🤖 Agent: Gate check by Ludovic's AI agent (Claude). Checked against 26.3 Soul Speed: the speed on soul blocks, no soul sand slowdown, the particles and sound, and boot wear on 4% of ticks, now also for each player on a server (`guest_server_tick`), with the particle sent to the players around. I pushed one fix: the random draws come in vanilla's order (the 0.04 wear roll first, then the particles, then the 0.35 sound roll and its pitch). It then conflicted with main (#2190 added the elytra glide wear in the same tick), so I merged main in: Soul Speed comes first, as `LivingEntity.baseTick` runs before `aiStep`, then the glide wear, so both draw from the level random in vanilla order. On that merge: clippy `-D warnings` clean, horizon-client tests pass, the fast build links, and `lan_e2e.py` elytra, damage and breath pass with the host in-frame and on its server thread. CI parity runs the harnesses this head reaches. Thanks @kacpherek!
🤖 Agent: #2318 has merged and this PR's base is main. The official Recipes mod (needs API level 6): pr_ready --quick green; the e2e passes 25/25 in-frame and on a dedicated server on this head (61 of 63 in five modes before; the 2 failures were #2501's bug, now merged). Queued at 92e3f496f.
🤖 Agent: #2306 has merged and this PR's base is main. recipes.place and recipes.give at API level 6, give cut to one stack. Sandbox (31), api_reference and clippy -D warnings pass; the actions e2e passed 58/58 in five modes. Queued at bd1436f16.
🤖 Agent: #2305 has merged and this PR's base is main. Recipe pages with station tabs and an ellipsis on a long title; clippy -D warnings with --features mods and --no-default-features is clean, and the browser e2e passed 98/98 in five modes. Queued at 0cba2edbd.
🤖 Agent: verify passed (incl. Linux clippy for the glibc malloc_trim path); chunk_generation, worldgen, worldgen_random, save_roundtrip and world_reopen pass on this head; pr_ready --quick green. Preload starts on world select per the orchestrator's decision. Perf fix, no label.
🤖 Agent: #2301 has merged and this PR's base is main. browser.set-recipes (with station tabs) at API level 6; sandbox (31), api_reference and clippy -D warnings pass. Queued at a373eea18.
A crate split reaches every harness (#2160: 352 harnesses, 71 min on a Mac
mini runner) and held the Mac mini's runners while other pull requests'
parity jobs queued for over an hour. With the full-parity label, the job
runs on the desktop's runner only (8 at a time) and waits for it while the
desktop is in use; its timeout is 300 min, since a job under way when
someone comes back goes on with 2 threads.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: #2300 has merged and this PR's base is main. Merged up the renumbered stack (browser at API level 6); clippy -D warnings with --features mods and --no-default-features is clean, and the browser e2e passed 98/98 in five modes before. Queued at a2cd70e8e.
🤖 Agent: #2284 has merged and this PR's base is main. browser is at API level 6 (recipes took 5). Sandbox (31, the UI guest at level 6), api_reference and clippy -D warnings pass. Queued at b836eea06.
🤖 Agent: #1895 has merged and this PR's base is main. Merged main: recipes moves to API level 5 (world took 4). Sandbox (30, including the level assertion), api_reference and clippy -D warnings on horizon-mods pass. Queued at b360e3c26.
🤖 Agent: LAN guests (host on its server thread) and a dedicated server's new players joined inside each other on the exact spawn, so they clicked the other player instead of a table or chest. New players now get PlayerSpawnFinder.findSpawn as in vanilla. lan_e2e spawned/crafting/chest pass in both host modes, a dedicated two-guest check passes, and the 32 affected harnesses pass.
glibc keeps the pages the preload thread freed in that thread's arena, and
macOS keeps them in its zones until memory runs short, so dropping the
preload didn't lower the footprint: horizon-threads now asks the allocator to
return free pages (malloc_trim, malloc_zone_pressure_relief).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The command ran on after remote.sh was stopped or its connection dropped,
holding the name's lock, and a game or daemon it left behind outlived the
run. It now runs in a scope named after the run (remote-<name>.scope),
stopped when the command ends, when the ssh session ends first, and before
the next run under that name.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The parity client's chat, villager and furnace steps went on 1.5 s after
sending, so a busy server's late reply landed in the next step (or a click
carried a stale state id). Each step now waits until the server answered:
its chat line, both slot packets of a shift-click, and the furnace totals
the click changes (up to the same 30 s patience). The respawn window keeps
only the player's own entity packets: a squid's swim event or a mob coming
into view falls into it as each server's ticks fall.
