackets`, corpse subclasses (`MilitiaFighterCorpse`, `SchmendrickApprenticeCorpse`), and the packet test mirrors.
- `VirtualHair.cs` renamed to `OutgoingVirtualHairPackets.cs` (the only type left in it after `VirtualHairInfo` was removed).
## Test Plan
- [x] Full solution build: **0 warnings, 0 errors** (`TreatWarningsAsErrors`).
- [x] `Server.Tests`: **708 passed** (incl. new `RemoveHairUsesEquippedSerial` / `RemoveFacialHairUsesEquippedSerial` proving the serial survives removal + hue clears).
- [x] `UOContent.Tests` corpse/hair: **6 passed** (incl. `CorpseHairMigrationTests` asserting the legacy hair bytes are consumed exactly — the loader's length invariant).
- [x] Generated migration code inspected: V14/V15/V16 readers consume the legacy block byte-for-byte; serial never written to disk.
## Upgrade notes
- Old Mobile (v30–v36) and Corpse (v13–v16) saves load losslessly.
- Minor cosmetic-only change: `SchmendrickApprenticeCorpse` hair/facial-hair RNG draws shift order within each pair (same draw count); irrelevant for a quest NPC corpse.
## Summary
Closes the Cold-cache regression flagged in PR #2450. `StepCache.TryGetMask` no longer eagerly runs `BuildChunk` on the first miss for a chunk that isn't in a `.swb` lazy reader. Instead it returns `Fallthrough_NotBuilt` and the caller (`BitmapAStarAlgorithm`) takes the per-cell slow path. The chunk is only promoted to the bitmap fast path after the **second** miss within a 30-second window, filtering single-touch pass-throughs.
This makes BitmapAStar's worst-case (cold cache + short hops) collapse from **12–47× slower** than FastAStar to **roughly the same**, which is the floor the slow path can deliver. Steady-state warm performance (the actual deliverable) is unchanged from PR-5 — it was always the cache fast path.
## The pet-follow scenario this fixes
A mounted player at ~4 tiles/sec with a pet/hireable following will trigger an NPC pathfind every 100–300 ms. Each pathfind is 1–6 tiles. As the player crosses chunk boundaries (~4 sec/chunk), the pet's first pathfind in the new chunk under the previous behavior triggered a full ~700 µs `BuildChunk` for a chunk the player would leave shortly after. At 50–100 mobiles per shard, this exceeded the 8 ms tick budget. PR-5 BDN data showed scenarios 6–9 (2–8 tile NPC perception) at 2,300–3,700 µs Cold vs FastAStar's 80–200 µs.
Under the new gate:
- First miss → `Fallthrough_NotBuilt` → caller uses slow path (~30–50 µs short path). No `BuildChunk`. No allocation.
- Player keeps moving → chunk never gets a second touch within window → never promoted, no rot.
- NPC patrolling a fixed territory → repeatedly hits the same chunks → second touch within window → promote → cache fast path on subsequent calls.
## What changed
- **`CacheHitKind.Fallthrough_NotBuilt = 6`** + **`CacheStats.FallthroughNotBuilt`** counter. `IsHit=false`, so the caller routes to slow path.
- **`StepCache._chunkMissTracker`** — `Dictionary<long, ChunkMissState>` capped at 4096 entries. State is `(byte missCount, uint lastMissTickStamp)` keyed by chunk key. Window-expired entries reset count to 1; capacity overflow prunes window-old entries first.
- **`StepCache.MissPromotionThreshold`** (default `2`) and **`StepCache.MissPromotionWindowMs`** (default `30_000`) — tunable, can be wired through `ServerConfiguration` if shards want different policy. Setting threshold to `1` restores legacy eager-build behavior (used by tests that prime chunks via single `TryGetMask` call).
- **`StepCache.TryGetMask` miss branch** — try lazy reader first (file-loaded chunks bypass the tracker entirely; an `.swb` represents an explicit prior decision to keep the chunk warm). Otherwise consult the tracker.
- **`BitmapAStarAlgorithm.GetSuccessorsSlowPath`** now layers `IsBlockedByDynamic` on top of `CalcMoves.CheckMovement`. Previously the slow path only ran for `CanFly` creatures and rare cache fallthroughs — `CheckMovement` doesn't iterate same-cell mobiles, so the bitmap fast path's `IsBlockedByDynamic` was the only mobile-blocking check. Now first-touch pathfinds run through the slow path, so the gap had to close.
## Tests
50 pathfinding tests pass (was 47). New / updated:
- **`TryGetMask_FirstTouchOnUnbuiltChunk_DefersBuildAndReturnsFallthrough`** — single TryGetMask call returns `Fallthrough_NotBuilt`, no chunk built, no allocation.
- **`TryGetMask_SecondTouchWithinWindow_PromotesAndBuilds`** — second call inside the 30s window builds + serves.
- **`TryGetMask_SecondTouchAfterWindow_RestartsCounterAndDefers`** — second call outside the window restarts the count, returns Fallthrough again.
- **`TryGetMask_DistinctChunks_TrackedIndependently`** — counters are per-chunk; one touch on each of two adjacent chunks both stay in fallthrough.
- **`LazyReaderHit_BypassesMissTrackerOnFirstTouch`** — open `.swb` + first touch hits without consulting the tracker. Production with `.swb` loaded skips the gate entirely.
- **`MultisVersion_Bump_TriggersDirtyRebuild`** — updated to reflect the new 3-step flow (Fallthrough → Miss_NotBuilt → Miss_DirtyRebuild).
- Tests that prime chunks via a single `TryGetMask` call (multi-Z, Tier4, lifecycle, parity, BitmapAStar uses-cache) set `MissPromotionThreshold = 1` to opt into eager behavior.
## Expected BDN impact
The Cold column from PR-5's BDN should change as follows once the bench's submodule pointer is updated to this branch:
| # | Scenario | Cold (PR-5) | Cold (PR-6 expected) | FastAStar Cold |
|--:|-----------------|-------------:|---------------------:|---------------:|
| 2 | sewer corridor | 1,627 µs | ~36 µs | 36 µs |
| 4 | causeway | 1,533 µs | ~39 µs | 39 µs |
| 6 | pet 2-tile | 2,364 µs | ~80 µs | 81 µs |
| 8 | npc 5-tile | 3,708 µs | ~140 µs | 141 µs |
| 9 | npc 8-tile | 2,386 µs | ~200 µs | 197 µs |
WarmNoFile and LazyWarm rows should be unchanged — they were always cache-warm. The miss tracker only fires when neither resident chunks nor the lazy reader can satisfy the request.
## Future work (not in this PR)
- **Background-thread bake**: builds outside the game thread so even promoted chunks don't pay the 700 µs build cost on the main thread. Rule 10 (no Task.Run) applies, so this needs careful design — the bake is a pure data transform but main-thread synchronization on chunk-state transitions has to be threaded through. Defer to a follow-up.
- **Long-traverse BDN scenario**: a multi-Find benchmark simulating 50 pet repaths across chunk transitions. Requires restructuring the bench harness; the existing 10-scenario corpus + Cold provider already exercises the gate.
- **Swim sourceZ bake**: scenario 5 (sea serpent) shows 56 B alloc on warm paths because the cache's SourceZ is computed under default-walker rules. Swim creatures fall through to slow path. Independent of this PR.
## Summary
Creatures (pets following, monsters chasing, NPCs approaching) would **oscillate — "pace back and forth really fast"** at concave obstacles (reported at the Britain Inn L-desk: a pet at `(1493,1614,20)` never reaching its master at `(1494,1605,21)`) instead of routing around them.
**Root cause** (the A* pathfinder itself was correct): all goal-seeking funnels through `MoveTo` and `WalkMobileRange` → `MoveTowardsOrAwayFrom`, which step greedily via `DoMove(dir, badStateOk:true)`. `DoMove` returns `true` even when the direct step was blocked and the creature merely **auto-turned and sidestepped** (`MoveResult.SuccessAutoTurn`), and the caller then set `Path = null`, discarding the `PathFollower`. So at a concave obstacle a non-progressing sidestep was mistaken for progress and the creature never committed to a route. (AOS pet-follow runs at `CurrentSpeed = 0.1`, hence the "really fast" shuffle.)
## What changed
- **New centralized `BaseAI.ApproachTarget(target, run, range)` primitive.** A greedy step is committed only when it **fully succeeds (`MoveResult.Success`) and actually gets closer**; otherwise the creature commits to a **persistent `PathFollower`** that routes around the obstacle and is never discarded by a greedy step. The open-terrain fast path (one greedy step, no pathfinding) is preserved. `MoveTo`, `MoveTowardsOrAwayFrom`, and `MoveToWithCollisionAvoidance` all delegate to it — public signatures unchanged, so no AI-class call site changes.
- **Best-distance give-up + idle.** A creature that cannot reach a **stationary** in-range goal stops shuffling and idles after `ApproachGiveUpTicks` (40) ticks without lowering its closest-ever distance; a **moving** goal (active chase) never gives up. It resumes the moment the goal moves.
- **Pathfinder fix (required):** `BitmapAStarAlgorithm.IsBlockedByDynamic` now skips the dynamic mobile-block check **at the goal cell only** (`MoveImpl.Goal`). Previously A* returned `null` whenever the target mobile stood on the goal cell, so creatures could never pathfind *toward* another mobile — only toward empty ground. The follower stops within `range` short of it. Static/item blocking and all non-goal mobile blocking are unchanged.
## Tests
New AI-loop integration tests in `ApproachTargetTests.cs` drive the real `BaseAI` primitives against live Britain Inn map statics: exact-repro pet follow, open-terrain (asserts zero pathfinding), `MoveTo` chase (static + walking-away target), route-around-a-dynamic-wall, and walled-off give-up-and-idle.
- Pathfinding + AI subset: **52/52** pass.
- Full `UOContent.Tests`: **301/301** pass. (Note: the test host lingers on shutdown — a pre-existing infra quirk unrelated to this change; all tests complete and pass.)
