fix(rust_brain): preserve HLC on snapshot restore and gossip receive

[cursor]

Bug and impact

restore_from_file() and gossip.receive() did not preserve or apply Hybrid Logical Clock (HLC) timestamps introduced in v0.6.0. After snapshot restore, any causal successor write with a pre-crash HLC was rejected with TimestampRegression, causing silent data loss in disaster-recovery and cross-node gossip scenarios.

Concrete trigger:

1. Node writes remember("k", "v1", hlc=(5000, 10, "nodeA"))
2. Snapshot is taken and restored on a fresh brain
3. Restored node gets a fresh HLC (wall, 0, uuid) instead of (5000, 10, "nodeA")
4. Causal successor remember("k", "v2", hlc=(5000, 11, "nodeA")) raises TimestampRegression

Root cause

restore_from_file() reconstructed MemoryNode without the hlc field, falling back to _hlc.now(). gossip.receive() similarly omitted hlc/ts_ns from incoming events.

Fix

• Add _parse_hlc() for wire/snapshot normalisation (with legacy ts_ns fallback)
• Restore hlc field and advance global HLC during restore_from_file()
• Pass hlc through bulk_write and gossip.receive
• Skip gossip updates missing HLC on existing keys (cannot establish causal order)
• Hold lock during restore to prevent concurrent writes

Validation

• Reproduced bug before fix (successor write fails after restore)
• Added regression tests: test_restore_preserves_hlc_for_replication, gossip HLC ordering tests
• Full suite: 191 passed, 8 skipped