## Summary
- Adds the Endless Decanter of Water (introduced in Publish 66.2 / SA era)
- Players throw a full Pitcher of Water at a Water Elemental for a 10% chance to receive the decanter; the pitcher is always destroyed on impact
- Each Water Elemental can only yield one decanter; the state is persisted and previously saved elementals are migrated to the new serialization version
- The decanter auto-refills from a linked water trough when the owner empties it within 10 tiles of the stored trough location
- Linking stores a Point3D + Map snapshot, supporting both static tile troughs and addon troughs
- The decanter is blessed and displays Linked/Unlinked status in its tooltip
## Summary
- **Timer-aware idle sleep**: Exposes `Timer.MillisecondsUntilNextTick()` to calculate remaining ms until the next timer wheel tick (0–8ms). The game loop sleeps for that duration minus a 1ms safety margin, instead of spinning at 100% CPU.
- **I/O completion wakeup**: Replaces `Thread.Sleep` with `NetState.WaitForCompletion()`, which uses platform-native completion notification (RIO `RIONotify` on Windows, `eventfd` on Linux, `kevent` timeout on macOS) to wake immediately when network data arrives during sleep.
- **Always-on**: Removes the debug-only `core.enableIdleCPU` config gate. The sleep is self-regulating — under load, `MillisecondsUntilNextTick` returns 0 so no sleep occurs (zero overhead). On idle, CPU drops from ~100% to ~1%.
- **CPS calculation cleanup**: Replaces the 128-element ring buffer with an EMA (exponential moving average) for `CyclesPerSecond`/`AverageCPS` — fewer allocations, no LINQ `.Average()` call each sample.
- **IORingGroup 1.0.6**: Adds `WaitForCompletion(int timeoutMs)` to the `IIORingGroup` interface with platform implementations:
- **Windows**: `RIONotify` arms the CQ event, `WaitForSingleObject` with timeout
- **Linux**: `eventfd` registered with io_uring, `poll()` with timeout
- **macOS**: `kevent()` with timeout
## Test plan
- [ ] Build succeeds on all platforms (`dotnet build`)
- [ ] Empty server: verify CPU usage drops from ~100% to ~1% idle
- [ ] Loaded server: verify no added latency — `MillisecondsUntilNextTick` returns 0 when timers are firing, sleep is skipped
- [ ] Connect a client during idle — verify connection accepted within one timer tick (~8ms)
- [ ] Verify `[admin` gump shows reasonable CPS values (EMA convergence)
## Summary
Fixes three bugs in `GuardedRegion.CallGuards`:
- **Operator precedence bug**: The condition `!m.Region.IsPartOf(this) && !m_GuardCandidates.ContainsKey(m)`
was inverted from intent. Replaced with explicit split: dictionary members are targeted regardless of
region; permanent candidates (reds/AlwaysMurderer) must be inside the region.
- **Premature `break`**: Only the first guard candidate was ever processed per "guards" call.
Removed the `break` so all valid candidates in range get a guard spawned.
- **Misleading message for permanent reds**: "Guards can no longer be called on you." (502276)
was sent to permanent reds, but guards can *always* be called on them. Now only sent to
temporary criminals whose guard window is actually consumed.
Also extracts `IsAlwaysGuardCandidate()` helper and simplifies `IsGuardCandidate()`.
## Edge cases verified (by analysis)
- **Multiple players call guards on same red**: `BaseGuard.Spawn` dedup (scans 15 tiles for
existing guard with same `Focus`) prevents duplicate guards. Message suppression eliminates spam.
- **Red fights spawned guard → criminal → guards called again**: Same dedup prevents infinite
guard spawns. Existing guard already has `Focus == red`.
## Test plan
- [ ] Temporary criminal in guarded region → guard spawns, receives "Guards can no longer be called on you."