Off the known-flaky list: it passes under 2.5 s server stalls every 5 s,
where main's client fails.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: verify passed; chunk_generation, worldgen, worldgen_random, save_roundtrip and world_reopen pass locally on this head plus #2481; pr_ready --quick green. Perf fix, no label.
The worldgen bundle, the loot tables and the registries a server sends are
read on a helper thread from the moment the player opens Singleplayer or
Create World, so a world opening finds them ready. A world opened meanwhile
waits for the part being read. If the player goes back to the title or to
Multiplayer and opens no world for a minute, the preload is dropped: the three
caches can now let go of what they hold.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The host fetches the commit from GitHub (nothing else is sent), keeps one
checkout and target per worktree, runs the command niced in its CI slice on
CPUs 4-7 with scratch homes, prepares vanilla data for harness commands as
CI does, and streams the output back. One run per name at a time; no lock
leaks into daemons the run starts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The server put every player new to the world exactly on the world spawn.
On the server thread the host's own player joins that way too, so the host
and each new guest stood inside each other: the guest's client targets
the host's player (its eye inside that box, as vanilla's getEntityHitResult
does), so it could open no crafting table, chest or furnace, and Entity.push
never parts two players with the same centre. A dedicated server stacked
its new players the same way.
Vanilla places each new player with ServerPlayer.adjustSpawnLocation:
PlayerSpawnFinder.findSpawn, a free column at a random offset within the
respawn radius. Game::server_player_spawn does it for a server's new
players with the window's own finder, passing over the candidates whose
chunk isn't loaded yet (vanilla waits for them; a server tick can't).
lan_e2e's new check `spawned` asserts that the host and the guest join on
different spots, then has the guest open a crafting table, a chest, a
furnace and a stonecutter where it joined.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A title wider than the panel ended at its edge mid-word ("Recipes: Oak P");
it now ends with vanilla's "..." (CommonComponents.ELLIPSIS).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
pr_ready --quick passed #1895 while CI's clippy failed on an unused import in horizon-mods' tests: clippy on the
touched crates and their dependents gives them other features than CI's workspace-wide run (Cargo unifies
features over the packages it builds).
- clippy: CI's two commands as they are. The whole workspace runs with the game's code mods, and the server build
(horizon-client --no-default-features) runs when horizon-client is affected. Warm on the Mac mini, a change to
hud.rs took 14-16 s against 17 s before, and food.rs 22-32 s against 14 s. Each worktree's first run costs
about 3 min once.
- tests stay on the touched crates, with the features CI's workspace run gives them: horizon-client's code mods,
and horizon-mods' runtime, pack and index.
- The Python tests always run, as in CI: some read the crates, the roadmap or the workflows.
- The launcher steps run when CI's launcher job would (tools/ci/rust_changes.py's rule, which also covers the
root manifests and the check itself).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: merged main for #2316 (the campfire fix: this head had #2015 without it, so dispenser and block_removal would fail parity). Clippy -D warnings on horizon-mods (all targets, all features) and on horizon-client --features mods --bins is clean. Requeued at 8ec47d641.
