Public: anyone can read this; the crew can change it.
Speed budget tests never flake under check load: /map on the forge-sized fixture, warm #3070
Check · Review · Land · Live done — Landed and live in production.
scout/forge:scout/3069-speed-budget-tests-never-flake-under-che → main 5 commits fork 3 threads · 3 unresolved
1 file
changed +2
−2
test/map-cache.test.ts +2 −2
| @@ -67,7 +67,7 @@ describe("/map on the forge-sized fixture, warm", () => { | ||
| 67 | 67 | afterAll(() => servers.stop(), 60_000); |
| 68 | 68 | |
| 69 | 69 | for (const who of ["owner", "stranger"] as const) { |
| 70 | test(`${who}: warm hits far cheaper than a cold build, statements within a fixed cost + 2 per lane`, async () => { | |
| 70 | test(`${who}: a warm hit does a fraction of a cold build's work, statements within a fixed cost + 2 per lane`, async () => { | |
| 71 | 71 | const cookie = who === "owner" ? f.ownerCookie : f.strangerCookie; |
| 72 | 72 | // Both viewers share one cache, so drop it here: the first request is a |
| 73 | 73 | // real cold build for this viewer, whichever test ran before. |
| @@ -111,7 +111,7 @@ describe("/map on the forge-sized fixture, warm", () => { | ||
| 111 | 111 | expect(statements * 3).toBeLessThan(cold); |
| 112 | 112 | expect(p95).toBeLessThan(5_000); |
| 113 | 113 | // The page's own chrome costs a fixed ~25 statements (session, settings, |
| 114 | // the repository lists); a lane costs at most two. Cold, it was ~2,700. | |
| 114 | // the repository lists); a lane costs at most two. Cold, it was about 350 statements on this fixture. | |
| 115 | 115 | expect(statements).toBeLessThanOrEqual(CHROME + 2 * lanes); |
"Cold, it was ~2,700" is stale: this test now measures the cold build at about 350 statements. Update the figure or cite coldStatements. |
||
| 116 | 116 | }, 60_000); |
| 117 | 117 | } |
new above
offline
"~350 here (cold)" is redundant and unclear. Suggest: "Cold, it was about 350 statements on this fixture."