[GCU] Using simulated config instead of target config when validating replace operation in NoDependencyMoveValidator

[ghooo]

What I did

Using SimulatedConfig instead of TargetConfig for validating a move using NoDependencyMoveValidator

SimulatedConfig: config after applying the given move
TargetConfig: the final config state that's required by the patch

The problem is if the moves is to update a list item, the list item location in the TargetConfig might have different location in the CurrentConfig.

The reason for that is the TargetConfig has the final outcome after applying the patch, but the move might be just making a small change towards this goal.

Example:
Assume current_config

{
    "VLAN": {
        "Vlan100": {
            "vlanid": "100",
            "dhcp_servers": [
                "192.0.0.1",
                "192.0.0.2"
            ]
        }
    }
}

TargetConfig:

{
    "VLAN": {
        "Vlan100": {
            "vlanid": "100",
            "dhcp_servers": [
                "192.0.0.3"
            ]
        }
    }
}

Move:

ps.JsonMove(diff, OperationType.REPLACE, ["VLAN", "Vlan100", "dhcp_servers", 1], ["VLAN", "Vlan100", "dhcp_servers", 0])

The move means:

Replace the value in CurrentConfig that has path `/VLAN/Vlan100/dhcp_servers/1` with
the value from the TargetConfig that has the path `/VLAN/Vlan100/dhcp_servers/0`

Notice how the array index in CurrentConfig does not exist in TargetConfig

Instead of using TargetConfig to validate, use SimulatedConfig which is the config after applying the move. In this case it would be:

{
    "VLAN": {
        "Vlan100": {
            "vlanid": "100",
            "dhcp_servers": [
                "192.0.0.1",
                "192.0.0.3"
            ]
        }
    }
}

How I did it

Replace diff.target_config with simulated_config

How to verify it

added a unit-test

[ghooo]

small change not related to the PR.

The MemoizationSorter is not used so we have not caught this before. This sorter is just added to run experiments on the performance.