- Full solution build: clean (0 warnings / 0 errors).
## Notes
- Branched off `main`; independent of the in-flight step-cache work.
- Out of scope (future work): proactive "SmartAI" look-ahead pathfinding so clever creatures plan a route before walking into the obstacle, rather than reacting after they hit it.
## Test Plan
- [X] In-game: order a pet to `follow`/`come` across the Britain Inn L-desk; confirm it routes around and reaches you instead of pacing.
- [X] Aggro a monster and kite it around a building/treeline; confirm it chases around obstacles.
- [X] Confirm open-terrain following/chasing feels unchanged (no extra latency).
- [X] Confirm a creature with a genuinely unreachable target idles rather than shuffling forever.
## Summary
Multi-Z cells (bridges, stairs, paver-over-ground, multi-floor structures) now carry **per-stratum walkability data** in the cache instead of falling through to the slow path. The data is computed at chunk-build time, persisted in the `.swb` file, and selected at query time by matching the request's `sourceZ` against each stratum's `zCenter` (within `StepHeight` tolerance).
This is the Tier 4 strata feature, deferred from PR #2447 / PR #2448 / PR #2449. Builds on PR #2449's lazy backing store and public bake helpers.
## Wire format change (v1 → v2)
`StepCacheFile.FormatVersion = 2`. `MinSupportedVersion = 2`. v1 `.swb` files are silently rejected at open time (treated as missing) and overwritten on the next `SaveToFile` / `BakeMap`. **No migration** — older files just get re-baked.
The `MinSupportedVersion` sentinel is the model going forward: bump the constant when an incompatible change lands; admins re-bake on the next deploy. No matrix of v1↔v2↔v3 migration logic to maintain.
## What changed
- **`StepProbe.ComputeStrataAt(map, x, y)`** — enumerates walkable standing-Zs at the cell (one per land surface plus one per walkable static), collapses Zs within `2*StepHeight`, runs `ComputeMaskAt` at each surviving Z. Returns `null` for single-Z cells (caller uses the chunk's main mask).
- **`StepChunk`** — replaces the old `MultiZCells` bitmap with a **strata storage pair**:
- `ushort[256] StrataOffsetByCell` (sentinel `NoStrata = 0xFFFF` = "no strata for that cell")
- `byte[] StrataData` packed: `u8 stratumCount`, then `count × 19-byte stratum`
- `sbyte zCenter, byte walkMask, byte wetMask, sbyte walkZ_N..NW (8), sbyte swimZ_N..NW (8)`
- `IsCellMultiZ` derives from `StrataOffsetByCell[cell] != NoStrata` — same semantics, single source of truth.
- **`StepCache.BuildChunk`** — populates strata for cells flagged multi-Z via `SetStrata`. Chunks with zero multi-Z cells pay zero strata overhead (offset array + data array stay null).
- **`StepCache.TryGetMask`** — for multi-Z cells, scans strata with `TryStratumHit`; returns the matching one with `HitKind=Hit`. Falls through to slow path only when no stratum matches the query `sourceZ`.
- **`StepCacheFile`** — v2 serialization with strata trailer per chunk + `recordLength` in index entry. Lazy reader sizes scratch per-chunk-record using the recorded length, growing on demand for multi-Z-heavy chunks. Patches `IndexOffset` on `w.Buffer` (BufferWriter's current backing array) since it grows during variable-size chunk writes.
## File layout v2
```
Header (48 bytes):
u32 Magic = 0x42575300 ('SWB\0')
u32 Version = 2
u32 MapId
u64 Fingerprint XxHash3 over LandTable + ItemTable flags + map files (mapX.mul/.uop, staidxX.mul, staticsX.mul)
u64 BakeTimestamp informational
u32 ChunkCount
u64 IndexOffset position where chunk index begins
Per chunk (variable size):
u16 ChunkX, ChunkY
u32 BuiltMultisVersion
u8 HasStrata 0 = no strata trailer; 1 = strata trailer follows
byte WalkMask[256], WetMask[256]
sbyte SourceZ[256], WalkZN..WalkZNW[256], SwimZN..SwimZNW[256]
// Strata trailer (only when HasStrata == 1):
u16 StrataOffsetByCell[256] // NoStrata sentinel = 0xFFFF
u32 StrataDataLength
byte StrataData[StrataDataLength]
Per multi-Z cell: u8 count, then count × Stratum (19 bytes)
Index trailer (20 × ChunkCount bytes):
per chunk: { u64 chunkKey, u64 fileOffset, u32 recordLength }
```
## Summary
Adds two pieces of pathfinding tooling on top of PR #2448's lazy `.swb` infrastructure:
- **`PathfindRecorder`** — admin-toggled JSONL telemetry capture; one record per `BitmapAStarAlgorithm.Find` call. Output format matches the BDN harness corpus, so production traffic can be captured and replayed in benchmarks without an adapter.
- **Public bake helpers on `StepCache`** — `ComputeLiveTileDataHash`, `TryReadTileDataHashFromFile`, `BakeMap`, `ClearResidentChunks`. Lets the benchmark project (and any future bake utility) drive cache fill + persist without exposing internal types.
The companion BDN harness update lives in [ModernUO-Benchmarks#kb/pathfinding-pr4-bench](https://github.com/modernuo/ModernUO-Benchmarks/tree/kb/pathfinding-pr4-bench): porting `Benchmarks/PathfindInGame/` from the `kb/ai_pathfinding` branch to the API shipped in #2446–#2448.
## What's in this PR
### `PathfindRecorder` (`PathfindRecorder.cs`)
- Holds a single `StreamWriter` open while recording; its internal buffer absorbs per-record writes without per-call `File.AppendAllText`.
- Single `bool` check on the hot path; cheap when disabled.
- Disabling flushes + disposes; an IO failure during write also disables the recorder.
- Server config:
- `pathfinding.recorder.enable` — bool, default `false`. Read on boot via `GetOrUpdateSetting`.
- `pathfinding.recorder.path` — default `<basedir>/Data/Pathfinding/recordings/pathfinds.jsonl`.
- Hooked into `BitmapAStarAlgorithm.Find` — runs once per call, does nothing when disabled.
- Admin command: `[PathRecord [on|off|flush|status]` (default `status`).
### Public cache helpers
- `static ulong StepCache.ComputeLiveTileDataHash()` — wraps the file module's hash function for staleness checks.
- `static bool StepCache.TryReadTileDataHashFromFile(string, out ulong)` — peeks at a `.swb` file's hash field (20 bytes).
- `int StepCache.BakeMap(int, string)` — walks every chunk in the map, populates resident set, saves. Offline / fixture use; blocks for many seconds on a full-map walk.
- `void StepCache.ClearResidentChunks()` — drops chunks + zeros counters but keeps lazy readers open. Lets benchmark loops measure "first query after boot" cost across iterations without the lazy-reader reopen overhead.
## Summary
Adds a binary disk format + lazy reader so the step cache can warm-start from a precomputed file without paying chunk-build cost on the first pathfind through a region. **Resident memory stays bounded by `MaxResidentChunks` regardless of file size** — opening a `.swb` reads only the header + chunk-offset index (~16 bytes per indexed chunk), and individual chunks are seeked + deserialized only when `ResolveMissingChunk` asks for them.
The lazy design (vs. an eager bulk load): a 250 MB bake on a RAM-constrained shard never materializes more than the LRU cap (~40 MB at the default 8192-chunk cap), and unwanted regions never enter memory at all.
Builds on PR #2447.
## What changed
- **`StepCacheFile`** — binary reader/writer module. Writer emits header → chunks (offsets recorded) → index trailer, then patches the header's `IndexOffset` field. Reader is `OpenForLazy(path)` returning a `LazyReader` that holds an open `FileStream` + offset dictionary.
- **`StepCacheFile.LazyReader`** — `TryReadChunk(chunkX, chunkY)` does a single seek + bulk read for one record. `Dispose` releases the underlying stream. Files are opened with `FileShare.Read | FileShare.Delete` so admin tooling can replace them.
- **TileData fingerprint via XxHash3.** The `.swb` header carries a hash of `LandTable + ItemTable` flags. Load rejects any file whose hash doesn't match the running server. Computed via `HashUtility.ComputeHash64` (engine-blessed hasher) — adds a `ReadOnlySpan<byte>` overload alongside the existing `ReadOnlySpan<char>` one for parity.
- **`StepCache.SaveToFile(path, mapId)`** — writes resident chunks for the given map.
- **`StepCache.TryOpenLazyReader(path, mapId)`** — opens the file, validates header, holds the reader for the map's lifetime.
- **`StepCache.ResolveMissingChunk`** — now consults the lazy reader before invoking the runtime baker. A loaded chunk whose `BuiltMultisVersion` doesn't match the live sector falls through to the baker (snapshot was made before a multi was added/removed in that sector).
- **`StepCache.Clear` closes lazy readers.** Test cleanup can delete `.swb` files cleanly.
- **Auto-load at startup.** `PathCacheCommands.Configure()` opens `Data/Pathfinding/<mapId>.swb` as a lazy reader for every map.
- **`[PathCacheSave`** / **`[PathCacheLoad`** — admin commands for the same workflow.
- **`pathfinding.maxResidentChunks` shard-tunable.** Read from `server.cfg` at boot via `ServerConfiguration.GetOrUpdateSetting` (default 8192 ≈ 40 MB). Small shards can tune down; large shards with substantial bakes can tune up to reduce eviction churn. Default is written back to `server.cfg` on first boot, matching the engine pattern used by other settings.
