🤖 Agent: Proven in union R1 (branch union/cg19-r1, main + 83 of kacpherek's PRs): the 98 harnesses it can affect, clippy -D warnings, tests, lan_e2e in-frame and on the server thread.
🤖 Agent: Proven in union R1 (branch union/cg19-r1, main + 83 of kacpherek's PRs): the 98 harnesses it can affect, clippy -D warnings, tests, lan_e2e in-frame and on the server thread.
🤖 Agent: Merged by Ludovic's AI agent (Claude, contrib-gate-19). Thanks @kacpherek: Open the command block screen on a server's client and send its edits. I checked it against vanilla 26.3, and it passed every harness it reaches in a union with your other PRs, plus lan_e2e in both modes.
🤖 Agent: verified by Ludovic's AI agent (contrib gate 20): recipe book settings persist on the server. It passed in this wave's union with main on the Mac mini: the selected parity harnesses (332/332, then 122/122 on the final union), clippy -D warnings and the tests, and lan_e2e with the host on its server thread and in-frame (18/18 each). Thanks @rdbtCVS!
🤖 Agent: verified by Ludovic's AI agent (contrib gate 20): pause menu server links. It passed in this wave's union with main on the Mac mini: the selected parity harnesses (332/332, then 122/122 on the final union), clippy -D warnings and the tests, and lan_e2e with the host on its server thread and in-frame (18/18 each). Thanks @socksprox!
Using a compass on a lodestone now sets its lodestone target (in place for a single compass in survival, otherwise one is split off), and the target clears when the lodestone is gone. The target also travels in item packets on the server thread.
Send open/filter changes to the authoritative player's recipe book and cover packet settings plus save/reopen in both single-player modes.
Co-Authored-By: OpenCode GPT-6.1 Sol <noreply@opencode.ai>
ClientboundServerLinksPacket (configuration and play) decoded and kept,
the pause menu gains its "Server Links..." button when a server sends
links, opening the server_links dialog screen: scrollable link buttons,
ConfirmLinkScreen on click (chatLinks/chatLinksPrompt honoured), and the
custom-screens warning with its disconnect. A dedicated server sends its
`bug-report-link` property as the BUG_REPORT link, like vanilla.
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
🤖 Agent: Gate check by Ludovic's AI agent (Claude). `tools/bundle_tooltip_e2e.py` passes in-frame. On the server thread (what players run) it showed every bundle full, holding more bundles, so I pushed two commits: (1) BundleContents' stream codec (component 52, a list of ItemStackTemplate, checked with javap) plus sending it to clients in `network_stack`; (2) `Hub::stack` gave every stack inside an item the outer item's id, which also turned a crossbow's charged projectiles into crossbows on clients. It is fixed now: each inner stack looks up its own id. With both commits the e2e passes on the server thread and in-frame (306 px tooltip frame; main shows 117 px and fails). The screenshot shows 32 stone with a half-full bar, as in vanilla. Unit tests: `bundle_contents_are_item_stack_templates` and `a_bundle_keeps_its_contents_across_the_wire`. Thanks @levzzz5154!
BundleContents had no stream codec here: the host never sent what a bundle
holds, so on the server thread, LAN and a dedicated server the new tooltip
showed every bundle empty, and a vanilla server's bundle made its packet
undecodable. Its STREAM_CODEC is a list of ItemStackTemplate, the same as
ChargedProjectiles' (component 52).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A Horizon client sends a client mod's batch of command lines as one
`horizon:quiet_commands` payload, which Horizon's server announces only to
Horizon clients. Each line runs exactly as the player typing it would
(the same checks, their own permission level), with vanilla's suppressed
output: no feedback and no failure in anyone's chat. Batches wait in a
queue: at most 64 lines run per player per tick, and at most 4096 wait.
Each finished batch is one line in the server's log.
tools/quiet_commands_e2e.py checks it on the dedicated server: a non-op's
quiet setblock is refused, an op's runs, and nothing reaches the chat.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Thanks socksprox! Mannequins render as the player model with their profile skin. I pushed d3af1e04e: a mannequin is a LivingEntity, not a Mob, so it runs no goals and never despawns. Checked in game: a summoned mannequin stays put for 400 ticks on the server thread and falls to the ground in-frame (it wandered 14 blocks before). Tested together with the other wave 2 contributor PRs: fmt, clippy -D warnings, and the horizon-client, horizon-world and horizon-protocol tests pass. Of the 127 harnesses the batch affects, 124 pass; the other three are a player-head mismatch fixed in #1934 and two worldgen timeouts on a loaded machine with no chunk differing. lan_e2e chat, fireball, cursor, looks, thrown, eat, shield, lava, contacts, projectiles and mobs pass with the host in-frame and on its server thread.