🤖 Agent: Gate check by Ludovic's AI agent (Claude). Checked against 26.3: turtles lay (`TurtleLayEggGoal` with the digging effect), `TurtleEggBlock` hatches, and a growing baby drops a scute. I pushed one fix: the home is set at spawn and kept on load (`finalizeSpawn`, `home_pos`), and breeding drops one orb of 1 to 7, drawn from the mother's random and only with `mob_drops`. After #2137 the gift loot event names its mob. I merged main into your branch (6ef61ac13): main's `Reinforcements` stays beside your `BlockBreakEffect`, and the grown scute now goes through #2137's `drop_gift_loot` (the turtle's id added to both `GiftLoot` events) behind your mob_drops check; clippy clean, same code as the tested tree. Gate: this wave's 55 PRs merged on main in one tree: clippy `-D warnings` clean, 1186 horizon-client and 163 horizon-world tests pass, and of the 101 harnesses `parity.py --changed-files` picked, 100 pass (`block_removal` failed on one spawner case from #2199's orb merging, fixed there). `lan_e2e.py` passes 24 checks with the host on its server thread and in-frame (chest, xp, pvp, doors, villagers, beds, frames, projectiles, arrows, fishing, damage, mobs, blast, drops, elytra, breath, predict, cracks, barter, kills, hurt, contacts, underwater, plates), plus forced-levels on the server thread. Thanks @kacpherek!
🤖 Agent: Merged by Ludovic's AI agent (Claude). The loot tables now load before a world's first chunks are read, so a loot chest near spawn gets its roll. The new spawn_loot harness fails on main (0 stacks) and passes here with vanilla 26.3's exact roll, on the server thread and in-frame. I merged main after #2316; dispenser and block_removal pass again. Follows @FusselTV's #2073.
Jobs queued while the loading screen was up carried ring-order priorities,
far above the plain distances used once it closed, so they waited behind
every later job. The proto chunks they generated held back the decoration
of the chunks around the player: at render distance 32 in-frame, nothing
was decorated until every proto chunk had been generated. Queued jobs now
keep their column and bias and are ranked again from the current centre,
extra areas, band and loading screen when those change, as vanilla's
chunk task queue follows the current ticket levels.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Keeps main's Reinforcements beside BlockBreakEffect, and the turtle's grown
scute goes through #2137's drop_gift_loot (the mob's id added to its two
GiftLoot events) behind the mob_drops check.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Gate check by Ludovic's AI agent (Claude). Checked against 26.3: `BreakDoorGoal` (the difficulty gate, 240 ticks, the crack progress that your last commit now sends to nearby players) and the vindicators' door rules in raids. I pushed one fix: zombies the game spawns now roll `canBreakDoors` (`random < special * 0.1`), as `Zombie.finalizeSpawn` does. One difference is left for later: crack packets go to players within 64 blocks, and vanilla's `destroyBlockProgress` uses 32. I merged main into your branch (ac4309d3b): #2162 moved the zombie attributes into a free function, and its leader bonus now goes through `set_can_break_doors`; clippy clean, door tests pass, same code as the tested tree. Gate: this wave's 55 PRs merged on main in one tree: clippy `-D warnings` clean, 1186 horizon-client and 163 horizon-world tests pass, and of the 101 harnesses `parity.py --changed-files` picked, 100 pass (`block_removal` failed on one spawner case from #2199's orb merging, fixed there). `lan_e2e.py` passes 24 checks with the host on its server thread and in-frame (chest, xp, pvp, doors, villagers, beds, frames, projectiles, arrows, fishing, damage, mobs, blast, drops, elytra, breath, predict, cracks, barter, kills, hurt, contacts, underwater, plates), plus forced-levels on the server thread. Thanks @kacpherek!
tools/parity_hosts.json's new `known_flaky` lists the harnesses that still flake for a cause not fixed yet,
each with its reason: dedicated_server (its Java client's 1.5 s step timer), remote_fireball (wall-clock
sampling) and world_reopen. One of them that fails runs once more from scratch, and a pass then counts.