## File layout (v1)
```
Header (48 bytes):
u32 Magic = 0x42575300 ('SWB\0')
u32 Version = 1
u32 MapId
u64 TileDataHash XxHash3 over LandTable + ItemTable flags (HashUtility)
u64 BakeTimestamp informational
u32 ChunkCount
u64 IndexOffset file position where the chunk index begins
Chunk records (fixed size, ~5,393 bytes each, +32 if multi-Z):
u16 ChunkX
u16 ChunkY
u32 BuiltMultisVersion
u8 HasMultiZ
byte WalkMask[256], WetMask[256]
sbyte SourceZ[256], WalkZN..WalkZNW[256], SwimZN..SwimZNW[256]
[byte MultiZCells[32] when HasMultiZ == 1]
Index trailer (16 × ChunkCount bytes):
(u64 chunkKey, u64 fileOffset)
```
## Memory math
| Scenario | Disk file | RAM at boot | Notes |
|---|---|---|---|
| Empty / no `.swb` files | — | 0 | Silent; cache builds on demand. |
| Admin-curated towns (5K chunks) | 25 MB | 0 + per-query | Index ≈ 80 KB. Resident grows to the configured cap under steady-state queries. |
| Full-map bake (50K chunks) | 250 MB | 0 + per-query | Index ≈ 800 KB. Same configured cap. Cold areas never load. |
| All 5 maps fully baked | 1.25 GB | 0 + per-query | Index ≈ 4 MB total. Same configured cap. |
## Hash choice (FNV-1a → XxHash3)
The original draft used inlined FNV-1a-64. Switched to XxHash3 via `HashUtility`:
- ~30× faster on this workload (~30 GB/s SIMD vs ~2 GB/s byte-by-byte). Boot-time only, so absolute saving is microseconds — the real wins are elsewhere.
- Stronger collision resistance and distribution.
- Drops ~25 lines of inlined hash code; matches the rest of the codebase's hashing pattern.
- Hash is stable as long as `HashUtility`'s `xxHash3Seed` constant doesn't change (already marked `// DO NOT CHANGE THIS NUMBER`).
## Summary
Builds on PR #2446's cache-direct A*. The previous PR conservatively routed players + creatures with capability flags entirely through the slow path. This PR pushes that line: most mobile classes now use the cache, with the right rule set layered on top per-mobile, and the cache fast-path now does the dynamic items / mobiles check that PR #2446 had silently skipped.
## What changed
- **Non-GM players** now use the cache. Diagonal corner-cut applies the strict AND-rule (BOTH cardinal partners walkable) by reading the same source-cell mask byte the creature OR-rule reads — both rules are evaluable from one byte.
- **Creatures with `CanOpenDoors` / `CanMoveOverObstacles`** now use the cache. Reading `MovementImpl` confirmed those flags only affect dynamic items, never static tiles, so they were over-conservatively excluded before.
- **Swim creatures** now use the cache via a capability overlay. `StepProbe` bakes a second rule set (`canSwim=true, cantWalk=true`) producing `WetMask` + `SwimZ_*`. The algorithm composes `effectiveMask = (walkMask & !cantWalk) | (wetMask & canSwim)` per direction; walk Z preferred when both apply.
- **Dynamic-obstacle pass.** Cache fast-path now mirrors `MovementImpl`'s per-cell items + mobiles collision check (`GetItemsAt` / `GetMobilesAt` at the target cell, with `CanOpenDoors` / `CanMoveOverObstacles` / spell-field overrides). This closes a correctness gap from PR #2446 — the cache fast-path was silently skipping dynamic obstacles entirely.
- **`StepCache.TryGetMask` returns `StepMask` struct** instead of 11 out parameters. `HitKind` rolls into the struct with an `IsHit` accessor. Sets up wet/swim without ballooning the call site.
- **`StepChunk.MultiZCells` is lazy-init.** Most chunks are entirely single-Z; allocating the 32-byte bitmap up-front wasted ~256KB at full cap.
- **Admin commands.** `[PathCacheStats` (resident chunks + hit/miss/eviction counters) and `[PathCacheClear` (drop everything, zero counters).
- **Feature flag.** `bitmap_pathfinding_cache` (default true) gates the cache fast-path. Flipped off, every cell expansion routes to `MovementImpl` — equivalent to PR #2446's slow-path-only behavior. Safety net for shipping the new behavior.
`RequiresSlowPath` shrinks to just `CanFly` — flying creatures Z-jump arbitrarily, which the cache's static-Z model can't accommodate.
## Summary
Replaces `FastAStarAlgorithm` with `BitmapAStarAlgorithm`: one cache lookup per cell expansion (8-direction mask + per-direction destination Z) instead of 8 separate `MovementImpl.CheckMovement` calls. Adds the supporting cache infrastructure to back it.
Public API unchanged — `MovementPath` / `Mobile.Move` / `CalcMoves.Find` return the same shapes; the algorithm swap is internal.
## What's in this PR
- **`BitmapAStarAlgorithm`** — A* that issues one `StepCache.TryGetMask` call per cell expansion. Inline fallthrough to the per-cell slow path for multi-Z, off-map, source-Z mismatch, and non-default walkers.
- **`StepCache`** — singleton chunk store keyed by `(mapId, chunkX, chunkY)`. Lazily built on first query, invalidated by `Sector.MultisVersion` mismatch, memory-bounded by sampled probabilistic LRU.
- **`StepProbe`** — computes static-only walkability for a single cell, mirroring `MovementImpl.Check` minus the item / mobile collision phases.
- **`StepMask` / `StepChunk`** — value / storage types for the per-cell results.
- **`CacheEvictionTimer`** — periodic cap backstop (60s interval; early-returns when not over cap).
- **`Map.Sector.MultisVersion`** promoted to `public` so the cache can detect dynamic-static invalidations cheaply.
## Eviction strategy
Sampled probabilistic LRU (Redis-style). Per eviction, sample 5 random keys from a parallel `List<long>` kept in lockstep with the chunk dictionary; evict the oldest of the sample via swap-and-pop. O(1) per eviction regardless of resident count, so sustained cap pressure has no perpetual perf hit.
## Capability handling (interim)
Non-default walkers (non-GM players, creatures with `CanSwim` / `CanFly` / `CanOpenDoors` / `CanMoveOverObstacles`) route entirely through the per-cell slow path via `BitmapAStarAlgorithm.GetSuccessorsSlowPath`. The 2-pass design (cache + capability overlay + dynamic-obstacle pass) lands in the follow-up PR.
Adds an opt-in 'youngPlayerSystem.enabled' server setting (default true) that, when set to false, disables the Young player system server-wide:
- Account.Young and PlayerMobile.Young getters short-circuit to false.
- New characters no longer receive Young status or a NewPlayerTicket.
- All downstream Young checks (notoriety beneficial-action restriction, CheckYoungProtection, stealing penalties, YoungDeathTeleport, death-item movement, poison immunity, '(Young)' name suffix, CanLogout, renounce-young keyword/gump, BaseCreature.OnDeath fame penalty, OnLogin time-remaining message) become inert.
Setters are intentionally left untouched so serialized account/player flags round-trip cleanly when the setting is later re-enabled.
## Summary
Phase 3.3 of the message-interpolation cleanup. Eliminates the `rank.ToString().ToLower()` two-allocation pattern in ConPVP trophy-award messages.
- Adds `TrophyRank.LowerName()` extension returning a static lowercase string per enum value via switch expression.
- Updates 10 call sites across Tournament, KingOfTheHill, DoubleDom, CTF, BombingRun (2 each - cash and no-cash branches).
The handler now appends a static interned string directly into the packet buffer; no `ToString()` formatter and no `ToLower()` allocation per call. Source comment in BombingRun.cs ("There is no formatting flag for Lowercase, we may need a custom interface to get rid of it") is now resolved at the call-site level.
## Summary
Phase 3 PR B of the message-interpolation cleanup. Handles the multi-line restructure sites flagged in `dev-docs/string-handling-message-interp-audit.md` (Phase 2). Phase 3.1 (PR #2436) handled trivial sweeps; this PR handles sites that needed an `if/else` hoist or switch restructure to eliminate `string.Format` while preserving exact message text.
Each site previously allocated an intermediate `string.Format(...)` result before passing to the message handler, despite Phase 1 making the handler accept interpolated string handlers natively.
## Sites fixed
- **`Projects/Server/Mobiles/Mobile.cs:7911`** - Title/guild header was using `string.Format` with a conditional template (`"[{1}]{2}"` vs `"[{0}, {1}]{2}"`). Split into `if (title.Length <= 0)` / `else` with direct `$"..."` interpolation.
- **`Projects/UOContent/Engines/ConPVP/DuelContext.cs:1337`** - View-ladder rank text used `string.Format(text, from == pm ? "You" : "They")`. Split into `if (from == pm)` / `else` with direct `$"..."` interpolation in each branch.
- **`Projects/UOContent/Engines/ConPVP/DuelContext.cs:1463`** - Showladder text reused a single format string for both `LocalOverheadMessage` ("You ... are ...") and `NonlocalOverheadMessage` ("`{pm.Name}` ... is ..."). Each call now uses an inline `$"..."` directly; no shared template.
- **`Projects/UOContent/Engines/ConPVP/Gumps/ConfirmSignupGump.cs:518`** - The signup confirmation message used a `switch` expression assigning a literal format string to `fmt`, then `string.Format(fmt, from.Female ? "Lady" : "Lord", timeUntil)`. Converted to a `switch` statement where each case calls `_registrar.PrivateOverheadMessage(...)` directly with an inline `$"..."`. Lady/Lord branching is hoisted to a `title` local.
## Exempted
- **`Projects/UOContent/Engines/ConPVP/Participant.cs:138`** - The `nonLocalOverhead` format string is a parameter passed in by callers of `Participant.Broadcast`. Investigation found 5 call sites in `DuelContext.cs` (lines 782, 802, 1187, 1196, and three at 1535/1564/1608) that pass distinct literal format strings. Refactoring would require changing all 5 callers and the method signature - out of scope for this PR. Marked with a `// Phase 3 audit:` comment per the audit's exemption convention.
