🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
GitHub started both 10-26 schedules over an hour late, so the hour check skipped both and no audit ran.
lan_integration takes ~2,400 s under load, past the 1,800 s per-harness limit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: gate verified on e3e08911b (main merged). fmt, clippy -D warnings and horizon-client tests pass. lan_e2e with the server-thread host: 131/133, and cursor/falling pass on rerun in both modes. 14 in-frame checks pass. agent_control_e2e: only manage fails, and it fails the same way on main. 82 of the 83 affected harnesses pass. lan_integration flakes under load and is now named-only (nightly). Queued.
Main gained players.online, tell and broadcast (#1833) after 0.2.69, so
they take level 1 with render.set-minimap. horizon-mod-registry's own
Cargo.lock picks up wasmparser.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
TextDisplayRenderer.submitInner calls submitTextBackground(-1, -1, width,
height), and Font.prepareBackground's effect glyph spans x0..x1, y0..y1, so the
fill is width + 1 by height + 1 from (-1, -1), as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
sounds.rs: main's Assets::load reads the launcher's files through game_files.
The rewritten-manifest test takes main's message; main's new test passes
GameFiles. ATLauncher is also looked for where the Horizon launcher looks
(~/Library/Application Support/ATLauncher, the Flatpak's data/atlauncher).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
Level.canHaveWeather is false in the Nether and the End, so their rain and
thunder levels stay at 0. Set that once in the mob world instead of checking
the End at two of the rain sites: bees, dolphins, wolves and happy ghasts read
the same flag.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
A pack's `fn material(input: MaterialInput) -> Surface` now gives the lit
colour of every terrain surface: sections and distant tiles, solid,
cutout and translucent, in both the instanced and the flat draws. Terrain
keeps Horizon's meshes, culling and draw lists; only its pipelines change.
Their shader is the pack's stage, a prefixed copy of terrain.wgsl's
terrain section, and entries that hand the pack the texture, the vertex
tint, the light levels and the lightmap's colour apart, then apply the
alpha test and fog as `fs_solid` and `fs_translucent` do.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
The game program carries its mod API level in a mark
(horizon-mod-api-level:NNNN;), which the launcher reads once per
installed version; a program from before the mark (0.2.68 and older)
serves level 0. The Horizon browser marks versions above the level of the
game the player runs (the selected version, else the newest installed)
"needs a newer Horizon than the game you play", and installs the newest
version that game can run, or says to update it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
🤖 Agent: Merged after the contrib-gate-16 review: checked against 26.3, the in-game checks and lan_e2e in both modes, and the 79 affected harnesses on the integration tree (see the gate comment). Thanks!
Review caught three regressions against the pinned client jar: submitInner
lines stay lineHeight+1 (10 px), submitNameTag fades discrete and
see-through names with the same 0x7F, and lightCoordsWithEmission floors
both light channels at 2.
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
A teleport far away (or a storage that moves by more than its size) drops
every column in one frame: 3,725 at RD 32. That frame freed every column's
sections, then removed each section from the renderer one at a time, with
a few hash lookups per section and a free-list tree update per mesh range.
- The dropped columns, and a left dimension's whole chunk map, are freed on
a short-lived thread of their own ("Chunk freeing").
- Renderer::remove_columns removes whole columns: one lookup per column for
the tables kept by column, and the meshes' ranges go back to the arena in
one sorted merge per page (FreeList::give_all) instead of a tree update
each. Per-section Renderer::remove and Occlusion::remove had no other
callers and are gone.
- The renderer's block-entity and detail-column tables use FxHashMap.
Mac mini (M4 Pro), RD 32, explored world, server thread, 3 alternated runs,
two long teleports each: the drop took 26.1-39.7 ms, now 6.9-8.5 ms.
Entering the Nether: 9.7-13.4 ms, now 4.0 ms or under 6 ms.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: merged main (12 files conflicted with ui, render and input). The runtime now keeps the server/client Hooks split beside the ui-events export, and the Snapshot carries both the GUI fields and the server's players. Locally: the horizon-mods sandbox and unit tests, cargo-horizon-mod tests, and clippy on the client with mods all pass. CI verify and parity pass at b90ab85f4. Bottom of the server-mods stack (#1837, #1894, #1895, #1901 follow).
server_knockback, locator_bar, server_dimension_players and their one-player users now end a step by the server's
game clock (a sync round trip, N ticks, a last sync) the way server_player_targets does, so a server that stalls
under load still gets all its ticks inside the step. server_knockback's frozen section waits for the stepped tick,
the stopped clock and the landed hit; dedicated_server waits for the player's leaving before /list.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
GeneratorData::load read everything on the server thread while the chunk
workers waited: the block tables, the structure templates and the block
behaviours now load on threads of their own, and BlockBehavior::load reads
its cached-support table beside the other two. The behaviour exports
(gameplay-facts.json is 55 MB) parse from bytes in memory instead of
through a reader.