Every retry prints a RETRIED line at the end of the run and in the CI job's summary, so a harness that
passes only on its retry stays visible on PR parity and parity-main alike. The existing fresh rerun of a
failure that reused oracle results is reported the same way. A known flaky harness skips that rerun and keeps
reusing oracle results, because flaking doesn't mean its cached oracle runs were wrong. Repeated runs
(--repeat, --flake-check) are never retried.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Keeps #2184's glide wear with Unbreaking after Soul Speed's tick, the order
of LivingEntity.baseTick and aiStep (both draw from the level random), and
both the can_glide_through and soul_speed_blocks tags.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Gate check by Ludovic's AI agent (Claude). Checked against 26.3: `Shulker.hitByShulkerBullet` teleports an open shulker and may split it, less likely the more shulkers are near. I pushed one fix: the split roll uses the level's random, as vanilla's `level.getRandom()` does. Your update (d7ec7b607: the open state is read at hit time, the split chance from the level's random) checked too: clippy clean, and the 79 harnesses `parity.py --changed-files` picks pass apart from the campfire cases #2316 fixes. Gate: this wave's 55 PRs merged on main in one tree: clippy `-D warnings` clean, 1186 horizon-client and 163 horizon-world tests pass, and of the 101 harnesses `parity.py --changed-files` picked, 100 pass (`block_removal` failed on one spawner case from #2199's orb merging, fixed there). `lan_e2e.py` passes 24 checks with the host on its server thread and in-frame (chest, xp, pvp, doors, villagers, beds, frames, projectiles, arrows, fishing, damage, mobs, blast, drops, elytra, breath, predict, cracks, barter, kills, hurt, contacts, underwater, plates), plus forced-levels on the server thread. Thanks @kacpherek!
🤖 Agent: Gate check by Ludovic's AI agent (Claude). Checked against 26.3: `PiglinAi.mobInteract` takes only gold ingots from a hand, and other loved items go to its inventory. One difference is left for later: the comparison of two items it likes equally isn't ported yet. Your update (e60cbe8d4: whole nugget stacks stay in the inventory, picked-up enchantments are kept) checked too: clippy clean, the 79 harnesses `parity.py --changed-files` picks pass apart from the campfire cases #2316 fixes, and lan `barter` passes in both modes. One gap left for later: vanilla can put a nugget stack in an empty main hand, and Horizon's hand slot holds one item. Gate: this wave's 55 PRs merged on main in one tree: clippy `-D warnings` clean, 1186 horizon-client and 163 horizon-world tests pass, and of the 101 harnesses `parity.py --changed-files` picked, 100 pass (`block_removal` failed on one spawner case from #2199's orb merging, fixed there). `lan_e2e.py` passes 24 checks with the host on its server thread and in-frame (chest, xp, pvp, doors, villagers, beds, frames, projectiles, arrows, fishing, damage, mobs, blast, drops, elytra, breath, predict, cracks, barter, kills, hurt, contacts, underwater, plates), plus forced-levels on the server thread. Thanks @kacpherek!
🤖 Agent: Gate check by Ludovic's AI agent (Claude). Checked against 26.3: `LivingEntity.handleFallFlyingCollisions` hurts on a horizontal hit with `fly_into_wall`. Gliding stops on climbable blocks and with Levitation, and the elytra wears through `hurtAndBreak`, so Unbreaking applies. Your update (a4407a4bc: Unbreaking for guests, wall hits kept on the server copy) checked too: clippy clean, the 79 harnesses `parity.py --changed-files` picks pass apart from the campfire cases in `dispenser` and `block_removal` that #2316 fixes, and lan `elytra` passes in both modes. Gate: this wave's 55 PRs merged on main in one tree: clippy `-D warnings` clean, 1186 horizon-client and 163 horizon-world tests pass, and of the 101 harnesses `parity.py --changed-files` picked, 100 pass (`block_removal` failed on one spawner case from #2199's orb merging, fixed there). `lan_e2e.py` passes 24 checks with the host on its server thread and in-frame (chest, xp, pvp, doors, villagers, beds, frames, projectiles, arrows, fishing, damage, mobs, blast, drops, elytra, breath, predict, cracks, barter, kills, hurt, contacts, underwater, plates), plus forced-levels on the server thread. Thanks @kacpherek!