## Summary
Phase 3.1 of the message-interpolation optimization series. Fixes 9 of the 28 sites flagged in the Phase 2 audit (PR #2435):
| File | Fix |
|---|---|
| `Commands/StaffAccess.cs:88,99` | Drop redundant `.ToString()` on enum holes |
| `Commands/Handlers.cs:102` | `builder.ToString()` -> `builder.AsSpan()` |
| `World Saves/SaveCommands.cs:71-75` | Merge 3 concatenated `$"..."` into one literal |
| `Server/Items/Item.cs:4213` | Hoist nested ternary `$"..."` to if/else |
| `Engines/Factions/Mobiles/Guards/BaseFactionGuard.cs:140-150` | Convert switch expression to switch statement |
| `Mobiles/Monsters/LBR/Jukas/JukaLord.cs:85` | Restructure `string.Format(toSay.RandomElement(), ...)` into switch |
| `Misc/AttackMessage.cs:30-41` | Inline `AggressorFormat`/`AggressedFormat` constants |
No functional changes. Each site emits identical text; the only difference is that the message string is now built into a pooled char buffer instead of being allocated as a `string` first.
## Summary
Adds a custom `:L` format specifier to `RawInterpolatedStringHandler`. When the format string is `"L"`, the handler lowercases the formatted value's chars in-place after the underlying `ISpanFormattable.TryFormat` / `IFormattable.ToString` path completes. Zero allocation, single-pass.
## Usage
```csharp
mob.SendMessage($"You earned a {rank:L} trophy!"); // "gold"
mob.SendMessage($"Welcome, {playerName:L}"); // lowercased
mob.SendMessage($"{count:L} kills"); // ints unchanged ("42")
```
## Motivation
Eliminates the `value.ToString().ToLowerInvariant()` two-allocation idiom that appears across the codebase for any type that goes through an interpolation handler. After this lands, content code can use the `:L` specifier directly instead of helper extensions or per-enum lookup tables.
## Coverage
- `AppendFormatted<T>(T value, string? format)` — generic path (covers IFormattable, ISpanFormattable, .ToString fallback)
- `AppendFormatted(ReadOnlySpan<char> value, int alignment, string? format)` — span path with alignment-aware lowercase range (only the value range is lowercased, not padding)
- `AppendFormatted<T>(T value, int alignment, string? format)` and `AppendFormatted(string? value, int alignment, string? format)` and `AppendFormatted(object? value, int alignment, string? format)` — inherit via delegation
The `format == "L"` comparison is case-sensitive — `:l` (lowercase L) is NOT recognized. `:L` matches the convention of e.g. `:N0` / `:F2` (numeric format specifiers traditionally use uppercase). `char.ToLowerInvariant` is used (not locale-dependent) for predictable game text.
## Future cleanup
Phase 3.3 (#2438) introduced a per-enum `TrophyRank.LowerName()` extension to eliminate `rank.ToString().ToLower()` allocations at 10 ConPVP sites. Once this PR lands, those sites can be simplified to `{rank:L}` and the `TrophyRankExtensions` helper can be removed. Tracked as a follow-up.
## Summary
Phase 1 of a multi-phase optimization to eliminate intermediate string allocations between `$"..."` interpolation and the packet text region for ModernUO's player-facing message APIs.
- Adds `[InterpolatedStringHandler]` overloads to every `Send*`/`Public/Local/Private/NonlocalOverheadMessage`/`Say`/`Emote`/`Whisper`/`Yell`/`SendLocalizedMessageTo` API in `OutgoingMessagePackets`, `Mobile`, and `Item`. Each overload is a 3-line shim that forwards `handler.Text` to the existing span-based path then calls `handler.Clear()` to return the rented `STArrayPool<char>` buffer (matches the established `SpanWriter.WriteAscii(ref RawInterpolatedStringHandler)` precedent).
- Converts `string text/args/affix/name` parameters to `ReadOnlySpan<char>` for consistency with the handler path. `lang` intentionally stays `string` (it's never interpolated and the `??= "ENU"` fallback stays cleaner).
- Adds `int charCount` overloads of the three `GetMaxMessage*Length` helpers so stackalloc sizing can avoid the redundant `ROS<char>` round-trip.
- Moves `Mobile` (17 methods) and `Item` (4 methods) message methods into new partial-class files (`Mobile.Messages.cs`, `Item.Messages.cs`) for organization.
No UOContent call sites change in this PR — existing `string`/`ROS<char>` calls compile unchanged via implicit conversion. Phase 2 (intermediate-string audit) and Phase 3 (cleanup PRs) follow.
## Files
- `Projects/Server/Network/Packets/OutgoingMessagePackets.cs` — `string` → `ROS<char>` for text params, `int charCount` length helpers added, class made `partial`
- `Projects/Server/Network/Packets/OutgoingMessagePackets.Interpolated.cs` (new) — 3 `ref RawInterpolatedStringHandler` extension overloads
- `Projects/Server/Mobiles/Mobile.cs` — message methods extracted (-262 lines)
- `Projects/Server/Mobiles/Mobile.Messages.cs` (new, 463 lines) — moved + ROS-converted methods + 25 handler overloads
- `Projects/Server/Items/Item.cs` — message methods extracted (-93 lines)
- `Projects/Server/Items/Item.Messages.cs` (new, 142 lines) — moved + ROS-converted methods + 4 handler overloads
- `Projects/Server.Tests/Tests/Network/Packets/Outgoing/MessagePacketTests.cs` — 3 new regression tests verifying byte-equivalence for the handler overloads
### Summary
- Add InteractiveTeleporter that teleports on double-click
- Add support to decorate command
- Add rope teleporters to the New Haven mines in the decoration file
## Summary
Converts `HouseRaffleManagementGump` from legacy `Gump` to `DynamicGump` with the static `DisplayTo` entry-point pattern.
The gump has paginated entries (up to 10 rows per page), conditional prev/next navigation buttons (vs. inactive image when at page boundary), and per-entry conditional layout (account-bearing vs. raw name). DynamicGump is the right choice.
Constructor is private; `DisplayTo` validates `from` / `NetState` / `stone.Deleted` before constructing. The list+sort runs eagerly in `DisplayTo` so paging math stays consistent across rebuilds.
Builder labels use `$"{value}"` interpolated-string-handler form for zero-allocation text.
Updates the caller in `HouseRaffleStone.ManagementEntry.OnClick`.
## Summary
Converts `RewardGump` and the inner `RewardConfirmGump` from legacy `Gump` to `DynamicGump` with the static `DisplayTo` entry-point pattern.
Both gumps have variable per-instance layout — the reward grid loops over `_rewards` and adds `AddItem(itemID, hue)` / `AddTooltip(tooltipID)` calls whose values are baked into the layout buffer per entry, so a cached static layout would be incorrect. DynamicGump is the right choice and avoids the placeholder dance for non-text values.
Constructors are private; `DisplayTo` validates `NetState`, null `rewards`/`onPicked`, and empty arrays before constructing.
Builder labels use `$"{value}"` interpolated-string-handler form for zero-allocation text.
## Summary
- Converts `HouseGumpAOS` from legacy `Gump` to `DynamicGump` with a private constructor and the static `DisplayTo` entry-point pattern.
- `AddPageButton`, `AddButtonLabeled`, and `AddList` helpers now write directly to the `DynamicGumpBuilder` via `ref` parameters.
- All internal re-display calls and the `HouseSign` / `ConfirmResizeHouseGump` callers route through `HouseGumpAOS.DisplayTo`.
- Field naming updated to underscore-prefix convention (`_house`, `_page`, `_from`, `_list`, `_hangerNumbers`, `_foundationNumbers`, `_postNumbers`, `_houseSigns`).
## Summary
- Converts the legacy pre-AOS `HouseGump` to `DynamicGump` with a private constructor and the static `DisplayTo` entry-point pattern.
- `HouseListGump` and `HouseRemoveGump` now route back through `HouseGump.DisplayTo` instead of constructing the gump directly.
- Updates the `HouseSign` caller accordingly.
Splits BarkeeperGump (DynamicGump) into two StaticGump<T> variants selected by body type — Human (modifiable appearance) and NonHuman (no appearance/gender controls). Each variant gets its own cached static layout via a CRTP base; dynamic per-instance text (rumor messages, keywords, tip message) is filled via slot placeholders in BuildStrings.
Moves PlayerBarkeeper, BarkeeperGump, and BarkeeperTitleGump into a dedicated Mobiles/Vendors/Barkeeper/ folder.
Pulls the Back button on the appearance-categories page out of the ModifyAppearance branch so non-human barkeepers no longer hit a dead end on that page.
BaseCreatures are deleted on death (Mobile.OnDeath calls Delete for non-players), so after save/restart the corpse's _owner reference resolves to null. CorpseNotoriety gated its entire creature branch on `target.Owner is BaseCreature`, falling through to player-corpse logic once the reference vanished. That made monster corpses turn red (body.IsMonster -> Murderer) and innocent NPC corpses turn grey (null is not PlayerMobile -> CanBeAttacked) on the next restart.
Snapshots the relevant owner state into CorpseFlag at corpse creation: OwnerWasBaseCreature, OwnerWasSummoned, OwnerWasAnimatedDead. Folds the standalone _murderer bool into CorpseFlag.Murderer for consistency with Criminal. CorpseNotoriety now consults the flags so the creature branch stays correct without a live mobile reference.
Bumps Corpse serialization to v16 with a MigrateFrom(V15Content) that maps the old Murderer bool onto the new flag. Pre-fix corpses already on disk decay within 7 minutes; their first post-restart color may be wrong, which is acceptable.
Also documents that the schema generator must be run after every version bump (`dotnet tool run ModernUOSchemaGenerator -- ModernUO.slnx`) since `dotnet build` does not emit migration JSON files.
## Summary
Migrates 14 ConPVP lobby/tournament gumps from legacy `Gump` to modern `DynamicGump`/`StaticGump<T>`. Layouts move into `BuildLayout(ref DynamicGumpBuilder)`, constructors become private, and validation moves into static `DisplayTo` entry points (empty-gump rule).
