From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <v.shpilevoy@tarantool.org>
Received: from smtpng3.m.smailru.net (smtpng3.m.smailru.net [94.100.177.149])
 (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256
 bits)) (No client certificate requested)
 by dev.tarantool.org (Postfix) with ESMTPS id 4A4BB4696C3
 for <tarantool-patches@dev.tarantool.org>;
 Tue, 28 Apr 2020 00:29:17 +0300 (MSK)
References: <cover.1587600640.git.v.shpilevoy@tarantool.org>
 <0d4129e35ccfa9cbfd8e4204e8179acb545cf46b.1587600640.git.v.shpilevoy@tarantool.org>
 <20200427151408.GD30870@tarantool.org>
From: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
Message-ID: <8c6e8f5f-f5b4-48f3-6214-cbc62be3af8f@tarantool.org>
Date: Mon, 27 Apr 2020 23:29:15 +0200
MIME-Version: 1.0
In-Reply-To: <20200427151408.GD30870@tarantool.org>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Subject: Re: [Tarantool-patches] [PATCH v2 2/3] box: introduce
	box_return_mp() public C function
List-Id: Tarantool development patches <tarantool-patches.dev.tarantool.org>
List-Unsubscribe: <https://lists.tarantool.org/mailman/options/tarantool-patches>, 
 <mailto:tarantool-patches-request@dev.tarantool.org?subject=unsubscribe>
List-Archive: <https://lists.tarantool.org/pipermail/tarantool-patches/>
List-Post: <mailto:tarantool-patches@dev.tarantool.org>
List-Help: <mailto:tarantool-patches-request@dev.tarantool.org?subject=help>
List-Subscribe: <https://lists.tarantool.org/mailman/listinfo/tarantool-patches>, 
 <mailto:tarantool-patches-request@dev.tarantool.org?subject=subscribe>
To: Nikita Pettik <korablev@tarantool.org>
Cc: tarantool-patches@dev.tarantool.org

Thanks for the question!

On 27/04/2020 17:14, Nikita Pettik wrote:
> On 23 Apr 02:12, Vladislav Shpilevoy wrote:
>> +
>> +int
>> +test_return_mp(box_function_ctx_t *ctx, const char *args, const char *args_end)
>> +{
>> +	(void) args;
>> +	(void) args_end;
>> +	char buf[512];
>> +	char *pos = mp_encode_uint(buf, 1);
>> +	int rc = box_return_mp(ctx, buf, pos);
>> +	if (rc != 0)
>> +		return rc;
>> +
>> +	pos = mp_encode_int(buf, -1);
>> +	rc = box_return_mp(ctx, buf, pos);
>> +	if (rc != 0)
>> +		return rc;
>> +
>> +	pos = mp_encode_uint(buf, UINT64_MAX);
>> +	rc = box_return_mp(ctx, buf, pos);
>> +	if (rc != 0)
>> +		return rc;
>> +
>> +	const char *str = "123456789101112131415";
>> +	pos = mp_encode_str(buf, str, strlen(str));
>> +	rc = box_return_mp(ctx, buf, pos);
>> +	if (rc != 0)
>> +		return rc;
>> +
>> +	pos = mp_encode_array(buf, 1);
>> +	pos = mp_encode_uint(pos, 2);
>> +	box_tuple_t *tuple = box_tuple_new(box_tuple_format_default(),
>> +					   buf, pos);
>> +	if (tuple == NULL)
>> +		return -1;
>> +	rc = box_return_tuple(ctx, tuple);
> 
> Probably I'm missing something (since I've never used C API) but
> when I do:
> 
> res = net:connect(box.cfg.listen):call(name)
> print(type(res[5]))
> 
> I get table. But shouldn't it be tuple (cdata), since it (last member)
> is explicitly wrapped into tuple)? I guess it is due to certain
> convention but failed to find it in docs. Could you please clarify?

That is not about C functions. It is about tuple data type lost during
its marshaling over network. You will get the same result, if you
write that in a Lua stored function. We solve that problem by introducing
MP_EXT to be able to restore types on the client side for uuids, errors,
decimals. But not for tuples. So they turn into tables when
serialized-deserialized from MessagePack during sending over the network.

Exception is IPROTO_SELECT and other DML/DQL mapped onto exact IPROTO_*
commands, because client always knows they return tuples, and can create
them on the client side instead of tables.