A head tested alone keeps failing harnesses that main already fixed. Since 10-25 evening, 9 of 21 failed PR
parity jobs failed only on such fixes: server_knockback (heads without #2114, 7 times), mob_ai's oracle heap
(without #2270), save_roundtrip (without #2212). #2160 needed 4 parity runs this way.
The job now merges main's current tip into the head and selects harnesses against that main commit, so only
the pull request's own changes count. A branch that doesn't change the runner scripts or the selection gets
main's, which replaces the two steps that copied main's vanilla-data.sh and parity.py in by hand. A head that
conflicts with main is tested alone, with a warning. The job summary names the main commit, and the check is
still `parity / parity` on the head.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🤖 Agent: Gate check by Ludovic's AI agent (Claude). Checked against 26.3: `Witch.aiStep` stands still while drinking (the -0.25 speed modifier) and only drinks an effect it doesn't already have, and `performRangedAttack` skips a potion the target already has. It merges after #2138, since both add fields to `Seen`. I merged main into your branch (f8415efe2): `Seen` keeps #2138's sitting/ridden fields beside your effects; clippy clean. Gate: this wave's 55 PRs merged on main in one tree: clippy `-D warnings` clean, 1186 horizon-client and 163 horizon-world tests pass, and of the 101 harnesses `parity.py --changed-files` picked, 100 pass (`block_removal` failed on one spawner case from #2199's orb merging, fixed there). `lan_e2e.py` passes 24 checks with the host on its server thread and in-frame (chest, xp, pvp, doors, villagers, beds, frames, projectiles, arrows, fishing, damage, mobs, blast, drops, elytra, breath, predict, cracks, barter, kills, hurt, contacts, underwater, plates), plus forced-levels on the server thread. Thanks @kacpherek!
🤖 Agent: Gate check by Ludovic's AI agent (Claude). Checked against 26.3: damage that doesn't bypass armor (cactus, fire, lava, campfire, berry bushes) wears the armor through `hurtArmor`. Here the wear lands when the damage events drain at the next player tick; vanilla applies it inside `actuallyHurt`. That is close enough for now. #2144 merged first and conflicted here, so I merged main into the branch: the hurt_armor call this PR takes out goes, and main's Thorns code stays. Gate: this wave's 55 PRs merged on main in one tree: clippy `-D warnings` clean, 1186 horizon-client and 163 horizon-world tests pass, and of the 101 harnesses `parity.py --changed-files` picked, 100 pass (`block_removal` failed on one spawner case from #2199's orb merging, fixed there). `lan_e2e.py` passes 24 checks with the host on its server thread and in-frame (chest, xp, pvp, doors, villagers, beds, frames, projectiles, arrows, fishing, damage, mobs, blast, drops, elytra, breath, predict, cracks, barter, kills, hurt, contacts, underwater, plates), plus forced-levels on the server thread. Thanks @kacpherek!
🤖 Agent: Gate check by Ludovic's AI agent (Claude). Checked against 26.3: the `ExperienceOrb` pickup box, merging by count, and the lava bounce. Your last commit's fix for a merged orb picked up by a guest or a server's client (one count at a time, the client keeping the orb) matches what I found. I pushed one fix to the harness: whether orbs merge depends on entity ids, which Horizon doesn't share with vanilla, so a spawner case in `block_removal` merged here and not in vanilla. The oracle and the fixture now print a merged orb once per count, and that case passes. (`block_removal`'s campfire cases fail on main until #2316 lands.) Gate: this wave's 55 PRs merged on main in one tree: clippy `-D warnings` clean, 1186 horizon-client and 163 horizon-world tests pass, and of the 101 harnesses `parity.py --changed-files` picked, 100 pass (`block_removal` failed on one spawner case from #2199's orb merging, fixed there). `lan_e2e.py` passes 24 checks with the host on its server thread and in-frame (chest, xp, pvp, doors, villagers, beds, frames, projectiles, arrows, fishing, damage, mobs, blast, drops, elytra, breath, predict, cracks, barter, kills, hurt, contacts, underwater, plates), plus forced-levels on the server thread. Thanks @kacpherek!