At render distance 32 the single section arena grew past the 4 GiB buffer
limit and the renderer panicked. Sections now live in pages of at most
256 MiB (or the device's buffer limit); draws keep their order and bind each
page once per run of consecutive draws in it, so translucent sections stay
back to front. docs/graphics.md describes the settings, the benchmark and
what changed; graphics_bench.py --compare prints a before/after table and
keeps each game's log.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Unloaded chunks and entity chunks are encoded and written by a background
region writer, in order; loads see queued records first, and synchronous
saves wait for the queue. Deleting entity chunks a region doesn't hold no
longer rewrites the region file.
- Mobs render within Entity.shouldRenderAtSqrDistance (64 x bounding-box size
x entity distance) and inside the view frustum (EntityRenderer.shouldRender);
block entities outside the frustum are skipped, beams and structure blocks
excepted.
- Minecarts and dropped items tick only in entity-ticking chunks, like mobs.
- Villager POI searches skip sections whose palette has no POI block.
- Streaming bookkeeping uses the Fx hash and checks a column's light
readiness once per frame instead of once per dirty section.
- HORIZON_SPIKES=MS prints the timed spans of slow frames; frame summaries
give the phase split of the slowest 1%.
RD 32 village flight on the 4080: 81 -> 230 fps; 1% low 5.8 -> 18.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Terrain sections live in one vertex arena (first-fit free list, grown on the
GPU) and draw with one multi_draw_indexed_indirect per pass instead of a
buffer bind and draw per section; per-frame scratch vectors are reused.
- The SectionOcclusionGraph search runs over dense arrays around the camera
instead of hash sets, with per-column visibility.
- Mobs outside the entity-ticking chunks (Chebyshev distance of the
simulation distance, DistanceManager.inEntityTickingRange) only check
despawning, as in ServerLevel.tick; before, every loaded mob ticked.
- World chunks use the Fx hash; sky darkening is sampled once per time and
weather instead of once per mob; creaking-heart scans compare state ids.
- Horizon-only graphics settings (horizon-graphics.txt): frame pacing without
spinning, present mode, animated textures, worker threads.
RD 32 village flight on the 4080: 3.1 -> 81 fps.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
horizon-ctl gains 'fly' (a constant-speed camera path that replies with the
frame summary); frame summaries add CPU time per frame (waits for the GPU,
display and limiter excluded), per-phase averages and GPU time from timestamp
queries. --offscreen renders into a texture on the discrete GPU so the 4080
can be measured under the headless compositor. Baselines: lavapipe on vscreen
and the RTX 4080 offscreen, seed 1234.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Removed (each unreferenced by code, workflows, tests or living docs; no harness
dependency record names any of them, and parity.py --list still finds all 281):
- 22 one-off write-ups of the Java-pack experiments in docs/: adaptive-terrain,
client-graphics-smoke, compact-headers, core-performance, culling-empty,
culling-rays, debounce-queue, dispatch-retry, foreground-first,
graphics-api-comparison, heap-budget, mac-first-test, optimization-plan,
playable-validation, quad-reuse, server-graphics-context, shader-world-smoke,
terrain-measurements, terrain-position, terrain-read-wait, tick-membership,
worker-pacing. Their findings are folded into one paragraph of java-packs.md.
- 24 benchmarks/ observation dumps only those write-ups linked (client graphics,
culling, queues, heap, Mac Vulkan, shader worlds, software identity, and the
unlinked rust-core, rust-worldgen-2026-10-05, rust-native-client and
rust-cuboid-model-decoding runs), plus 12 unlinked files inside kept
report folders. The scoped-parity reports the living docs cite stay.
- debug/: one server log from a chunk investigation.
- The manual macos-shader-research and macos-vulkan-research workflows.
- tools/logs/latest.log, an empty log committed by accident (now ignored).
Kept: research-lock.json (parity.py, packaging, the launcher and datafix read
it), the Java-pack how-tos (vulkan, collectors, profiling, process-memory,
tick-state) and tools/test_client_graphics.py (ci.yml runs it).
README: leads with what Horizon is and "Download the launcher" (the
launcher-v* releases; the launcher updates itself), then build from source
for developers. Direct game downloads and PLAY.txt steps for players are gone;
the launcher has its own section with two screenshots. The roadmap merges the
known differences, the final checks and the plans after parity, under the
"Roadmap" heading the launcher's What's next page reads. The parity table and
its 100%-of-what-the-harnesses-compare note are unchanged.
docs/README.md indexes what remains: build and run, parity (status and
workflow), world generation, world data and simulation, the old Java track.
Links to the removed pages and README anchors are fixed.
packaging/PLAY.txt now speaks to players who came through the launcher and
points anyone with a bare archive to it, keeping the manual steps short.
The launcher's README link points at main instead of the launcher branch.
