polygon/heimdall: fix snapshot store RangeFromBlockNum

[taratorio]

fixes #13674

issue was that after rm -rf chaindata:

1. txEntityStore.EntityIdFromBlockNum returns an entity id (in this case a Checkpoint entity id) whose corresponding entity is no longer in the DB (in this case erigon/datadir/heimdall/mdbx.dat) because it has been retired into its corresponding snapshot file
2. then, in turn, txEntityStore.RangeFromBlockNum returns only the Checkpoint entities which it has in the DB (they are all with Start block num greater than the one in 1) - this creates a gap of X number of checkpoints
3. then, in turn, when we download all the corresponding blocks for the checkpoints in 2) we hit a pos sync failed: block gap inserted error which is because of the gap in checkpoints between start of downloading new blocks and start of first checkpoint in 2)

Managed to reproduce the above by:

• deleting last 10k worth of snapshots for headers.seg, bodies.seg, transactions.seg and corresponding .idx files
• rm -rf chaindata
• adding the pos sync failed: unexpected first checkpoint with id 15274 has start 17673843 which is greater than download start 17660000 err check and hitting it after the above 2 steps

[taratorio]

wait until #13702 is merged and issue with bor checkpoints snapshots generation is resolved (needs a bug fix and regeneration of snapshots)

[taratorio]

need to re-test once we fix snapshot generation logic for borcheckpoints and bormilestones to avoid gaps

[taratorio]

e.g. problem is in a gap like this (FirstInDb and LastInSnapshots should be connecting)

entityId_FromBlockNum: 15265
entityId_FirstInDb: 15274
entityId_LastInSnapshotsInReverse:15261 kind=*heimdall.Checkpoint