4fe648b674
playhorizon.dev now deploys from the private ludovic111/horizon-site, which holds website/'s history. The files its build still reads from here move out of website/: - the design tokens to crates/horizon-theme/tokens.css (xtask tokens writes them, tokens_css_is_up_to_date checks them); - the roadmap's JSON to roadmap/roadmap.json (tools/roadmap.py build); - the benchmark fixtures to crates/horizon-telemetry-server/tests/fixtures/benchmarks/ (the site keeps a copy); - the web icons to docs/media/brand/web/ (xtask icons). pr_ready loses its website area and the tokens special case; rust_changes, the roadmap's links and every doc that pointed into website/ follow. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Benchmark summary fixtures
Used by the telemetry server's GET /v1/benchmarks/summary test (tests/summary.rs) and, as a copy, by the
website's aggregator test (scripts/benchmarks.test.mjs in ludovic111/horizon-site, which checks its copy against
this one on main): the same benchmark items in, the same groups out. Change both together.
items.jsonl: one stored item per line (install_id,day,version,data: a benchmark report). Group RTX 4080 + i7-13700KF at render distance 12 has three installs (one with two runs); the same install at 32, and the RTX 3060 group, have fewer than three and are left out. The other runs are rejected: VSync on, offscreen, a frame cap, a missing scene, a one-scene scenario, a report of another version, no install ID.- Breakdowns: the RTX 4080 group's three installs share a VRAM class (16 GB), the OS and a discrete GPU, so those cells are published; a3 has 16 GB of RAM and an SSD where a1 and a2 have 32 GB and NVMe, so neither the RAM nor the drive cells reach three installs and those dimensions are left out.
expected.json:min_installsandgroupsof the summary (generatedvaries).