Consolidate run_hash and KARVA_RUN_ID into a single identifier

[MatthewMckee4]

Summary

Closes #729. The runner used to mint two identifiers per karva test invocation: run_hash (a millisecond timestamp encoded into the cache directory name and used as a sort key for read_recent_durations / prune_cache), and KARVA_RUN_ID (a UUID v4 exposed to test code for log and artifact correlation). Two identifiers doing nearly the same job meant "where do I get the run id?" had two answers.

RunHash now carries both a millisecond timestamp and a UUID v4. The cache directory becomes run-<ms>-<uuid> so the timestamp prefix preserves the chronological sort that prune and read-recent-durations rely on, and KARVA_RUN_ID becomes the bare <ms>-<uuid> form, derived from the same source of truth. As a side benefit, dense CI matrices that start multiple runs in the same millisecond no longer collide on the cache directory.

from_existing continues to accept legacy run-<ms> directories on disk: a missing UUID falls back to Uuid::nil() and an unparseable timestamp falls back to 0, so prune cleans those up first rather than panicking.

Test plan

ci

[codspeed-hq]

Merging this PR will not alter performance

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

✅ 1 untouched benchmark

---

_{Comparing worktree-ticklish-orbiting-pancake (89ab32f) with main (2cc8ff8)}