Replace net.IPNet to netip.Prefix for performances purposes

[tomMoulard]

This PR replaces the usage of net.IPNet to netip.Prefix for performances purposes.

Following the netip doc:

Compared to the net.IP type, Addr type takes less memory, is immutable, and is comparable (supports == and being a map key).

here are a few benchmarks (results may vary)

> go test -bench=. github.com/tomMoulard/fail2ban/ipchecking -test.run=Benchmark
goos: linux
goarch: amd64
pkg: github.com/tomMoulard/fail2ban/ipchecking
cpu: Intel(R) Core(TM) i7-10750H CPU @ 2.60GHz
BenchmarkIPChecking_new-12    	 1000000	      1042 ns/op
BenchmarkIPChecking_old-12    	  512394	      2307 ns/op
PASS
ok  	github.com/tomMoulard/fail2ban/ipchecking	2.618s

func BenchmarkIPChecking_new(b *testing.B) {
	tests := []struct {
		ip1    string
		ip2    string
		expect bool
	}{
		{ip1: "127.0.0.1", ip2: "127.0.0.1", expect: true},
		{ip1: "127.0.0.1", ip2: "127.0.0.2", expect: false},
		{ip1: "127.0.0.1/24", ip2: "127.0.0.2", expect: true},
		{ip1: "127.0.0.1/32", ip2: "127.0.0.2", expect: false},
		{ip1: "10.0.0.1/32", ip2: "127.0.0.2", expect: false},
		{ip1: "10.0.0.1/0", ip2: "127.0.0.2", expect: true},
		{ip1: "0.0.0.0", ip2: "127.0.0.2", expect: false},
	}

	for n := 0; n < b.N; n++ {
		for _, tt := range tests {
			ipv4, _ := ipchecking.ParseNetIP(tt.ip1)
			if ipv4.Contains(tt.ip2) != tt.expect {
				b.Logf("ip1: %s, ip2: %s", tt.ip1, tt.ip2)
				b.Errorf("wanted '%v' got '%v'", tt.expect, ipv4.Contains(tt.ip2))
			}
		}
	}
}

func BenchmarkIPChecking_old(b *testing.B) {
	tests := []struct {
		ip1    string
		ip2    string
		expect bool
	}{
		{ip1: "127.0.0.1", ip2: "127.0.0.1", expect: true},
		{ip1: "127.0.0.1", ip2: "127.0.0.2", expect: false},
		{ip1: "127.0.0.1/24", ip2: "127.0.0.2", expect: true},
		{ip1: "127.0.0.1/32", ip2: "127.0.0.2", expect: false},
		{ip1: "10.0.0.1/32", ip2: "127.0.0.2", expect: false},
		{ip1: "10.0.0.1/0", ip2: "127.0.0.2", expect: true},
		{ip1: "0.0.0.0", ip2: "127.0.0.2", expect: false},
	}

	for n := 0; n < b.N; n++ {
		for _, tt := range tests {
			ipv4, _ := ipchecking.BuildIP(tt.ip1)
			if ipv4.CheckIPInSubnet(tt.ip2) != tt.expect {
				b.Logf("ip1: %s, ip2: %s", tt.ip1, tt.ip2)
				b.Errorf("wanted '%v' got '%v'", tt.expect, ipv4.CheckIPInSubnet(tt.ip2))
			}
		}
	}
}

There is one breaking change in the ipchecking API (except for the functions renaming to be closer to netip function calls) that is StrToIP now breaks when some IPs are faulty instead of ignoring them.