**Per-gump base type decisions:**
- `BeginGump` → `StaticGump<BeginGump>`: layout is fully fixed (no dynamic content). All other gumps below are `DynamicGump` because they bake dynamic player names, guild abbreviations, ruleset titles, arena names, tournament participant names, ladder rankings, or per-instance rule modifications. Per the cliloc/dynamic-text rule, dynamic content forces `DynamicGump`.
- `ReadyGump`, `ReadyUpGump` → `DynamicGump` (per-instance participant rosters).
- `AcceptDuelGump`, `AcceptTeamGump`, `ConfirmSignupGump` → `DynamicGump` (challenger/registrar/team names, dynamic rule modifications).
- `PickRulesetGump`, `RulesetGump` → `DynamicGump` (ruleset titles and option labels per instance).
- `ParticipantGump`, `DuelContextGump` → `DynamicGump` (player rosters/team labels).
- `LadderGump` → `DynamicGump` (ladder entries: ranks, levels, guild abbrs, names, wins/losses).
- `ArenaGump` → `DynamicGump` (arena names with active player names).
- `PreferencesGump` → `DynamicGump` (arena name list).
- `TournamentBracketGump` (~1k LOC) → `DynamicGump`. The whole gump is one type-switched view that re-renders on every button press across `Index`, `Rules_Info`, `Participant_List`, `Participant_Info`, `Round_List`, `Round_Info`, `Match_Info`, `Player_Info`. All branches bake per-instance content.
**Refresh-via-this conversions (the big perf wins):**
- `LadderGump`: page +/- now mutates `_page` and calls `from.SendGump(this)` instead of allocating a new `LadderGump`.
- `PickRulesetGump`: ruleset apply / flavor toggle now refreshes via `this`.
- `ParticipantGump`: increase/decrease team size, remove player, target failure all refresh via `this`.
- `DuelContextGump`: failed-start and add-participant refresh via `this`.
- `ConfirmSignupGump`: every signup-validation rejection branch in `OnResponse` and every `AddPlayer_OnTarget` rejection branch refreshes via `this` (was allocating a new gump per branch).
- `TournamentBracketGump`: every navigation button (back/forward, type change, page change, drill-down) mutates `_type`/`_object`/`_list`/`_page` and refreshes via `this`. Previously each click allocated a new 1k LOC gump.
All gumps are `Singleton`, use private constructors with static `DisplayTo` entry points that null-check `NetState` before allocation. External callers in `DuelContext`, `TournamentBracketItem`, `TournamentController`, `TournamentSignupItem`, and the cross-references between `AcceptDuelGump`/`ParticipantGump`/`AcceptTeamGump`/`ConfirmSignupGump` are all updated to use `DisplayTo`. Legacy `m_X` fields renamed to `_x` per coding standards.
## Summary
Migrates the eight concrete New Guild System gumps (CreateGuild, GuildInfo,
GuildMemberInfo, GuildRoster, GuildDiplomacy, WarDeclaration,
GuildAdvancedSearch, GuildInvitationRequest) and three abstract bases
(BaseGuildGump, BaseGuildListGump, OtherGuildInfo) from the legacy `Gump`
class to `DynamicGump`. Layout work moves from constructor-side `AddX(...)`
calls into `BuildLayout(ref DynamicGumpBuilder builder)` — the abstract
`BaseGuildGump` now provides a `BuildContent` callout for shared
tab-strip chrome, and `BaseGuildListGump<T>` adds another
`BuildListExtras` hook so subclasses can paint highlighted titles after
the filter/sort/pagination chrome.
The headline win is the **self-refresh pattern** on the list gumps and
diplomacy advanced search. Previously each filter/sort/back/forward
click allocated a brand new gump via `GetResentGump`. After migration,
those handlers mutate `_filter`, `_startNumber`, `_comparer`, `_ascending`,
or `_display` on the existing gump and call `from.SendGump(this)`,
letting the singleton path in `NetStateGumps.Send` swap in the same
instance with the new layout. The original list is preserved separately
from the per-render filtered/sorted `_displayList`, so refreshes pick up
the latest state without losing the source list.
All guild gumps deal with per-instance dynamic strings (guild names,
member names, war declarations, alliance names), which would defeat
`StaticGump<T>` caching per the cliloc rule, so every concrete subclass
migrates to `DynamicGump`. `AllianceRosterGump` (in `Misc/Guild.cs`)
is a `GuildDiplomacyGump` subclass and inherits the new behavior; its
unused override and stored alliance reference were dropped along with
the now-obsolete `GetResentGump` abstract.
## Summary
Migrates the Old Guild System (pre-AOS guild stones) gumps from the legacy `Gump` class to `DynamicGump`, following the same pattern used for the Quest gump migration in #2416. All concrete gumps now have private constructors gated by static `DisplayTo` entry points (empty-gump rule), and `Singleton => true` is set across the board so reopening a sibling dialog automatically closes the previous one.
**Migrated gumps:**
- `GuildGump` - main guild dialog
- `GuildmasterGump` - guildmaster functions
- `GuildCharterGump` - charter and website display
- `GuildWarGump` - warfare status (kept as player-facing)
- `GuildWarAdminGump` - war menu (retained as player-facing - reachable from `GuildmasterGump`'s WAR button by guildmasters)
- `GuildChangeTypeGump` - Standard/Order/Chaos selection
**Abstract bases:** `GuildListGump` and `GuildMobileListGump` keep their shared list-rendering chrome inside a single concrete `BuildLayout` on the abstract class and expose a `protected abstract void BuildHeader(ref DynamicGumpBuilder builder)` hook for subclasses (replacing the old `Design()` override). This mirrors the abstract-base treatment used for the ML quest base in the quest-gump migration PR.
**Concrete subclasses migrated alongside the abstract bases:**
- `GuildListGump` subclasses: `GuildAcceptWarGump`, `GuildDeclarePeaceGump`, `GuildDeclareWarGump`, `GuildRejectWarGump`, `GuildRescindDeclarationGump`
- `GuildMobileListGump` subclasses: `DeclareFealtyGump`, `GrantGuildTitleGump`, `GuildAdminCandidatesGump`, `GuildCandidatesGump`, `GuildDismissGump`, `GuildRosterGump`
**Cliloc rule:** Every gump bakes per-instance dynamic content (guild names, member names, war declarations, candidate lists), which would defeat `StaticGump<T>` caching. Per the cliloc rule, all are `DynamicGump`.
**External callers updated:** the prompt files (`GuildAbbrvPrompt`, `GuildCharterPrompt`, `GuildDeclareWarPrompt`, `GuildNamePrompt`, `GuildTitlePrompt`, `GuildWebsitePrompt`), `RecruitTarget`, the `Guildstone` item, and the New Guild System `GuildInfoGump`'s Order/Chaos handler all now go through static `DisplayTo` entry points instead of `new XGump(...)`.
## Summary
Migrates the four ConPVP game board (scoreboard) gumps from legacy `Gump` to `DynamicGump`:
- **`BRBoardGump`** (Bombing Run) — variable layout: row-per-team based on `Participants.Count`. Migrated to `DynamicGump`, `Singleton`, private constructor + `DisplayTo`, `SetNoClose()`.
- **`CTFBoardGump`** (Capture the Flag) — variable layout: row-per-team filtered to only teams with a flag. Same migration shape.
- **`DDBoardGump`** (Double Domination) — variable layout: row-per-team. Same migration shape.
- **`KHBoardGump`** (King of the Hill) — variable layout: row-per-team. `sealed`. Same migration shape.
### Refresh-via-this decision
For all four boards, **score data lives on the `*Game` / `*TeamInfo` objects, not the gump**. The gump just renders a snapshot of those values at the moment it is sent. The three call sites per board are:
1. `OnDoubleClick` on the in-world scoreboard item — one-shot manual open.
2. After death/kill score events (in `OnDeath`) — game logic pushes a fresh board to the dying player so they see updated scores.
3. End-of-game broadcast loop — sends final results to every participant.
None of these are button-driven refreshes from inside the gump, and the gump owns no mutable state. Therefore each event allocates a fresh gump (now via `DisplayTo(...)`) rather than calling `SendGump(this)` on a long-lived instance — that pattern doesn't fit when the data source is external. The win comes from `DynamicGump`'s ref-struct builder writing directly to buffers, eliminating the legacy `GumpEntry` list allocations on every send.
### Mechanics
- `: Gump` → `: DynamicGump`; layout moved from constructor to `BuildLayout(ref DynamicGumpBuilder)`.
- `Closable = false` → `builder.SetNoClose()`.
- Constructors are `private`; static `DisplayTo(Mobile, *Game, ...)` validates `mob?.NetState != null && game != null` before constructing.
- Team-section-mode parameter (`section`) preserved on BR / CTF / DD as an optional `DisplayTo` param even though no current caller uses it.
- The four `m_Game` / similar fields are renamed to `_game`; new fields use `_camelCase` per CLAUDE.md §12.
- `AddBorderedText` / `AddColoredText` helpers became `static` and take `ref DynamicGumpBuilder`.
- Updated all 12 internal call sites (3 per file) to go through `DisplayTo`. No external callers.
- No `OnResponse` was defined on any of these gumps (the only button is a close button), so no `RelayInfo` signature changes were needed.
Touches 4 game files, but only the gump classes — game logic (BR death/scoring, CTF flag handling, DD domination, KH king timer) is untouched.
## Summary
Migrates three legacy `Gump`-based dialogs to the modern builder API:
- **PlayerBBGump** (`Projects/UOContent/Items/Misc/PlayerBulletinBoards.cs`) — `DynamicGump`. The bulletin board renders a different post per page (variable per-instance state), so the layout cannot be cached. Now `Singleton`, with a private constructor and a static `DisplayTo` entry point. Scroll/banish/delete/post-props buttons mutate `_page` and self-refresh via `SendGump(this)` instead of allocating a fresh gump on every click. Prompt-driven flows (post message / set title / post greeting) re-enter through `DisplayTo` after the prompt completes.