🤖 Agent: Thanks levzzz5154! Matches ItemCooldowns and the slot sweep in Gui.renderSlot; thrown, eat and shield pass on both hosts. I merged main and dropped orphaned doc comments. Tested together with the other wave 2 contributor PRs: fmt, clippy -D warnings, and the horizon-client, horizon-world and horizon-protocol tests pass. Of the 127 harnesses the batch affects, 124 pass; the other three are a player-head mismatch fixed in #1934 and two worldgen timeouts on a loaded machine with no chunk differing. lan_e2e chat, fireball, cursor, looks, thrown, eat, shield, lava, contacts, projectiles and mobs pass with the host in-frame and on its server thread.
Servers on 26.3 spawn NPCs as minecraft:mannequin — the avatar entity
vanilla added for exactly this — instead of fake players (gommehd.net's
"Gommunity" lobby NPCs are invisible today). The client never registered
the type: MobTypes.load only builds species::ALL, so AddEntity produced a
tracked entity but nothing to draw.
- EntityDataSerializers.RESOLVABLE_PROFILE (41) decodes into
Value::Profile. Before, the entry list ended at it and the mannequin's
DATA_IMMOVABLE and DATA_DESCRIPTION after it never decoded.
- species::ALL gains mannequin (player#main model, Humanoid anim), and
the player-only render paths — skin texture, sleep pose,
AvatarRenderer's 0.9375 scale — take an avatar() check instead of the
id string.
- SetEntityData's apply stores DATA_PROFILE on the mob, and
update_remote_skins resolves its textures property through
PlayerSkinRenderCache's rules (unsigned allowed, SkinManager's
insecure path) the way a tab player's skin resolves.
- Mannequins get the shared-flags crouch and sleeping-position handling
other players already had.
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Keep main's item_sync.rs and lan::guest::stack_to_network for stacks on
the wire (#1874) and move the firework component decoding there:
Component::Fireworks and Component::FireworkExplosion in
item_sync::components_tag, with fireworks_tag.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Nothing tracked `ItemCooldowns`: `draw_item_decorations` passed a cooldown of
0.0 with a note that there was none to shade, and ender pearls were held back
by an ad-hoc `pearl_cooldown` counter.
`horizon_client::cooldowns` keeps the `use_cooldown` component's entries by
cooldown group the way `Player.getCooldowns` does (start and end ticks, the
percent `getCooldownPercent` returns), `ItemInfo` now carries the item's
`use_cooldown`, and the game uses it where vanilla does:
`ServerPlayerGameMode.useItem`/useItemOn refuse a stack on cooldown,
`ItemStack.applyAfterUseComponentSideEffects` starts the entry when a use
finishes (a thrown pearl, a wind charge, an eaten chorus fruit),
`GuiGraphicsExtractor.itemCooldown` shades the slot, and a server's guests
keep their own and are sent `ClientboundCooldownPacket` (started, and 0 when
it ends), which a client applies.
Fixes#1763
Co-Authored-By: [CC] <deepseek-v4.1-flash> <noreply@deepseek.com>
Main gained PotionContents on the wire (#1796) in the same two functions
this branch hands to the shared codec. Resolved by keeping the codec and
teaching it the potion: components_tag writes it by its built-in registry
name, and the remote's id lookup asks the built-in registries first, as a
hub's SyncData::id already does. The round-trip test now covers a splash
potion.
Co-Authored-By: Hermes Agent (Nous Research) <noreply@nousresearch.com>
`Game::to_network` wrote damage alone into SetCreativeModeSlot, and
`item_from_network` read back damage and enchantments, so anything else a
creative stack carried was gone once the server echoed the inventory: a
named sword lost its name, trimmed armour its trim, an enchanted book its
stored enchantments and its repair cost.
`item_sync` is the pair for the components Horizon models: `encode` writes
them with the synced registries' ids (an unknown name is left out rather
than sent as id 0, which is a different entry) and `apply` reads them back
into the same `StackData` fields `item_data::decode_components` fills.
Potions are not carried yet: `minecraft:potion_contents` has no id or codec
in horizon-protocol, so a creative potion still arrives as an uncraftable
one (#1832). Nor are containers, charged projectiles, written books and a
map id, which the packet models but this patch does not write.
Refs #1832
Co-Authored-By: Hermes Agent (Nous Research) <noreply@nousresearch.com>
ClientboundCookieRequestPacket in play state decoded as Unknown and was
dropped; StoreCookie and Transfer likewise, and the configuration-phase
CookieRequest was always answered empty. Proxies (BungeeCord, Velocity)
store and read cookies around a server switch and send Transfer to move
a player to another address — both ended in a kick.
Play-state cookie_request/store_cookie/transfer now have packet types,
cookie requests answer from the stored cookies in both configuration and
play (ClientCommonPacketListenerImpl.handleRequestCookie), stored cookies
land in the session's map (handleStoreCookie), and a transfer reconnects
the game to the target address carrying them (handleTransfer →
ConnectScreen.startConnecting with TransferState).