- [ ] Permanent red (5+ kills) in guarded region → guard spawns, does NOT receive the message
- [ ] Multiple players call "guards" on same red → only 1 guard spawns
- [ ] Criminal outside region but in dictionary → still targeted by guards call from inside region
- [ ] Red outside guarded region → NOT targeted by guards call (must be inside region)
## Summary
- Adds `Mobile.Murderer` virtual property and consolidates kill-threshold checks across the codebase
- Tracks ping-pong count: how many times a player crosses the 5-kill murderer threshold (T2A/UOR/UOTD only, disabled on LBR+)
- Adds `[CommandProperty]` to view a player's ping-pong count via the admin panel
- After enough ping-pongs, player is permanently flagged as a murderer regardless of kill count
- Accounts for perma-red players with low kills in murderer status transition notifications
- Implements era-appropriate "I must consider my sins" speech responses:
- **T2A**: contextual cliloc flavor text (502122–502126)
- **UOR–AOS**: raw short/long-term murder counts + ping-pong count if applicable
- **SE+**: localized stats message (1114370)
- Refactors kill-report logic out of `Keywords.cs` into `PlayerMurderSystem.ReportKillsToSelf`
## Testing
- [x] Thoroughly tested and self reviewed
- [x] Test T2A "I must consider my sins" behaviour over all scenarios.
- [x] Test UOR "I must consider my sins" behaviour over all scenarios.
- [x] Test that LBR does not have ping pongs enabled (I must consider my sins)
- [x] Test serialization cross over from v0 -> v1 increments 1 ping pong if player is already red.
## Notes
* Manually setting kills to 5 does not trigger a ping pong, it must go through the actual murder system. This includes if the kills were manually set to 5 and then migrated (as manually setting kills to 5 never adds the player into the murder system - it only happens via ReportMurderer). This is arguably a bug in the existing system, but one that currently only ever happens via staff interaction.
* Thieves guild SuspendOnMurder specifically checks for kills > 0. This means a person with 0 shorts but 5 ping pongs (flagged as murderer) can steal. This may be accurate, as according to a forum post this is how it works on UOSA which is the T2A gold standard.
* This doesn't implement Pre-T2A behaviour which should be that "I must consider my sins" does nothing at all. The reason I didn't implement it for Pre-T2A is then it 100% have to sit behind a feature flag. I don't mind adding it as a feature flag, just let me know.
## Summary
Server-side movement throttle that prevents speed hacking while accurately identifying cheaters with detection of lagging connections.
**Key features:**
- Credit buffer (200ms) absorbs timing jitter from legitimate players
- Movement queue handles larger bursts, draining at proper game-tick intervals
- RTT measurement distinguishes network lag from speed hacks
- Queue depth detection catches ACK-throttled speed hacks (going straight)
## How It Works
**Throttle** (prevention): Movements arriving too early either consume credit or get queued. The queue drains at
correct intervals, so speed hackers can't move faster regardless of what they send.
**Detection** (identification): Combines multiple signals to identify cheaters:
| Signal | What it catches |
|--------|-----------------|
| Queue depth ≥4 sustained | ACK-throttled speed hacks (client limits unacked moves to 5) |
| Movement rate >1.05x | Direction-change speed hacks where timing is visible |
| Stable RTT + high queue | Eliminates false positives from laggy players |
**RTT-Aware Logic:**
- Probes only sent to players actively moving (event-driven, not global loop)
- Stable low-latency + problems = suspicious
- Unstable/high-latency + problems = probably just lag, throttle handles it
## Configuration
```json
{
"movementThrottle.maxCredit": 200,
"movementThrottle.softQueueLimit": 6,
"movementThrottle.hardQueueLimit": 10,
"movementThrottle.debugLogging": false
}
```
## Summary
- Adds `public virtual bool Murderer => Kills >= 5` property to `Mobile`, replacing ~30 scattered `Kills >= 5` / `Kills < 5` magic-number checks across the codebase
- `BaseCreature` overrides `Murderer` to also return `true` when `AlwaysMurderer` is set
- Updates `Corpse` serialization to v15, storing `_murderer` as a `bool` field (migrated from `int Kills >= 5` in earlier versions)
### Summary
Updates all calls to container.EnumerateItems() to properly dispose of the underlying PooledRefQueue so that we are properly recycling pooled arrays.
