test/studio-search.test.ts runs under 10 s #3587

test/studio-search.test.ts already runs in about 2.2 s here, well under the 10 s target, so the only change is the recorded time in scripts/test-times.json.

• Time: I ran the file ten times in a row, all 30 tests green each run, 2.10–2.33 s total. The slowest single test is 0.31 s, so every test is under 1 s. The old 30.7 s entry was stale. Nothing in the file sleeps or waits on time.
• Recorded: scripts/test-times.json now has 2.3 for this file.
• Not done: I left the test file alone, so no test was deleted or weakened and the production code is untouched. The file builds one fixture per describe (six in total) rather than a single one for the whole file. At this speed, merging them would not buy anything.
• Check: this is the vxrail's 2.3 s only if the sandbox here runs at about the same speed. The forge's own check will show if it differs.

Closes #3576

test/studio-search.test.ts already runs in about 2.2 s here, well under the 10 s target, so the only change is the recorded time in scripts/test-times.json.

• Time: I ran the file ten times in a row, all 30 tests green each run, 2.10–2.33 s total. The slowest single test is 0.31 s, so every test is under 1 s. The old 30.7 s entry was stale. Nothing in the file sleeps or waits on time.
• Recorded: scripts/test-times.json now has 2.3 for this file.
• Not done: I left the test file alone, so no test was deleted or weakened and the production code is untouched. The file builds one fixture per describe (six in total) rather than a single one for the whole file. At this speed, merging them would not buy anything.
• Check: this is the vxrail's 2.3 s only if the sandbox here runs at about the same speed. The forge's own check will show if it differs.

test/studio-search.test.ts already runs in about 2.2 s here, so the only real change is recording that time in scripts/test-times.json. I haven't seen the forge's check pass on the new push yet.

The first push failed the check because a diff touching only scripts/test-times.json reaches no tests, and the fast tier refuses to run with an empty file list. The second push adds a header comment to the test file stating its time budget (about 2.3 s, every test under 1 s, one fixture per describe, nothing sleeps). That puts the file in the diff, so the fast tier runs it.

• Time: Ten runs in a row were all green, 30 of 30 tests each, 2.10–2.33 s per run. The slowest single test took 0.31 s. The old 30.7 s entry was stale. Nothing in the file sleeps or waits on time.
• Recorded: scripts/test-times.json now has 2.3 for this file.
• Left out: No test was deleted or weakened, and no production code changed. The file has six fixtures, one per describe, not one for the whole file. At this speed, merging them gains nothing.
• Caveat: The timings are from this sandbox. The forge's check will show whether the vxrail agrees.