Light client fcU should not spam the EL

[arnetheduck]

INF 2026-02-25 23:23:14.770+01:00 Forkchoice requested sync to new head      topics="beacon engine" number=24536880 hash=c004e97148ca base=0 finHash=000000000000 safe=000000000000 pendingFCU=000000000000 resolvedFinNum=0 resolvedFinHash=000000000000 comp=ec
INF 2026-02-25 23:23:14.824+01:00 Forkchoice requested sync to new head      topics="beacon engine" number=24536880 hash=c004e97148ca base=0 finHash=000000000000 safe=000000000000 pendingFCU=000000000000 resolvedFinNum=0 resolvedFinHash=000000000000 comp=ec
INF 2026-02-25 23:23:14.877+01:00 Forkchoice requested sync to new head      topics="beacon engine" number=24536880 hash=c004e97148ca base=0 finHash=000000000000 safe=000000000000 pendingFCU=000000000000 resolvedFinNum=0 resolvedFinHash=000000000000 comp=ec
INF 2026-02-25 23:23:14.880+01:00 Forkchoice requested sync to new head      topics="beacon engine" number=24536881 hash=945ef8a5d1f1 base=0 finHash=000000000000 safe=000000000000 pendingFCU=000000000000 resolvedFinNum=0 resolvedFinHash=000000000000 comp=ec
INF 2026-02-25 23:23:14.933+01:00 Forkchoice requested sync to new head      topics="beacon engine" number=24536881 hash=945ef8a5d1f1 base=0 finHash=000000000000 safe=000000000000 pendingFCU=000000000000 resolvedFinNum=0 resolvedFinHash=000000000000 comp=ec
INF 2026-02-25 23:23:14.977+01:00 Forkchoice requested sync to new head      topics="beacon engine" number=24536881 hash=945ef8a5d1f1 base=0 finHash=000000000000 safe=000000000000 pendingFCU=000000000000 resolvedFinNum=0 resolvedFinHash=000000000000 comp=ec

When syncing, fork choice is called quite often, like in the above case where it happens every block - however, the light client has already established a newer head and keeps repeating it to the execution client. We should detect when this happens and only send actually updated heads.