- **MessageGump** + **OldMessageGump** (`Projects/UOContent/Items/Skill Items/Fishing/Misc/SOS.cs`) — `StaticGump<T>`. Both render a fixed structure (background + body + button) where only the formatted sextant coordinate string varies per SOS bottle. The cliloc IDs that *appear in the gump packet* are constant (`MessageGump` uses 1018326; `OldMessageGump` uses no `AddHtmlLocalized` at all — its message is pre-formatted into a string before construction), so the layout cache is safe per the cliloc rule. The varying coordinate text is fed through an HTML placeholder via `BuildStrings`. Both gumps are now `Singleton` with private constructors and static `DisplayTo` entry points.
- **ShardPollGump** (`Projects/UOContent/Misc/ShardPoller.cs`) — `DynamicGump`. The gump's structure changes both with the number of poll options (variable loop) and with the `editing` flag (admin sees radios + add-option row + result percentages; players see only radios). The dual-purpose view is preserved as a single `DynamicGump` with the `editing` flag still controlling layout shape — staff path verified to still render the editor with the totals header, vote percentages, and "Create new option" radio. `Closable = false` becomes `builder.SetNoClose()`. Now `Singleton`, with private constructor and a `DisplayTo` that returns the gump instance so `EventSink_Login_Callback` can still call `QueuePoll` on it. The cancel/edit re-issue paths use `SendGump(this)` for self-refresh; the queued login flow uses `DisplayTo` so each queued poll gets its own gump.
External callers (`OnDoubleClick`, `PostPrompt`, `SetTitlePrompt`, `ShardPollPrompt`, `EventSink_Login_Callback`, the Timer-delayed queued poll send) all updated to use the new `DisplayTo` entry points. Legacy `m_X` field naming was already absent in two of the three files; the bulletin board fields kept their `_camelCase` names. `dotnet build Projects/UOContent/UOContent.csproj` reports 0 warnings, 0 errors.
## Summary
Migrates the five-step SoulStone wizard and the TreasureMapChest remove-confirmation dialog from legacy `Gump` to the modern builder API.
Per-gump base type:
- **`SelectSkillGump` -> `DynamicGump`** -- the skill picker iterates the player's skill list and emits one button per non-zero skill, so the layout shape varies per instance.
- **`ConfirmSkillGump` -> `DynamicGump`** -- skill name uses `AosSkillBonuses.GetLabel(...)` which returns dynamic clilocs in the `1044060 + (int)skill` range, plus current/cap skill values rendered as text labels.
- **`ConfirmTransferGump` -> `DynamicGump`** -- same dynamic skill cliloc plus per-instance Base/Cap/Stored values.
- **`ConfirmRemovalGump` -> `StaticGump<ConfirmRemovalGump>`** -- only fixed clilocs (warning text, Continue, Cancel), so the layout caches.
- **`ErrorGump` -> `DynamicGump`** -- title and message clilocs are constructor parameters that vary per call site.
- **`TreasureMapChest.RemoveGump` -> `StaticGump<RemoveGump>`** -- fixed-cliloc confirmation prompt (no item list, despite the name); `Closable=false`/`Disposable=false` are now `builder.SetNoClose()`/`builder.SetNoDispose()`.
All six gumps are now `Singleton => true`, have private constructors, and expose a static `DisplayTo` entry point that validates `from`, `NetState`, and the underlying entity before constructing -- prevents the empty-gump leak. Wizard navigation between steps now goes through `DisplayTo` (e.g. `ConfirmSkillGump.DisplayTo(from, _stone, skill)` from the skill picker, `ErrorGump.DisplayTo(...)` from absorption pre-checks, `SelectSkillGump.DisplayTo(...)` from the "make another selection" button on `ConfirmSkillGump` and from `ErrorGump` bounce-back). Because each gump is Singleton, sending the same type again automatically closes any prior instance instead of stacking; the explicit `gumps.Close<T>()` chain on `OnDoubleClick` is preserved so opening the soulstone still resets any orphaned step from another wizard.
`OnResponse` now uses `in RelayInfo info`. All inline `AddX(...)` calls move to `builder.AddX(...)` inside `BuildLayout`. `Skill.Base.ToString("F1")` etc. are converted to `$"{value:F1}"` interpolation passed to `AddLabel(ReadOnlySpan<char>)`. Skill picker pagination still uses client-side `AddPage` / `GumpButtonType.Page` -- no server-state pagination to migrate.
## Summary
Migrates the Quest system gumps from legacy `Gump` to `DynamicGump` / `StaticGump<T>`.
**Renames** `Engines/ML Quests/Gumps/BaseQuestGump` to **`BaseMLQuestGump`** to
disambiguate from the Core quest abstract (`Engines/Quests/Core/QuestSystem.cs`).
Both abstracts now extend `DynamicGump`.
**Abstract bases**:
- `Server.Engines.Quests.BaseQuestGump` (Core) - now `abstract DynamicGump`. Holds constants and a static `AddHtmlObject(ref DynamicGumpBuilder, ...)` helper. Concrete subclasses each provide their own `BuildLayout`.
- `Server.Engines.MLQuests.Gumps.BaseMLQuestGump` (ML, renamed) - now `abstract DynamicGump` with a `protected abstract BuildContent(ref DynamicGumpBuilder)` hook. The base's `BuildLayout` draws shared chrome (background art, frame, label header) and then invokes `BuildContent`; subclasses use `BuildPage`, `SetTitle`, `RegisterButton`, `SetPageCount`, and content helpers (`AddDescription`, `AddObjectives`, `AddObjectivesProgress`, `AddRewardsPage`, `AddRewards`, `AddConversation`).
**Concrete Core gumps** migrated to `DynamicGump`:
- `QuestCancelGump`, `QuestOfferGump` (Core), `QuestObjectivesGump`, `QuestConversationsGump`, `QuestLogUpdatedGump`, `QuestItemInfoGump`, `SheetMusicOfferGump` (Impresario), `PaintedImageGump` (renamed from `PaintedImage.InternalGump`).
**Concrete ML gumps** migrated to `DynamicGump` (extending `BaseMLQuestGump` or directly):
- `InfoNPCGump`, `QuestConversationGump`, `QuestLogDetailedGump`, `QuestLogGump`, `QuestOfferGump` (ML), `QuestReportBackGump`, `QuestRewardGump`, `QuestCancelConfirmGump`, `RaceChangeConfirmGump`.
**`StaticGump<T>` migrations**:
- `ScrollOfAbraxusGump` (Dark Tides). Its layout is a single hard-coded cliloc (1060116) with no per-instance dynamic content - safe to cache.
**Cliloc rule**: Every other quest dialog bakes per-instance cliloc IDs into its layout (quest titles, NPC names, race-specific prompts, era-conditional progress messages, escort destinations). Per the cliloc rule, baking different cliloc numbers into a cached layout would defeat `StaticGump<T>` caching - so all of these are `DynamicGump`.
**Signature changes** to support the migration:
- `QuestObjective.RenderMessage`/`RenderProgress` now take `ref DynamicGumpBuilder builder` (15 overrides updated across Collector, Solen Matriarch, Ambitious Solen Queen, Study of the Solen Hive, Terrible Hatchlings, The Summoning, Uzeraan Turmoil, Witch Apprentice, Emino's Undertaking).
- ML `BaseObjective.WriteToGump` / `BaseObjectiveInstance.WriteToGump` / `BaseReward.WriteToGump` now take `ref DynamicGumpBuilder` (KillObjective, GainSkillObjective, EscortObjective, CollectObjective, DeliverObjective, BaseReward).
**Empty-gump and `Singleton` rules**: All concrete gumps now have private constructors with static `DisplayTo` entry points that null-check the player NetState before allocation. All gumps that shouldn't stack are `Singleton => true`.
**External callers updated**: `MLQuest.SendOffer`/`OnRefuse`, `MLQuestEntry.SendProgressGump`/`SendRewardGump`/`SendReportBackGump`, `MLQuestSystem.QuestGumpRequest` and `ViewQuestsCommand`, `BoonCollector` (Darius/Nedrick), `SirHelper.OnDoubleClick` (no longer caches a single shared `InfoNPCGump` instance - `DisplayTo` constructs one per click and the gump's `Singleton => true` handles deduplication), `RaceChangeDeed.OnDoubleClick`, `PaintedImage.OnDoubleClick`, `ScrollOfAbraxus.OnDoubleClick`, `Impresario.OnTalk`, and `PlayerMobile`'s `BaseQuestGump` alias is now `BaseMLQuestGump`.
**Concrete subclasses found beyond the listed entry points**: `SheetMusicOfferGump` (in Impresario.cs), `PaintedImageGump` (was `PaintedImage.InternalGump`), `QuestObjectivesGump`, `QuestConversationsGump`, `QuestLogUpdatedGump`, `QuestItemInfoGump` (the Core base has these embedded across `QuestSystem.cs`/`QuestObjective.cs`/`QuestConversation.cs`/`QuestItemInfo.cs`).
**Code-standards cleanup**: Renamed legacy `m_X` private fields to `_x` in rewritten files; braces on all control flow.
## Summary
Migrates five legacy `Gump`-derived UI dialogs in the NPC and Skill domains to the modern `DynamicGump` builder pipeline. All five gumps were chosen as `DynamicGump` rather than `StaticGump<T>` because their layout shape varies per instance, and several of them carry per-instance localization numbers (cliloc IDs) that the static cache cannot bake (see CLAUDE.md gump-system rule and `dev-docs/gump-system.md`).
Per-gump rationale:
- **`TownCrierGump` (`Mobiles/Townfolk/TownCrier.cs`)** - DynamicGump. Announcement count varies, expiration text is rebuilt per render via `ValueStringBuilder`, and one button per entry is emitted in a loop.
- **`ClaimListGump` (`Mobiles/Vendors/NPC/AnimalTrainer.cs`)** - DynamicGump. The pet list and resulting background/alpha-region heights vary per stabling player.
