Decode: use unsafe []byte->string conversion for discarded value

For Decode's MLI4I, we invoke strconv.Atoi after string(*b)
which incurs a []byte->string allocation. The converted []byte->string
is discarded so anyways it is okay to use unsafe, as the string is
never used again.

Results:

* time/op (ns/op)
Encoding/Decoding_2I-8	7.5ns ± 1%  6.7ns ± 0%  -10.36%	(p=0.000 n=10+9)
Encoding/Decoding_4E-8	8.8ns ± 0%  7.9ns ± 0%  -10.35%	(p=0.000 n=10+10)
Encoding/Decoding_A4E-8	38ns ± 1%   17ns ± 0%   -54.84%	(p=0.000 n=10+10)

* allocs/op (B/op)
Encoding/Decoding_A4E-8	4.0B ± 0%   0	-100.00% (p=0.000 n=10+10)

* allocs/op (N/op)
Encoding/Decoding_A4E-8	1.0 ± 0%    0	-100.00% (p=0.000 n=10+10)

mli.go

@@ -71,6 +71,7 @@ import (
 	"encoding/hex"
 	"fmt"
 	"strconv"
+	"unsafe"
 )
 
 // empty is used as a quick return during errors
@@ -238,7 +239,7 @@ func Decode(key string, b *[]byte) (int, error) {
 		}
 
 		// Convert to integer from ASCII
-		n, err := strconv.Atoi(string(*b))
+		n, err := strconv.Atoi(unsafeByteToStr(*b))
 		if err != nil {
 			return 0, fmt.Errorf("unable to convert string values to integer - %s", err)
 		}
@@ -249,6 +250,10 @@ func Decode(key string, b *[]byte) (int, error) {
 	}
 }
 
+func unsafeByteToStr(b []byte) string {
+	return *(*string)(unsafe.Pointer(&b))
+}
+
 // Encode will accept a message length type and message length value desired. Encode will return a byte slice which
 // contains a MLI formatted for in the desired message length type.
 //