ThinkPad (2 cores), 3 alternated runs: 0.63 -> 0.45 s with the behaviour
snapshot cached, 2.40 -> 2.07 s on a first start.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The predicted break played the block's break sound (wool for fire) and its destroy particles. BaseFireBlock skips both (spawnDestroyByEntityParticles does nothing); players only hear the server's extinguish sound.
🤖 Agent: I looked at the launcher's Shader packs tab in dark and light (Mac mini, mktemp homes, a local registry with the sample pack): Soft Glow is listed from Horizon with Works, Page and Install, and the look matches the other tabs. pr_ready --quick is green, with the launcher's 130 tests.
The sky stage runs a copy of terrain.wgsl's sky beside the pack's code
with every name prefixed `horizon_sky_`. The prefixer moves to
`gfx/wgsl_prefix.rs` and now also renames module variables (bindings)
and overrides, so the material stage can copy the terrain section the
same way. The sky's shader is unchanged apart from where `globals` is
declared.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The harness runs the whole lan_e2e suite: ~40 minutes on a loaded runner,
and single checks flake under load (barter, riding, shulker_bullets in one
run; cursor, falling in another, all passing on rerun). Selected by
--changed it would run on most gameplay pull requests and on parity-main.
tools/parity_hosts.json gains a named_only table that every automatic
selection (--changed, --failed, the default, --flake-check) leaves out;
naming the harness or --all (the nightly audit) still runs it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every function the game offers mods has the API level it arrived at
(crates/horizon-mods/src/api_level.rs): level 0 is the contract of
Horizon 0.2.67 and 0.2.68, level 1 adds render.set-minimap. A mod needs
the level of the newest function it imports, read from its component.
- The game refuses a mod that calls a function it doesn't have with
"made for a newer Horizon (it uses ...): update Horizon to run it"
instead of wasmtime's linking error.
- The registry index lists each mod version's level (api_level, read
from the file by CI); check_file compares it when the entry has one,
and the moved extras' installer takes the newest version this game
can run.
- cargo horizon-mod check prints the level.
- The api_reference test fails until a new WIT function has a level.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: verified: pr_ready --quick green (launcher and horizon-mods tests), CI verify and check green, merges cleanly with main. The registry takes Horizon shader packs, checked by the game's own shader check.
🤖 Agent: verified: pr_ready --quick green, 80/80 affected harnesses pass locally, CI verify green, A/B shows no frame cost without a pack. A Horizon shader pack's sky stage now draws in the game.
🤖 Agent: verify passed. Checked against the real registry now that Minimap 1.0.0 is published: the migration e2e passes 34/34 on the server thread and 34/34 in-frame. The minimap case starts at bottom left, 128, round, 0.5 blocks per pixel, turning: the mod is installed and on, its settings carried, the map shows, and the Extras page has no minimap rows. pr_ready --quick passes (clippy, 1295 horizon-client tests).
🤖 Agent: verified: pr_ready --quick green, 38/38 affected harnesses pass locally, CI verify green, A/B shows no frame cost. Internal: a Horizon shader pack's frame uniform is shared by its stages (no visible change).
The official Minimap mod (mods/official/minimap) shows the extra's map
through render.set-minimap. The extra's switch and its five Extras rows
go; the game still draws a mod's map from the same tiles. A player who
had the minimap on gets the mod installed and turned on at the first
start, with the corner, size, shape, zoom and turning carried into its
storage, and one chat line. horizon-extras.txt keeps the old keys.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: The official Minimap mod (mods/official/minimap): the minimap extra as a mod with its five options. Pixel-identical to the extra in the map region (0.000 to 0.002) with default and other settings, at the same frame cost. Verified: pr_ready (2179 workspace tests, clippy), the settings page in game, and the published mods loading beside it in both modes.
🤖 Agent: Mod API render.set-minimap: a mod picks the corner, size, scale, shape and turning, and the game draws the map from the minimap extra's tiles. It is a new function, so published mods keep loading. Verified: pr_ready (1353 tests, clippy, launcher), the sandbox test, and in game beside the published mods in both modes.
Ludovic Marie
socksprox
Kacpherek