### Summary
- Bump IORingGroup 1.0.0 → 1.0.1 — fixes a disconnect handling bug in the native ring layer
- Fix ghost NetStates — Dispose() set _running = false before checking it, so the "force immediate disconnect" path
was dead code. Capture wasRunning before clearing it, add [Obsolete] guard, and route internal callers through
DisposeInternal()
- Fix unauthenticated socket cleanup — graceful disconnect on unauthed connections could get stuck with pending sends;
now force-immediate after Disconnect() if DisconnectPending is already set
- Replace ConcurrentQueue<NetState> _disposed with Queue<NetState> — server is single-threaded; moved the field into
the Network partial class where it's consumed
- Move ConnectingSocketIdleLimit into the Network partial class alongside DisconnectUnattachedSockets
- Reset activity timer on receive, not just send — receiving data directly proves liveness instead of relying on the
ping→pong→send round-trip to reset the timer
- Move CheckAllAlive from Timer into Slice — the timer fired before I/O completions were processed, so after server
stalls (world saves), buffered client pings hadn't reset timestamps yet, causing false disconnects. Now runs at the
end of Slice() after all recv completions are handled
- Lower inactivity timeout 90s → 30s, check interval 90s → 5s — clients ping every ~1s, so 30s of silence is ~30
missed pings; worst-case detection drops from ~180s to ~35s
- Simplify CheckAlive — early-return when socket is null or alive; force-kill stuck DisconnectPending sockets
immediately instead of calling Disconnect() again
- Remove unused imports from GameEncryption.cs
### Summary
- Fix AccountGold gold duplication exploit: When AccountGold.Enabled was true, double-clicking a BankCheck deposited
the full value to the account but then continued creating physical gold piles for the same amount, duplicating the
value
- Fix Deposit/DepositUpTo partial deposit exploit: Both methods created new max-size gold piles and checks without
first filling existing partial stacks, wasting container slots and causing premature "bank full" failures that could
be leveraged to manipulate gold distribution
- Fix BagOfSending bank stacking: Gold and BankCheck items sent via BagOfSending now use Banker.Deposit for efficient
stacking instead of naive TryDropItem, which could fail on a full bank even when existing piles had room
- Improve gold/check deposit efficiency: Banker.Deposit and Banker.DepositUpTo now top off existing gold piles (up to
60k) and bank checks (up to 1M) before creating new items, maximizing use of available container slots
### Summary
Changes component verification so the order is now housing.bin, then txt files in client, then txt files in Data/Components folder on the server.
> [!IMPORTANT]
> **Breaking Changes**
> - DecodePacket and EncodePacket delegates replaced with IClientEncryption interface
> - NetState.Connection (Socket) replaced with internal RingSocket management
> - NetState.RecvPipe and NetState.SendPipe removed (buffers managed internally)
## Summary
Upgrades the networking stack from PollGroup-based I/O to io_uring, significantly improving I/O performance on Linux.
This also adds native client encryption support for encrypted UO clients.
## Major Changes
io_uring Networking Architecture
- Replaced PollGroup with IORingGroup for async socket I/O operations
- Removed Pipe.cs (mirrored ring buffer) and TcpServer.cs in favor of RingSocketManager
- Added NetState.Network.cs - centralized network infrastructure handling accept, recv, send, and disconnect
completions
- Added SocketHelper.cs - platform-specific socket utilities for raw socket handle operations (getpeername,
getsockname)
- Buffer management now handled by RingSocketManager with configurable slab allocation
### Client Encryption Support
- Added full encryption stack in Network/Encryption/:
- EncryptionConfig.cs - configurable encryption modes (None, Unencrypted, Encrypted, Both)
- EncryptionManager.cs - encryption detection and initialization for login/game packets
- LoginEncryption.cs - handles login packet encryption with version-derived keys
- GameEncryption.cs - handles game server encryption using Twofish
- TwofishEngine.cs - optimized Twofish block cipher implementation
- LoginKeys.cs - encryption key table for client versions
- IClientEncryption.cs - interface for client encryption implementations
### NetState Improvements
- Replaced Socket Connection with RingSocket _socket for managed socket lifecycle
- Changed from GCHandle polling to event-based completion processing
- Disconnect handling now properly waits for pending sends to flush
- Simplified connecting socket management using lazy queue removal
### Configuration
- New settings: network.encryptionMode and network.encryptionDebug
- Encryption mode flags: Unencrypted, Encrypted, or Both
### Dependencies
- Replaced PollGroup NuGet package with IORingGroup
- Linux requires liburing-dev / liburing-devel package
