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 a snapshot restore, nodes were assigned fresh HLCs instead of the stored values. Legitimate causal successor writes (e.g. from the original writer or a peer node) were rejected as TimestampRegression, causing silent data loss in disaster-recovery and multi-node gossip scenarios.

Concrete trigger:

1. Node writes remember("k", "v1", hlc=(5000, 10, "nodeA"))
2. Snapshot 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")) is rejected

Root cause

restore_from_file() constructed MemoryNode without passing the snapshot's hlc field, so the dataclass default factory generated a new HLC. gossip.receive() similarly omitted HLC passthrough. bulk_write() also did not forward HLC from row data.

Fix

• Add _parse_hlc() for snapshot/wire HLC normalisation (with legacy fallback from ts_ns)
• Restore HLC on snapshot load and advance global _hlc via _hlc.update()
• Pass HLC through bulk_write and gossip.receive()
• Reject stale gossip updates; skip HLC-less updates on existing keys
• Hold lock during restore to prevent concurrent mutation

Validation

• Reproduced bug before fix (successor write rejected after restore)
• Added regression tests in test_enterprise_backup.py and test_gossip.py
• Full suite: 192 passed, 8 skipped