.gitignore: tools/logs/ and .DS_Store. No code behaviour changes; the compiler
and clippy flag no unused code and the tree has no stale TODOs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
horizon_client::chat::ChatScreen is now a port of vanilla's ChatScreen,
EditBox and CommandSuggestions: the input's scrolling, cursor, selection,
word moves and select all, history up and down, argument colouring, usage
lines and parse errors, the suggestion list (sorting, hover, Tab and
Shift+Tab cycling, clicks, wheel) and its tooltips, drawn as GUI calls the
game paints. tools/test_chat_screen_parity.py drives a real ChatScreen on
vanilla's client command tree (custom suggestions answered by the server
dispatcher): 3000/3000 cases match.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
EmptyItemInHotbarFix was missing from the data fixers. horizon_client::hotbar loads the nine rows of nine stacks of an older file as vanilla does: lenient counts, unregistered items empty, stray entries dropped from their row, a row that is no list ending the reading, short rows lost and long ones cut.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
horizon_client::pack_screen ports PackSelectionScreen, its two
TransferableSelectionLists and the ConfirmScreen of an incompatible pack as
vanilla's state machine (clicks, wheel, search typing, focus, selection,
scroll) and gives its GUI calls; screen_draw holds the widgets (StringWidget,
MultiLineTextWidget, Button, a font-aware EditBox port) and StringSplitter's
line splitting with legacy codes. Pack components keep their legacy codes and
style fields. tools/test_pack_screen_parity.py runs a real
PackSelectionScreen through the same actions: 1500/1500 cases match.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- ClientboundUpdateRecipesPacket at join: the recipe property sets and the
stonecutter's recipes, from the data pack (lan/recipes.rs).
- On a player's first tick the minecraft:tick criteria are met: the recipes
their advancements grant go out in ClientboundRecipeBookAddPacket, then
ClientboundUpdateAdvancementsPacket with what the player may see
(lan/advancements.rs).
- Command replies reach remote players as the components vanilla sends:
Text keeps the Component it was built from (literal, translatable,
siblings, style), sendFailure wraps the message as vanilla does, and the
NBT wraps the elements of mixed lists.
The dedicated-server harness drops its known gaps: it compares these
packets (sorted where vanilla's hash maps order them per run) and the
replies to three failing commands, keeps only the first set_time among the
joining packets and turns random ticks off in the generated world.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Book::ghost is the one place that places a recipe's ingredients into the grid or furnace slots; the harness runs fillGhostRecipe against the real menus. Fixes found: the result and the dyed, potion and map demo ingredients cycle through every stack (not the first), a furnace's fuels follow the registry's component lookup, and items keep their registry order.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The recipe book view harness now searches: names, parts of names, ids with a colon, odd texts.
Horizon's recipe tooltips lacked attribute modifiers, smithing template, potion, dyed, pot, rocket and disc lines, and the results a display cycles through (map copies, tipped arrows, dyed armour).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The LAN and dedicated servers built the tree from the data generator's
commands report, which sorts each node's children. lan/command_order.rs,
generated from vanilla's Commands(CommandSelection.ALL) by
tools/generate_lan_command_order.py, gives the registration order, and the
dedicated-server harness now compares the tree in order (the game rules,
registered in an identity-hashed order, aside).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
horizon_client::packs::SelectionModel ports PackRepository's selection and
PackSelectionModel (default positions, fixed and required packs, moving up
and down, commit), and the pack screen uses it. The new pack_selection
harness drives vanilla's model and Horizon's through 4,000 random action
sequences; every state matches.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The HUD harness's chat cases now also draw the chat as the open chat
screen does: foreground, at the focused height, scrolled, with the scroll
bar.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The world was saved 20 wall-clock seconds after force-loading the area, so
on a busy machine chunks were still generating or their fluids flowing,
and random ticks changed blocks while the client was being sent its
chunks: which chunks differed varied from run to run. Wait until every
force-loaded chunk is loaded and 600 game ticks have passed, with
random_tick_speed 0.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The golem's goals are replaced by vanilla's brain (core and idle activities,
the transport cooldown, look-at-player-sometimes, stroll or wait), and
TransportItemsBetweenContainers moves into the library: nearest wanted
chest in chunk and block entity map order, queueing behind golems with a
chest open, the 60-tick interaction with its states, sounds and opener
counts, picking up and putting down (with vanilla's overcounting), visited
and unreachable positions. Weathering draws from the level random after the
living tick; the spin sound is client-local. mob_ai gains seven copper golem
scenarios that print each golem's state, hand and opened chest and every
chest's contents and openers; all match vanilla.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Profiles' textures properties are unpacked as authlib does and looked up
as SkinManager does (signatures, the local player, allowed texture URLs,
downloads), skins are fetched and processed as SkinTextureDownloader does
and shown from atlas slots. The tab_list harness now runs a real
SkinManager and compares processLegacySkin over 400 PNGs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Beardifier, density pow/log, Marsaglia's Gaussian, dripstone, ores, fancy
trunks, icebergs, end spikes, strongholds, note pitches, sculk charge
particles, the warden's and brains' weighted shuffles.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>