- **`AnimalLoreGump` (`Skills/AnimalLore.cs`)** - DynamicGump. Page count itself varies (3 pages pre-AOS, 5 pages on AOS) and several `AddHtmlLocalized` calls use cliloc IDs computed from per-creature data (loyalty rating `1049595 + c.Loyalty / 10`, food preference, pack instinct), which violates the StaticGump cliloc-bake rule.
- **`DisguiseGump` (`Items/Skill Items/Thief/DisguiseKit.cs`)** - DynamicGump. Page count and entry order shift on `from.Female`, `Body.IsFemale`, and `startAtHair`.
- **`CommentsGump` (`Gumps/CommentsGump.cs`)** - DynamicGump. Comment list and pagination depend on `Account.Comments` size; the title label encodes the variable account username string.
All five are now `Singleton => true`, have `private` constructors, and expose static `DisplayTo(...)` entry points that validate prerequisites before constructing - guaranteeing no empty gumps (CLAUDE.md Sec.13). All internal refresh paths (prompts, `OnDoubleClick`, command handlers, target callbacks) were updated to call `DisplayTo` rather than `new XGump(...)`. Legacy `m_`-prefixed fields renamed to `_camelCase` (CLAUDE.md Sec.12), and `DisguiseEntry`'s `m_`-prefixed public readonly fields converted to PascalCase auto-properties. `OnResponse` signatures updated to `in RelayInfo info`. No external callers needed updating - all `new XGump(...)` sites lived inside the same files.
## Summary
Migrates 10 player-facing legacy `Gump` subclasses for holiday and decorative items to the modern `DynamicGump` / `StaticGump<T>` system. All migrated gumps are `Singleton`, use private constructors gated by static `DisplayTo(...)` entry points (empty-gump rule, CLAUDE.md §13), and replace legacy `m_X` fields with `_x` per coding standards.
Per-gump base type and rationale:
- **Mistletoe.cs** — `MistletoeAddonGump` -> `StaticGump<MistletoeAddonGump>`. Fixed re-deed confirmation layout.
- **StValentinesBears.cs** — Renamed `InternalGump` -> `StValentinesBearsGump`, base `StaticGump<T>`. Fixed sign-bear layout with three text entries; legacy `m_Bear` -> `_bear`. Switched legacy `AddTextEntry(..., size)` -> `AddTextEntryLimited(...)` (modern API split).
- **Wreath.cs** — `WreathAddonGump` -> `StaticGump<WreathAddonGump>`. Fixed re-deed confirmation layout.
- **HolidayPottedPlant.cs** — Renamed `InternalGump` -> `HolidayPottedPlantGump`, base `StaticGump<T>`. Fixed plant-picker layout.
- **SnowStatue.cs** — Renamed `InternalGump` -> `SnowStatueGump`, base `StaticGump<T>`. Fixed statue-picker layout. Dropped unused `Mobile from` ctor arg.
- **TapestryOfSosaria.cs** — Renamed `InternalGump` -> `TapestryOfSosariaGump`, base `StaticGump<T>`. Single image, fixed.
- **HouseRaffleDeed.cs** — `WritOfLeaseGump` -> `DynamicGump`. Description HTML is computed per-instance from deed expiration / days-left, so layout text varies per instance. Added `Singleton => true` (was missing implicitly via legacy default).
- **SpecialScroll.cs** — Renamed `InternalGump` -> `SpecialScrollGump`, base `DynamicGump`. **Cliloc rule**: `_scroll.Message`, `_scroll.Title`, `_scroll.SkillLabel` are dynamic cliloc numbers per scroll type — `AddHtmlLocalized` bakes the cliloc number into the cached layout, so `StaticGump<T>` cannot cache it.
- **BallotBox.cs** — Renamed `InternalGump` -> `BallotBoxGump`, base `DynamicGump`. Layout shape varies: variable topic-line count, owner-vs-voter buttons, and vote-tally bars all add/remove elements. Legacy `m_Box` -> `_box`. Updated the `TopicPrompt` callbacks to call `BallotBoxGump.DisplayTo(...)` instead of allocating a new gump directly.
- **AquariumGump.cs** — `AquariumGump` -> `DynamicGump`. Per-page layout depends on each item's `LabelNumber` and (for `BaseFish`) `GetDescription()` cliloc — those vary per fish/decoration. Two `DisplayTo` overloads (auto-detect access vs explicit edit flag) to mirror the original two call sites in `Aquarium.cs`. Legacy `m_Aquarium` -> `_aquarium`.
### Skipped
- **HouseRaffleManagementGump.cs** — Skipped as **staff-only**. Only invoked via `ManagementEntry` context entry inside `HouseRaffleStone.cs`, gated by `from.AccessLevel >= AccessLevel.Seer`. Per task instructions, staff-only gumps are out of scope for this PR.
### External callers updated
- `Aquarium.cs` — Two `new AquariumGump(...)` sites swapped to `AquariumGump.DisplayTo(...)`.
All other migrated inner classes were nested in the same file and only had local references, which were updated to call the new `DisplayTo(...)` static entry point.
## Summary
Migrates the three player-facing travel/moongate gumps from legacy `Gump` to the modern `DynamicGump` system.
- `GoGump` (Gumps/Go/GoGump.cs) → `DynamicGump`. The category-tree layout's row count varies with the current `GoCategory`'s child count and pagination. Refreshes via `SendGump(this)` after mutating `_node`/`_page` instead of allocating a new instance per nav/page click.
- `MoongateGump` (Items/Misc/PublicMoongate.cs) → `DynamicGump`. The destination tab strip and per-map pages are gated by ruleset (sigil bearer, murderer, faction facet) and expansion/young flag, plus the configured map selection. The set of pages and the active-map swap make the layout shape per-instance.
- `MoongateConfirmGump` (Items/Skill Items/Magical/Misc/Moongate.cs) → `DynamicGump`. Per the **dynamic-cliloc rule**, the gump bakes one of two different cliloc numbers (1062050 Felucca-warning vs 1062049 generic confirm) and selects between an AOS and pre-AOS layout shape — both characteristics force `DynamicGump` rather than `StaticGump<T>` because cached layout bytes would otherwise lock in the wrong cliloc/shape.
All three now use a `private` constructor with a `public static DisplayTo(...)` entry point that validates prerequisites before any gump is allocated (empty-gump rule, CLAUDE.md §13). All are `Singleton` and use `SendGump(this)` self-refresh on internal navigation. Updated callers: `PublicMoongate.UseGate` and `Moongate.BeginConfirmation` now call `DisplayTo(...)`.
## Summary
Second PR in the player-facing legacy gump migration. Converts the four Plants system gumps:
- `MainPlantGump`, `ReproductionGump`, `EmptyTheBowlGump` → `DynamicGump`. Layout varies by plant status, growth stage, health, and pollination/resource availability — cannot use cached `StaticGump<T>`.
- `SetToDecorativeGump` → `StaticGump<SetToDecorativeGump>`. Pure confirmation dialog with no per-instance variation.
All four:
- `Singleton => true` (auto-replace previous plant gump on re-open instead of stacking)
- Constructor `private`; entry is static `DisplayTo(Mobile, PlantItem)` per the empty-gump rule (CLAUDE.md §13)
- `OnResponse` self-refresh paths converted from `from.SendGump(new XGump(_plant))` to `from.SendGump(this)` — saves an allocation on every help/info button click and on every "gather resources/seeds/pollen" action
- Helper draw methods take `ref DynamicGumpBuilder builder` instead of mutating instance state
- Renamed legacy `m_Plant` to `_plant` per CLAUDE.md §12
Updated external callers to use the new entry points:
- `PlantItem.OnDoubleClick` → `MainPlantGump.DisplayTo(from, this)`
- `PlantPourTarget.OnTargetFinish` → `MainPlantGump.DisplayTo(from, m_Plant)` (also drops the now-redundant legacy `singleton: true` flag — `Singleton` property handles it)
- `PollinateTarget.OnTargetFinish` → `ReproductionGump.DisplayTo(from, m_Plant)`
## Summary
First PR in a multi-PR migration of player-facing legacy `Gump` subclasses to the modern `DynamicGump` / `StaticGump<T>` system. Plan covers ~95 files across ~14 PRs by system; this PR is the foundation (smallest, isolated, no cross-refs).
- `VirtueGump` → `DynamicGump`: per-instance virtue hues from `GetHueFor()` and the conditional self/other button block prevent layout caching.
- `VirtueStatusGump` → `StaticGump<VirtueStatusGump>`: layout is identical for every player; only used as a navigation hub.
- `VirtueInfoGump` → `DynamicGump`: dynamic cliloc IDs (`1051000 + (int)virtue`, the description cliloc, and the conditional `1052055`/`1052052` footer) cannot be cached by `StaticGump<T>` — cliloc *numbers* are baked into layout bytes, only HTML/label *text* can be deferred to placeholders.
All three:
- `Singleton => true` (replaces previous instance instead of stacking)
- Constructors made `private`; entry points are static (`RequestVirtueGump`, `DisplayTo`) per the empty-gump rule (CLAUDE.md §13)
- `OnResponse` updated to `in RelayInfo info` modern signature
- `VirtueInfoGump` self-refresh button now uses `_beholder.SendGump(this)` instead of allocating a new instance
- Removed the unused `VirtueGumpItem : GumpImage` nested class — replaced with direct `builder.AddImage(...)` calls; preserves the legacy `class=VirtueGumpItem` attribute for packet parity
The special-cased TypeID for VirtueGump (`BaseGump.cs:86`) is preserved because the type's full name (`Server.Engines.Virtues.VirtueGump`) is unchanged.
## Summary
Removes per-call heap allocations from `Container`'s consume / find / group hot paths and from `BaseCreature.OnDeath`'s fame/karma tracking. The headline wins: kill the `List<List<Item>>` + `Item[][]` + `int[]` grouping bridges in `ConsumeTotal*` / `ConsumeTotalGrouped*` / `GetBestGroupAmount*`, and kill the per-call `Predicate<Item>` allocations in `FindItemsByType(Type)` / `FindItemsByType(Type[])`.
