From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from localhost (localhost [127.0.0.1]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id 8F42B2E244 for ; Mon, 3 Jun 2019 17:15:15 -0400 (EDT) Received: from turing.freelists.org ([127.0.0.1]) by localhost (turing.freelists.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FMG6-n9ZS6Vj for ; Mon, 3 Jun 2019 17:15:15 -0400 (EDT) Received: from smtp1.mail.ru (smtp1.mail.ru [94.100.179.111]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTPS id 4F9E42DF7A for ; Mon, 3 Jun 2019 17:15:15 -0400 (EDT) Subject: [tarantool-patches] Re: [PATCH v5 3/6] sql: introduce tuple_fetcher class References: <5672cd4b-0c46-b668-efa6-a6a465e601fa@tarantool.org> <20190531194524.GH6141@atlas> <20c02296-1a80-a384-cafa-c061f51ad31d@tarantool.org> From: Vladislav Shpilevoy Message-ID: Date: Mon, 3 Jun 2019 23:15:10 +0200 MIME-Version: 1.0 In-Reply-To: <20c02296-1a80-a384-cafa-c061f51ad31d@tarantool.org> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit Sender: tarantool-patches-bounce@freelists.org Errors-to: tarantool-patches-bounce@freelists.org Reply-To: tarantool-patches@freelists.org List-Help: List-Unsubscribe: List-software: Ecartis version 1.0.0 List-Id: tarantool-patches List-Subscribe: List-Owner: List-post: List-Archive: To: tarantool-patches@freelists.org, Kirill Shcherbatov , Konstantin Osipov Hi! Thanks for the fixes! See 2 comments below, review fixes on the branch, and at the end of the email. > commit 80251ea01b0eeca9cf8bcb50f4b97d97bbcb41ca > Author: Kirill Shcherbatov > Date: Mon May 20 19:28:42 2019 +0300 > > sql: introduce vdbe_field_ref class > > Refactored OP_Column instruction with a new vdbe_field_ref class. > The vdbe_field_ref is a reusable object that speed-up field 1. 'speed' -> 'speeds'. > access for given tuple or tuple data. > Introduced OP_Fetch opcode that uses tuple_fetcher given as a 2. There are no 'tuple_fetcher' anymore. > first argument. This opcode makes possible to perform binding > of a new tuple to an existent VDBE without decoding it's fields. > > Needed for #3691 > I've pushed my review fixes in a separate commit. I didn't fix the commit message, because I can't do it without amendment. Below are my explanations, and diff. 1) I fixed indentation and width > 80. 2) I renamed vdbe_field_ref_tuple_field_fast to vdbe_field_ref_fast_fetch. I didn't like tautology with word 'field' used twice. 3) I dropped a couple of very outdated comments from fetch() function. 4) I fixed a comment for OP_Fetch with declared but unused P4 and P5. 5) I renames field_idx to fieldno in fetch() to be consistent with fast_fetch(). On the whole, the patch is great, even without CHECKs stuff. It strongly improves tuples integration into Vdbe, which beforehand treated them as mere arrays. ================================================================== diff --git a/src/box/sql.c b/src/box/sql.c index b4702fd78..f7b90ab5c 100644 --- a/src/box/sql.c +++ b/src/box/sql.c @@ -1368,7 +1368,7 @@ vdbe_field_ref_create(struct vdbe_field_ref *field_ref, struct tuple *tuple, void vdbe_field_ref_prepare_data(struct vdbe_field_ref *field_ref, const char *data, - uint32_t data_sz) + uint32_t data_sz) { vdbe_field_ref_create(field_ref, NULL, data, data_sz); } @@ -1377,5 +1377,6 @@ void vdbe_field_ref_prepare_tuple(struct vdbe_field_ref *field_ref, struct tuple *tuple) { - vdbe_field_ref_create(field_ref, tuple, tuple_data(tuple), tuple->bsize); + vdbe_field_ref_create(field_ref, tuple, tuple_data(tuple), + tuple->bsize); } diff --git a/src/box/sql.h b/src/box/sql.h index 066a0d322..3056d8f7e 100644 --- a/src/box/sql.h +++ b/src/box/sql.h @@ -430,7 +430,7 @@ struct vdbe_field_ref { */ void vdbe_field_ref_prepare_data(struct vdbe_field_ref *field_ref, const char *data, - uint32_t data_sz); + uint32_t data_sz); /** * Initialize a new vdbe_field_ref instance with given tuple diff --git a/src/box/sql/vdbe.c b/src/box/sql/vdbe.c index 256ee77a0..3f0b856b3 100644 --- a/src/box/sql/vdbe.c +++ b/src/box/sql/vdbe.c @@ -623,8 +623,8 @@ mem_type_to_str(const struct Mem *p) * offset to @a fieldno. */ static const void * -vdbe_field_ref_tuple_field_fast(struct vdbe_field_ref *field_ref, - uint32_t fieldno, uint32_t *field_size) +vdbe_field_ref_fast_fetch(struct vdbe_field_ref *field_ref, uint32_t fieldno, + uint32_t *field_size) { if (field_ref->tuple == NULL) return NULL; @@ -641,38 +641,30 @@ vdbe_field_ref_tuple_field_fast(struct vdbe_field_ref *field_ref, } /** - * Fetch field by field_idx using vdbe_field_ref and store result + * Fetch field by fieldno using vdbe_field_ref and store result * in dest_mem. * @param field_ref The initialized vdbe_field_ref instance to use. - * @param field_idx The id of the field to fetch. + * @param fieldno The id of the field to fetch. * @param[out] dest_mem The memory variable to store result. * @retval SQL_OK Status code in case of success. * @retval sql_ret_code Error code otherwise. */ static int -vdbe_field_ref_fetch(struct vdbe_field_ref *field_ref, uint32_t field_idx, - struct Mem *dest_mem) +vdbe_field_ref_fetch(struct vdbe_field_ref *field_ref, uint32_t fieldno, + struct Mem *dest_mem) { sqlVdbeMemSetNull(dest_mem); uint32_t *slots = field_ref->slots; - if (field_idx >= field_ref->field_count) { + if (fieldno >= field_ref->field_count) { UPDATE_MAX_BLOBSIZE(dest_mem); return SQL_OK; } - /* - * Make sure at least the first field_idx + 1 entries - * of the header have been parsed and valid information - * is in field_map[]. - * If there is more header available for parsing in the - * record, try to extract additional fields up through the - * field_map+1-th field. - */ const char *data; - if (field_ref->rightmost_slot <= field_idx) { + if (field_ref->rightmost_slot <= fieldno) { uint32_t field_sz; - data = vdbe_field_ref_tuple_field_fast(field_ref, field_idx, - &field_sz); + data = vdbe_field_ref_fast_fetch(field_ref, fieldno, + &field_sz); if (data != NULL) { /* * Special case for tarantool spaces: for @@ -685,27 +677,22 @@ vdbe_field_ref_fetch(struct vdbe_field_ref *field_ref, uint32_t field_idx, * visited in mp_next() cycle. */ uint32_t offset = (uint32_t)(data - field_ref->data); - slots[field_idx] = offset; - slots[field_idx + 1] = offset + field_sz; + slots[fieldno] = offset; + slots[fieldno + 1] = offset + field_sz; } else { uint32_t i = field_ref->rightmost_slot; data = field_ref->data + slots[i]; do { mp_next(&data); slots[++i] = (uint32_t)(data - field_ref->data); - } while (i <= field_idx); + } while (i <= fieldno); field_ref->rightmost_slot = i; } } - /* - * Extract the content for the p2+1-th column. Control - * can only reach this point if field_map[field_idx], - * field_map[field_idx+1] are valid. - */ assert(sqlVdbeCheckMemInvariants(dest_mem) != 0); uint32_t dummy; - data = field_ref->data + slots[field_idx]; + data = field_ref->data + slots[fieldno]; if (vdbe_decode_msgpack_into_mem(data, dest_mem, &dummy) != 0) return SQL_TARANTOOL_ERROR; @@ -714,7 +701,7 @@ vdbe_field_ref_fetch(struct vdbe_field_ref *field_ref, uint32_t field_idx, * Wrap it in a blob verbatim. */ if (dest_mem->flags == 0) { - dest_mem->n = slots[field_idx + 1] - slots[field_idx]; + dest_mem->n = slots[fieldno + 1] - slots[fieldno]; dest_mem->z = (char *) data; dest_mem->flags = MEM_Blob | MEM_Ephem | MEM_Subtype; dest_mem->subtype = SQL_SUBTYPE_MSGPACK; @@ -723,7 +710,8 @@ vdbe_field_ref_fetch(struct vdbe_field_ref *field_ref, uint32_t field_idx, * Add 0 termination (at most for strings) * Not sure why do we check MEM_Ephem */ - if ((dest_mem->flags & (MEM_Ephem | MEM_Str)) == (MEM_Ephem | MEM_Str)) { + if ((dest_mem->flags & (MEM_Ephem | MEM_Str)) == + (MEM_Ephem | MEM_Str)) { int len = dest_mem->n; if (dest_mem->szMalloc < len + 1) { if (sqlVdbeMemGrow(dest_mem, len + 1, 1) != 0) @@ -2801,7 +2789,7 @@ op_column_out: break; } -/* Opcode: Fetch P1 P2 P3 P4 P5 +/* Opcode: Fetch P1 P2 P3 * * * Synopsis: r[P3]=PX * * Interpret data P1 points at as an initialized vdbe_field_ref