Bug: Group sync test has tautological assertion and measures wrong client perspective

[api-Hypernova]

Describe the bug

The groupsync test in measurements/perf-matrix.test.ts has two bugs that make it ineffective:

1. Tautological assertion: The test asserts expect(members.length).toBe(members.length), which compares a value to itself and will always pass regardless of actual member count.
2. Wrong sync perspective: The test measures sync() time from the group creator's client immediately after group creation. Since the creator already has all group data locally, this measures a near-no-op rather than actual network sync performance.

// Current implementation (lines 162-166)it(`groupsync-${i}(${populateSize})[${installationPerMember}]:sync a large group of ${i} members ${i}`, async () => {  await newGroup.sync();  // Creator syncing their own just-created group  const members = await newGroup.members();  expect(members.length).toBe(members.length);  // Always true});

Expected behavior

1. The assertion should verify that the member count equals the expected value (i).
2. The sync should be measured from a receiver's perspective (a member who wasn't the creator) to capture actual network fetch time for group data.

Steps to reproduce the bug

1. Run the performance test suite with BATCH_SIZE=10
2. Observe the groupsync-10 test always passes
3. Modify the group creation to add fewer members than i
4. Observe the test still passes (because members.length === members.length is always true)
5. Check the timing results — sync duration will be minimal since creator already has all data locally