### `Container.cs`
- `ConsumeTotal`, `ConsumeTotalGrouped`, `GetBestGroupAmount` now share four streaming helpers (`HasAmount`, `TryFindGroupMeetingAmount`, `BestGroupTotal`, `ConsumeSlice`) backed by `PooledRefList` instead of allocating per-group lists and jagged arrays. Two-phase validate-then-consume pattern preserved — all-or-nothing semantics for spell reagents, vendor pay, and crafting still hold.
- `(Type)` / `(Type[])` / `(Type[][])` overload trios collapsed to single `ReadOnlySpan<Type>` + `ReadOnlySpan<int>` implementations. Implicit `T[] → ReadOnlySpan<T>` conversion means UOContent callers compile unchanged.
- Unused overloads deleted: `ConsumeTotalGrouped(Type)`, `ConsumeTotalGrouped(Type[][])`, `GetBestGroupAmount(Type)`, `GetBestGroupAmount(Type[][])`, plus the never-called `TryDropItems` hook and its private `ItemStackEntry` struct.
- Fixes a `PooledRefList` leak in `GetBestGroupAmount(Type[], …)` (missing `using`).
- `m_ContainerData` / `m_Items` / `m_TotalGold` / `m_TotalItems` / `m_TotalWeight` / `ContainerData.m_Table` / `ContainerData.logger` renamed to the underscored convention. `m_Items` cross-file rename for the Container-side references in `Item.cs`; `Item.CompactInfo.m_Items` deliberately left alone (separate effort).
- `CheckHold` parent walk simplified; trivial dispatch methods (`CheckHold` overloads, `OnItemAdded`, `OnItemRemoved`, `OnStackAttempt`) get `[MethodImpl(AggressiveInlining)]`; `Destroy` and `DisplayTo` cache `Items` outside the loop; dead comments removed.
### `Item.Enumerable.cs`
- `FindItemsByType(Type)` previously allocated a `Predicate<Item>` per call (method-group conversion). `FindItemsByType(Type[])` allocated a closure capturing `types`. Both now construct the enumerator with a `Type` / `ReadOnlySpan<Type>` field directly, no delegate.
- `FindItemsByTypeEnumerator<T>` gains two constructors plus a `Matches(T)` helper that picks the right filter inline. Constructor chaining via a private 2-arg seed constructor incidentally fixes a pre-existing bug where `PooledRefQueue` was always rented at capacity 0 because `_recurse` hadn't been assigned yet.
- `(Type[])` overload of `FindItemsByType` becomes `(ReadOnlySpan<Type>)`.
- `EnumerateItemsByType(Type)` / `EnumerateItemsByType(ReadOnlySpan<Type>)` / `ListItemsByType(Type)` / `ListItemsByType(ReadOnlySpan<Type>)` simplified to delegate to the new alloc-free overloads instead of filtering manually.
### `Utility.cs`
- `InTypeList<T>(this T, Type[])` and `InTypeList(this Type, Type[])` switched to `ReadOnlySpan<Type>`.
### `BaseCreature.cs`
- `OnDeath` per-death `List<Mobile>` / `List<int>` / `List<int>` for fame/karma tracking switched to `PooledRefList`.
## Summary
Overhauls the fame and karma system to be more era-accurate, based on original design documents and publish notes.
### Karma on player kill → karma on murder report
- Removes karma gain/loss on player kill (was immediate on death)
- Karma is now set to `Kills * -1000` on murder **report** instead
- Fame on player kill now uses the same formula as monster kills (`Fame / 100`)
This karma loss on murder report behaviour was tested on both the demo and live servers, behaviour was matching in terms of karma loss on report. Official UO servers karma loss AMOUNT match with my memory of T2A/UOR with one caveat - it's doubled on live servers (-2000 * kill count). I'm not sure when this changed and this behaviour has always had very poor and incorrect documentation, even 10-20 years ago. I was obsessed with the dread lord title on OSI and the only way I knew how to get it was reach 10 kills then macro them off. Even in publish 16 (when "The Murderer" title was removed) it still required 10 kills. Maybe it changed to -2000*Kills in AOS - that's where I've put the era gating diff.
### Era gates
- **Karma lock** (ankh toggle + auto-lock on negative karma) gated to `Core.UOTD && !Core.AOS` — [didn't exist before Jan 28, 2001](https://web.archive.org/web/20010128092700/http://update.uo.com/design_300.html)
- **Felucca fame/karma +30% bonus** gated to `Core.LBR` — [added in Publish 16, July 2002](https://uo.com/wiki/ultima-online-wiki/technical/previous-publishes/2002-2/publish-16-part-2-5-23rd-july/)
- **Fame/karma splitting** among damage dealers gated to `Core.UOR` — pre-UOR awards go to last hit only
### Skill karma penalties
- **Provocation** on innocent NPC: NPC says cliloc 501591, karma loss (floor -7500)
- **Stealing** attempt: karma loss on every attempt (floor -5000). Stealing did not cause karma loss at all before!
- **Summon Daemon**: karma loss on successful cast (floor -7000)
- **Corpse carving** (human): -70 for innocent corpses (floor -7000), -20 for freely-aggressable (floor -2000)
- **Bounty head turn-in**: karma gain capped at 2000 (was awarding flat +2000)
### Beneficial action karma
- **Beneficial spells** (heal, cure, etc.): `AwardKarma(caster, target.Karma / 5)` — healing good targets raises karma, healing evil targets lowers it - source is UO98 demo scripts
- **Bandages**: same formula but gain only (skipped if target karma ≤ 0) - see stratics link ("only ever gain karma, not lose it")
Sources: UO98 Demo scripts and playing, [Fame and Karma wiki](https://uo.com/wiki/ultima-online-wiki/player/fame-and-karma/), [UO design doc (Jan 2001)](https://web.archive.org/web/20010128092700/http://update.uo.com/design_300.html), [Publish 16](https://uo.com/wiki/ultima-online-wiki/technical/previous-publishes/2002-2/publish-16-part-2-5-23rd-july/), [Stratics healing reference](https://web.archive.org/web/20001209014200fw_/http://uo.stratics.com/heal.shtml)
### Refactoring
- Extracted `Titles.ComputeKillAwards(killed, map)` shared by player kill and creature kill paths
- Extracted `Titles.SetKarma(m, value, message)` for direct karma assignment (used by murder report)
- Extracted `SendKarmaMessage` and `CheckKarmaLock` helpers from `AwardKarma`
## Summary
- Fixes pets falling behind mounted masters in AOS+ by setting `CurrentSpeed = 0.1` when following master
- Fixes AI timer permanently stopping when `Obey()`/`Think()` returns `false` for transient conditions
- Fixes controlled pets losing AI in inactive sectors (pet follows owner across sector boundary, sector deactivates, AI dies)
- Adds defense-in-depth: AI timer restarts on pet resurrection and order changes
## AI Timer Permanent Stop (Bug Fix)
`AITimer.OnTick()` called `Stop()` when `Obey()` or `Think()` returned `false`. By that point, `ShouldStop()` had already validated the creature is alive, on a valid map, and in an active sector — so any `false` return was a **transient** condition, not terminal. The timer stopped permanently with no mechanism to restart it.
**Scenarios that triggered permanent AI death:**
- Dead bonded pet with attack order (`DoOrderAttack` returned `false` for `IsDeadPet`)
- Failed pet transfer — loyalty refusal, combat, disconnected player, or pending trade (`DoOrderTransfer` returned `false` for 5 different transient conditions)
- Unknown `OrderType` or `ActionType` (defensive defaults)
**Fixes:**
- Removed `Stop()` from the `Obey()`/`Think()` failure path — timer skips the tick and fires again next interval
- Changed `DoOrderAttack()` and all five `DoOrderTransfer()` failure paths to return `true` (correct semantics: these are recoverable states, not "stop AI forever" signals)
- Added `Activate()` call in `ResurrectPet()` — ensures dead bonded pets have AI running after resurrection
- Added `Activate()` call in `OnCurrentOrderChanged()` — self-heals timer if any voice command is issued to a pet with a stopped timer
## Controlled Pet Sector Deactivation (Bug Fix)
`ShouldStop()` stopped the AI timer for **all** `PlayerRangeSensitive` creatures in inactive sectors, including controlled pets. But `Deactivate()` intentionally exempted controlled pets. The exemption was dead code — `ShouldStop()` bypassed it.
This matters when a pet follows its owner across a sector boundary: the owner enters the next sector (active), the pet's old sector deactivates (no more players), and the pet's AI dies. The pet stops following and stands there until the player backtracks far enough to reactivate the sector.
**Fix:** Added `Controlled` check to `ShouldStop()` to match `Deactivate()`. Controlled pets now keep their AI running in inactive sectors. The overhead is negligible — controlled pets are bounded by follower slots.
## Movement Speed Simplification
- Simplifies `AITimer` to use `CurrentSpeed` directly as the tick interval (in seconds), removing the complex multiplier/floor logic in `GetBaseInterval`
- Refactors `DoMoveImpl` speed assignment into explicit if/else for clarity
- AOS+ pets following master use `CurrentSpeed = 0.1` (100ms), matching `RunMountDelay`
## Files Changed
- `AITimer.cs` — removed `Stop()` on Obey/Think failure, added `Controlled` exemption to `ShouldStop()`, simplified interval logic
- `BaseAI.cs` — renamed `_timer` to `AITimer` (public), simplified `Deactivate()`, fixed `ReturnToHome` to use `Activate()`
- `PetOrders.cs` — `DoOrderAttack` and `DoOrderTransfer` return `true` for transient failures
- `PetOrderHandlers.cs` — `OnCurrentOrderChanged()` calls `Activate()` to self-heal stopped timers
- `BaseCreature.cs` — `ResurrectPet()` calls `Activate()`, fixed `GoHome_Callback` PlayerRangeSensitive check
- `AIMovement.cs` — refactored speed assignment, AOS+ follow-master speed fix
