On Mar 27, 2013, at 2:44 PM, Ben Pfaff <b...@nicira.com> wrote:
> +static size_t
> +json_string_serialized_length(const char *string)
> +{
> + size_t length;
> + uint8_t c;
> +
> + /* "" */
> + length = 2;
I think this length represents the surrounding quotes, but until I looked at
json_serialize_string(), I thought it was for just a single quote based on this
comment.
> +/* Returns strlen(json_to_string(json, 0)).
> + *
> + * JSSF_SORT does not affect the length of json_to_string()'s output, but
> + * JSSF_PRETTY does. */
Is it worth mentioning that this is treating it as not pretty?
> +size_t
> +json_serialized_length(const struct json *json)
> +{
> + switch (json->type) {
> + case JSON_NULL:
> + return strlen("null");
> + break;
> +
> + case JSON_FALSE:
> + return strlen("false");
> + break;
> +
> + case JSON_TRUE:
> + return strlen("true");
> + break;
> +
> + case JSON_OBJECT:
> + return json_object_serialized_length(json->u.object);
> +
> + case JSON_ARRAY:
> + return json_array_serialized_length(&json->u.array);
> +
> + case JSON_INTEGER:
> + return snprintf(NULL, 0, "%lld", json->u.integer);
> +
> + case JSON_REAL:
> + return snprintf(NULL, 0, "%.*g", DBL_DIG, json->u.real);
> +
> + case JSON_STRING:
> + return json_string_serialized_length(json->u.string);
> +
> + case JSON_N_TYPES:
> + default:
> + NOT_REACHED();
> + }
> +}
Is there a reason for the inconsistency with some case blocks have a "break"
and others not? The json_serialize() function uses "break" in all of them.
--Justin
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