### Test plan
- Verify server starts and accepts connections on Linux with io_uring
- Verify server starts and accepts connections on Windows (fallback to IOCP)
- Test unencrypted client connections (ClassicUO with encryption disabled)
- Test encrypted client connections if available
- Verify graceful disconnect flushes pending data
- Confirm CI builds pass on all target platforms
## Summary
- Adds proper Latin1 encoding support, replacing CP1252 usage throughout the codebase
- Adds specialized, optimized string decoding methods with safe string filtering for each encoding type
- Filters invalid Unicode characters (C0/C1 control codes, non-characters) by removal rather than replacement since
the UO client renders nothing for these characters
- Fixes UTF-16 null terminator position handling to correctly advance by 2 bytes
## Changes
TextEncoding.cs
- Added SearchValues-based invalid byte/char detection for efficient filtering
- Added encoding-specific GetString methods: GetStringAscii, GetStringLatin1, GetStringUtf8, GetStringBigUni,
GetStringLittleUni
- Each method supports a safeString parameter for filtering invalid characters
- Little-endian UTF-16 uses direct memory cast for zero-copy decoding on LE systems
- Invalid characters are removed (not replaced with U+FFFD) since the client renders nothing for them
SpanReader.cs
- Added ReadLatin1() and ReadLatin1Safe() methods
- Rewrote encoding-specific read methods to use optimized TextEncoding.GetString* methods
- Fixed UTF-16 null terminator handling: position now correctly advances by byteLength (2) instead of 1
SpanWriter.cs
- Added WriteLatin1 and WriteLatin1Null methods
## Packet Updates
- Updated all packet code to use Latin1 encoding instead of CP1252
- Affected: account packets, equipment packets, menu packets, message packets, mobile packets, player packets, secure
trade packets, vendor packets, gump packets, book packets, mahjong packets
## Filtering Behavior
Invalid characters filtered in safe mode:
```
┌───────────────┬────────────────────────┐
│ Range │ Description │
├───────────────┼────────────────────────┤
│ 0x00-0x1F │ C0 control codes │
├───────────────┼────────────────────────┤
│ 0x7F │ DEL │
├───────────────┼────────────────────────┤
│ 0x80-0x9F │ C1 control codes │
├───────────────┼────────────────────────┤
│ 0xFFFE-0xFFFF │ Unicode non-characters │
└───────────────┴────────────────────────┘
```
Note: Surrogate pairs (0xD800-0xDFFF) are not filtered because proper validation requires context checking for paired
vs unpaired surrogates. The UO client renders nothing for these anyway.
## Test Plan
- All 631 Server.Tests pass
- Verified client rendering behavior using TestUnicodeGump command (pages 1-5)
- Confirmed U+FFFD, unpaired surrogates, and non-characters all render as blank in client
- Verified Latin1 characters (0xA0-0xFF) display correctly
- Verified C1 control codes (0x80-0x9F) are filtered and don't display
### Summary
* Fixes migration a socket from one poll group to another
* Fixes epoll/wepoll incompatibility
* Updates minimum support to Windows 10+ (Server 2019+)
* Enhances performance by supporting synchronous IOCP completions when available (up to 30% increase in performance).
Fix [add command failing with ambiguous type names + refactor for performance
### Problem
[add blight would fail with "No type with that name was found" because multiple types contain "blight" (e.g., Server.Items.Blight, Server.Ethics.Evil.Blight, Server.Items.BlightGrippedLongbow, Server.Items.QuiverOfBlight). The old code only succeeded when exactly one type matched the search regardless of constructability and inheriting Mobi les/Items.
### Solution
Exact match takes priority: If a type's name exactly equals the search string (case-insensitive), use it directly. Otherwise, show the AddGump with all partial matches.
- [add blight → Creates Blight (exact name match)
- [add bligh → Shows gump with Blight, BlightGrippedLongbow, etc.
### Refactoring
- CommandEventArgs context: Added GetContext<T>/SetContext<T> to pass resolved type through the command chain without method signature changes
- Removed TrySetupTarget duplication: Validation now happens only in ValidateArgs, eliminating redundant code paths
- Split type matching:
- ExactMatch(string) → Returns Type for exact name match (used by [add)
- MatchEmptyCtor(string) → Returns ConstructorInfo[] for gump display (empty-callable constructors only)
### Memory & Performance Improvements
| Optimization | Benefit |
|-------------------------------|----------------------------------------------------------------------------------------------|
| _mobileItemTypes cache | Filters Mobile/Item types once per assembly, reused on all subsequent searches |
| ReadOnlySpan<string> for args | Avoids string[] heap allocations when slicing arguments |
| ValueStringBuilder | Stack-allocated string building, avoids StringBuilder heap allocation |
| Single type resolution | Type resolved once in ValidateArgs, passed via context to Execute (was resolved 2-3x before) |
