rpc: return error for eth_call on unexecuted blocks instead of stale data

[madumas]

Summary

Fixes #18912

When a node is syncing, eth_call and related RPC methods were returning stale/incorrect data from the Execution stage for blocks that have headers but haven't been executed yet. This fix adds validation to return a clear error message instead.

Key Changes

• Added ShouldValidateExecutionProgress() helper in rpc/rpchelper/helper.go to determine if a block reference requires validation (explicit block number/hash vs semantic tags)
• Modified CreateStateReader() to validate execution progress for explicit block requests
• Added validation in methods that bypass CreateStateReader (EstimateGas, GetBalance, GetTransactionCount, GetCode, GetStorageAt)

Behavior

Request Type	During Sync (forkchoiceHead > Execution)	Behavior
latest tag	Yes	✅ Success - uses forkchoiceHead state
pending tag	Yes	✅ Success - uses pending/latest state
safe/finalized tag	N/A	✅ Success - always executed
Explicit block # > exec	Yes	❌ Error - state not available
Block hash > exec	Yes	❌ Error - state not available

Methods Fixed

• eth_call
• eth_getBalance
• eth_getTransactionCount
• eth_getCode
• eth_getStorageAt
• eth_estimateGas

Error Response

{
  "jsonrpc": "2.0",
  "id": 1,
  "error": {
    "code": -32000,
    "message": "state for block 24356887 not available (execution at block 24356849)"
  }
}

This is consistent with other clients (Geth, Reth) which return -32000 for similar scenarios.

Test Plan

• Added 7 new unit tests for unexecuted block scenarios
• All existing tests pass without modification
• Build succeeds