Report per operation timings in the editor performance tests

[vogella]

The editor performance tests measured opening plus closing in a single block, and twelve editor switches in another, so a change in one operation could not be attributed, and nothing at all was printed when no performance database is configured. Both tests now time the individual operations and log min, median, 90th percentile and maximum, they warm up before measuring so that class loading and JIT no longer land in the first measured block, and they are bounded by time instead of a fixed iteration count. EditorSwitchTest drops from roughly 145 to 35 seconds while collecting more samples than before.

The recorded series are unchanged, so the existing performance database path and the "UI - Open/Close Editor" summary keep working as before.

This is a first step towards measuring and then improving editor handling. Sample output of a local run:

OpenCloseEditor[perf_text] open      n=210  min=  32.23  p50=  39.72  p90=  56.98  max=  72.49 (ms)
OpenCloseEditor[perf_text] close     n=210  min=  10.29  p50=  15.91  p90=  21.73  max=  32.16 (ms)
EditorSwitch to [perf_outline]       n=180  min=   5.13  p50=  15.44  p90=  22.83  max=  29.38 (ms)
EditorSwitch to [perf_text]          n=180  min=   5.46  p50=  13.79  p90=  21.70  max=  35.02 (ms)

[github-actions]

Test Results

   858 files  ±0     858 suites  ±0   43m 39s ⏱️ - 8m 11s
 8 130 tests ±0   7 887 ✅ ±0  243 💤 ±0  0 ❌ ±0 
20 292 runs  ±0  19 638 ✅ ±0  654 💤 ±0  0 ❌ ±0 

Results for commit 5cfa04d. ± Comparison against base commit f78bd54.