Open-source platform for 3D AI agents. Turn text or a photo into a rigged, animated GLB avatar, give it an LLM brain, memory and a wallet, and embed it anywhere with one web component. Remote MCP server, x402 payments, Three.js, browser-native.
JavaScript 3d-generationai-agentsanimationavatardigital-humangenerative-aiglbgltfimage-to-3dllmmcpmcp-servermodel-context-protocolsolanatext-to-3dthreejsweb-componentswebglwebxrx402
test(x402-discovery): budget the catalog render for the cold import and live upstreams it does
nirholasc0012af3d8f2
diff --git a/tests/api/x402-discovery-green.test.js b/tests/api/x402-discovery-green.test.jsindex 7ae9e96c8..4191271e4 100644--- a/tests/api/x402-discovery-green.test.js+++ b/tests/api/x402-discovery-green.test.js@@ -99,11 +99,10 @@ function checkResource(r) {return { label: label(r), errors, warnings };}-// Rendered in a CHILD node process, loading api/wk.js natively — exactly how+// Rendered in a CHILD node process, loading api/wk.js natively, exactly how// production (server/index.mjs) loads it. Importing it in-process here would-// pull its multi-thousand-module graph through vitest's transform pipeline:-// ~55s in isolation and past the 120s hook budget under full-suite contention,-// so the file flaked on load, never on content. The native import is ~6s.+// pull its multi-thousand-module graph through vitest's transform pipeline and+// the file would fail on load rather than on content.const RENDER_ENV = {APP_ORIGIN: 'https://three.ws',X402_PAY_TO_BASE: '0x0000000000000000000000000000000000000001',@@ -151,10 +150,20 @@ function renderDiscovery() {describe('x402 discovery catalog is fully green', () => {let report;+ // The hook needs its own budget because the global 120s ceiling does not cover+ // what this render actually costs, and the file failed on that ceiling every+ // run, loaded or idle. Measured on this box with nothing else running: ~54s for+ // the child's cold native import of api/wk.js, then 18-26s to render. The+ // render walks live upstreams (CoinGecko, DeFiLlama) to enumerate the datapoint+ // families, and each of those fetches is bounded at 15s+ // (api/_lib/market-data/datapoints.js), so six families put the render's worst+ // case near 90s. Budget = that worst case plus the cold import plus room for+ // five sibling forks competing for the box. None of the four assertions below+ // says anything about speed.beforeAll(async () => {const doc = await renderDiscovery();report = (doc.resources || []).map(checkResource);- });+ }, 420_000);it('catalogs a non-trivial number of resources', () => {expect(report.length).toBeGreaterThan(20);