Decoding of UtxoId with multibyte characters can panic

[Dentosal]

The following query panics:

{ 
 coin(utxoId:  "0x00😎"  )  {  utxoId  } 
}

This is caused by careless use of string indexing here:

fuel-vm/fuel-tx/src/transaction/types/utxo_id.rs

Lines 104 to 118 in 651351f

	fn from_str(s: &str) -> Result<Self, Self::Err> {
	const ERR: &str = "Invalid encoded byte";
	let s = s.trim_start_matches("0x");
	let utxo_id = if s.is_empty() {
	UtxoId::new(Bytes32::default(), 0)
	} else if s.len() > 2 {
	UtxoId::new(
	Bytes32::from_str(&s[..s.len() - 2])?,
	u8::from_str_radix(&s[s.len() - 2..], 16).map_err(|_| ERR)?,
	)
	} else {
	UtxoId::new(TxId::default(), u8::from_str_radix(s, 16).map_err(|_| ERR)?)
	};
	Ok(utxo_id)
	}

We should probably use #[deny(clippy::string_slice)] to avoid these issues in the future.