fix(rust_brain): preserve HLC across snapshot restore and gossip

[cursor]

Bug and impact

restore_from_file(), bulk_write(), and gossip.receive() did not preserve or apply Hybrid Logical Clock (HLC) timestamps introduced in v0.6.0. After snapshot restore, gossip replays and causally valid successor writes were rejected with TimestampRegression, causing silent data loss in disaster-recovery and multi-node deployments.

Concrete trigger: Node A writes key k with HLC (5000, 10, nodeA), snapshots, crashes. Node B restores the snapshot. Node A (or gossip) attempts a successor write with HLC (5000, 11, nodeB) — rejected because restore assigned a fresh wall-clock HLC instead of the snapshot value.

Root cause

v0.6.0 added HLC monotonic enforcement to remember() and serializes HLC in to_dict(), but restore_from_file() created MemoryNode without reading hlc from the snapshot, bulk_write() did not pass hlc, and gossip.receive() ignored HLC on incoming events.

Fix

• Add _parse_hlc() for snapshot/wire normalization (with legacy fallback for pre-v0.6.0 snapshots)
• Restore HLC on snapshot load and advance the global HLC via _hlc.update()
• Pass HLC through bulk_write and gossip.receive()
• Hold lock during restore mutation
• Gossip: skip updates missing HLC when key already exists; apply HLC for causal ordering

Validation

• Reproduced pre-fix: restore assigned fresh HLC; successor write raised TimestampRegression
• Post-fix reproduction script passes
• 29 targeted tests pass: test_hlc_snapshot_gossip.py, test_enterprise_backup.py, test_gossip.py, test_rust_brain.py