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
A server's client (single player on the server thread, LAN guests, a
dedicated server's players) saw every potion as an "Uncraftable Potion"
in the water colour: the server never wrote `minecraft:potion_contents`
and the client read past it. A potion taken from the creative menu also
reached the server without its potion, so it splashed as water.
PotionContents now goes as vanilla's stream codec writes it: the potion
by its `minecraft:potion` id and the custom colour, with the custom
effects and name kept as their bytes. The server writes it for every
stack it sends, the client reads it into the item's components, and a
creative pick and the server's read of it keep the potion too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Decode source/event selection flags, including 26.3 UI source 10, and route them to Audio.stop_sounds. Compare all 36 native Java codec selections and round-trip their bytes; reject invalid sources and truncated packets.
Co-Authored-By: Codex (gpt-6.1-sol) <noreply@openai.com>
show_dialog and clear_dialog in configuration and play, and
custom_click_action, the answer, in both. Dialog.DIRECT_CODEC as
horizon_client::dialog: the five dialog types, their bodies, inputs and
actions with vanilla's defaults, the values inputs hold, and
Action.createAction (a command template's $(key) substitution with
escapeWithoutQuotes, a custom action's payload of every input's tag).
Nothing uses them yet: the screen comes next.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A loopback test listens on both [::1] and 127.0.0.1 and checks the
connection lands on IPv4 although IPv6 is listed first; it fails without
the sort. Another checks the fallback to IPv6 when IPv4 refuses. When every
address fails, the error is now the first address's (the one vanilla tries)
rather than an IPv6 fallback's, which is often only "no route to host".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The trim decode sat inside the read-past fallback and matched the kind
there; it now has a TRIM arm, and the fallback only reads past. Long
comment lines wrapped.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Main's SRV resolver (#1578) added Connection::connect_server; both connect paths now
go through connect_addresses, which puts IPv4 first, as ServerAddressResolver.SYSTEM's
InetAddress.getByName does for the original host and the SRV target alike.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Two halves of FireworkRocketEntity on a server were missing:
- SetCreativeModeSlot encoded only damage, so a creative-picked rocket
lost its minecraft:fireworks component (flight_duration and explosions
gone); the wire codec also had no decoder for it or firework_explosion,
so a stack carrying them failed to decode at all. The components now
follow Fireworks.STREAM_CODEC/FireworkExplosion.STREAM_CODEC both ways
(component ids 71/70).
- A remote rocket never read DATA_ATTACHED_TO_TARGET,
DATA_ID_FIREWORKS_ITEM or DATA_SHOT_AT_ANGLE, so the server-attached
boost rocket was an inert marker: attached_to now comes from
EntityDataSerializers.OPTIONAL_UNSIGNED_INT (the id plus one), the item
feeds the explosion list, and FireworkRocketEntity.tick's
isFallFlying boost applies to the local player.
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
boss_event_round_trips missed its #[test] attribute, failing clippy's
dead-code lint; unknown_packets_keep_their_payload decoded boss_event,
which this PR gives a real decoder, so the garbage payload no longer
comes back Unknown — it uses player_look_at, which still has none.
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Vanilla resolves a server with InetAddress.getByName, and Java's default
(java.net.preferIPv6Addresses=false) puts IPv4 addresses first. Horizon
tried addresses in the system resolver's order, which on macOS puts IPv6
first: a server whose IPv6 address resets the connection showed
"Connection reset by peer (os error 54)" where vanilla joined.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Decode the trim component as its material and pattern registry ids when both
are references (a direct holder stays read past) and name them through the
server's synced trim_material and trim_pattern registries, as enchantments
are, so worn and held trimmed armour draws its trim in multiplayer.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Vanilla's ServerRedirectHandler asks through JNDI's DNS client: the
system's name servers in turn, one 1 s UDP try each (retries = 1), TCP when
the reply is truncated, the first SRV record of the answer. resolver/dns.rs
does that by hand in ~200 lines with wire-format tests and a local stub
server, so the game, the dedicated server and the bots no longer pull in
tokio, hickory and ~25 other crates (none of them were in the game before).
Name servers come from resolv.conf on Unix, as JNDI reads them, and from
the network adapters on Windows (ipconfig, Windows only; the crate keeps
forbid(unsafe_code)). Each server waits 1 s, as vanilla does, instead of
the whole connect timeout.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The clientbound boss_event packet (add/update/remove/style/progress), a
BossOverlay that renders each bar, darkens the world under a wither, and
suppresses the create-the-world fog — vanilla's BossHealthOverlay and
GameRenderer behaviour on a real server.
Co-Authored-By: Devin <devin@cognition.ai>