Actual status-driver comparison for PR #19825

Head: 26aa1b72c1cf93c163e367db83b86d190269ab79
Base: 687bd555333fd44bb7eef27110b6c08b19e23b23
Linux, v24.21.0, git version 2.43.0

Eight alternating before/after pairs per scenario and driver lifetime, warm filesystem cache. 100 base files, one feature commit, 20 modified files, a local origin remote with main and feature tracking refs. Configured base main; fallback control missing-base. Local reads include Changes totals and disable divergence; remote reads have an upstream and disable fetching. Driver construction excluded. Warm driver reused within case; fresh driver constructed per sample. Actual source factories from the exact base/head, real Git subprocesses; proxy only records commands. All 48 pairs return identical status values. This measures the driver, not GUI rendering or network latency. Shared-host timings are descriptive; command elimination is deterministic.

| Scenario | Driver | Git calls main → head | Origin HEAD reads main → head | Median main → head (ms) | Median paired saving (ms) |
| --- | --- | --- | --- | --- | --- |
| configured-local | warm | 10 → 9 | 1 → 0 | 75.7 → 72.1 | 12.0 |
| configured-local | fresh | 15 → 14 | 1 → 0 | 103.5 → 95.8 | 7.0 |
| missing-local | warm | 12 → 12 | 1 → 1 | 73.3 → 73.5 | 2.6 |
| missing-local | fresh | 17 → 17 | 1 → 1 | 94.5 → 91.3 | 5.4 |
| upstream-remote | warm | 10 → 9 | 2 → 1 | 55.2 → 50.1 | 6.9 |
| upstream-remote | fresh | 10 → 9 | 2 → 1 | 55.7 → 47.5 | 4.9 |