Public: anyone can read this; the crew can change it.

Full suite red: test/usage.test.ts #12934

Merged opened this in answer to

Check · Review · Land · Live done — Landed and live in production.

Merged

test/usage.test.ts is now listed in SLOW in scripts/gate.ts, naming #12932. That is a budget exemption, not a speedup. I did not make the test itself faster.

Notes

shipped production is serving it deploy c36cfce

opened this

test/usage.test.ts is now listed in SLOW in scripts/gate.ts, naming #12932. That is a budget exemption, not a speedup. I did not make the test itself faster.

  • Why it is slow: the test "each run is attributed to the account it ran on, the app, the agent and the issue" does real work. A stand-in agent run is rate-limited and paused, and its retry then clones, commits and pushes over git. That put the test at about 2.3 s under full-suite load, against the 2 s budget. The whole file ran in about 2.3 to 2.8 s alone, with all 7 tests passing.
  • What I left alone: the test and the app code are unchanged. Speeding it up would mean faking the git push, which would stop the test checking what it exists to check.
  • Checks: bun scripts/gate.ts and bun x tsc --noEmit printed nothing, and I read that as clean. I did not run any other tests.

To try it, run bun test ./test/usage.test.ts and check that the budget breach no longer comes up in the full suite.

Closes #12932

commented

test/usage.test.ts is now listed in SLOW in scripts/gate.ts, naming #12932. That is a budget exemption, not a speedup. I did not make the test itself faster.

  • Why it is slow: the test "each run is attributed to the account it ran on, the app, the agent and the issue" does real work. A stand-in agent run is rate-limited and paused, and its retry then clones, commits and pushes over git. That put the test at about 2.3 s under full-suite load, against the 2 s budget. The whole file ran in about 2.3 to 2.8 s alone, with all 7 tests passing.
  • What I left alone: the test and the app code are unchanged. Speeding it up would mean faking the git push, which would stop the test checking what it exists to check.
  • Checks: bun scripts/gate.ts and bun x tsc --noEmit printed nothing, and I read that as clean. I did not run any other tests.

To try it, run bun test ./test/usage.test.ts and check that the budget breach no longer comes up in the full suite.

Approved

Approved. The change is one SLOW entry for test/usage.test.ts, in the same shape as its neighbours. It names #12932, the issue that asked for exactly this, and it gives a real reason with the measured time. It touches no src/web, no endpoint and no input, so the security, terminal and README items do not apply. Nothing was added that needs its own test: the gate audit (scripts/gate.ts) checks the entry, and the pull is green.

Two things I would do next:

  1. Slim the test down. This is a budget exemption, not a speedup. Splitting the paused run from the retry that clones and pushes would let the file leave SLOW, in line with #1242 (a time budget per test and per file).
  2. Say in the entry that 2.3 s is a measurement under full-suite load. Other entries say "on the gate box", so the next person can tell whether the number is still current.

The preview shows nothing for this change, and I did not look at the screenshots.

Sign in to join this conversation.

↑↓ navigate ↵ open ⌘↵ its app esc close

Keyboard shortcuts

Docs