utils/kr_cache_gc: minor nitpicks

It should be no real change, just better readability, hopefully.

utils/kr_cache_gc/db.c

@@ -90,20 +90,20 @@ const uint16_t *kr_gc_key_consistent(knot_db_val_t key)
 {
 	const static uint16_t NSEC1 = KNOT_RRTYPE_NSEC;
 	const static uint16_t NSEC3 = KNOT_RRTYPE_NSEC3;
-	uint8_t *p = key.data;
-	while(*p != 0) {
-		while(*p++ != 0) {
-			if (p - (uint8_t *)key.data >= key.len) {
-				return NULL;
-			}
-		}
+	// find the first double zero in the key; CACHE_KEY_DEF
+	const uint8_t *kd = key.data;
+	ssize_t i;
+	for (i = 2; !(kd[i - 1] == 0 && kd[i - 2] == 0); ++i) {
+		if (i >= key.len) return NULL;
 	}
-	if (p - (uint8_t *)key.data >= key.len) {
-		return NULL;
-	}
-	switch (*++p) {
+	// the next character can be used for classification
+	switch (kd[i]) {
 	case 'E':
-		return (p + 2 - (uint8_t *)key.data >= key.len ? NULL : (uint16_t *)(p + 1));
+		if (i + 1 + sizeof(uint16_t) > key.len) {
+			assert(!EINVAL);
+			return NULL;
+		}
+		return (uint16_t *)&kd[i + 1];
 	case '1':
 		return &NSEC1;
 	case '3':

utils/kr_cache_gc/kr_cache_gc.c

@@ -216,7 +216,7 @@ int kr_cache_gc(kr_cache_gc_cfg_t *cfg)
 		switch (ret) {
 		case KNOT_EOK:
 			deleted_records++;
-			const uint16_t *entry_type = ret == KNOT_EOK ? kr_gc_key_consistent(**i) : NULL;
+			const uint16_t *entry_type = kr_gc_key_consistent(**i);
 			assert(entry_type != NULL);
 			rrtypelist_add(&deleted_rrtypes, *entry_type);
 			break;