Tarantool development patches archive
 help / color / mirror / Atom feed
From: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
To: Kirill Shcherbatov <kshcherbatov@tarantool.org>,
	tarantool-patches@freelists.org
Subject: [tarantool-patches] Re: [PATCH v4 3/4] box: introduce column_mask for ck constraint
Date: Sun, 19 May 2019 19:02:10 +0300	[thread overview]
Message-ID: <148cfacb-757b-0c4a-7b7d-efaa18c6ffac@tarantool.org> (raw)
In-Reply-To: <16fabe763ee57c8cd5c807b24d0fc500ad5a8bae.1558014727.git.kshcherbatov@tarantool.org>

Thanks for the patch! See 3 comments below.

On 16/05/2019 16:56, Kirill Shcherbatov wrote:
> This patch introduces a smart bitmask of fields are mentioned in

1. Remove 'are'. 'mentioned' here works as adjective. The same in
other places. Please, find them by yourself.

> ck constraint. This allows to do not bind useless fields and
> do not parse more fields that is required.
> 
> Needed for #3691
> ---
>  src/box/ck_constraint.c  | 66 ++++++++++++++++++++++++++++++++++------
>  src/box/ck_constraint.h  |  6 ++++
>  src/box/sql.h            |  5 ++-
>  src/box/sql/build.c      |  2 +-
>  src/box/sql/resolve.c    | 19 +++++++++---
>  src/box/sql/sqlInt.h     |  6 ++++
>  test/sql/checks.result   | 59 +++++++++++++++++++++++++++++++++++
>  test/sql/checks.test.lua | 20 ++++++++++++
>  8 files changed, 167 insertions(+), 16 deletions(-)
> 
> diff --git a/src/box/ck_constraint.c b/src/box/ck_constraint.c
> index 43798be76..4756f7970 100644
> --- a/src/box/ck_constraint.c
> +++ b/src/box/ck_constraint.c
> @@ -46,17 +46,21 @@ const char *ck_constraint_language_strs[] = {"SQL"};
>   * tree traversal.
>   * @param ck_constraint Check constraint object to update.
>   * @param space_def Space definition to use.
> + * @param column_mask[out] The "smart" column mask of fields are
> + *                         referenced by AST.
>   * @retval 0 on success.
>   * @retval -1 on error.
>   */
>  static int
>  ck_constraint_resolve_field_names(struct Expr *expr,
> -				  struct space_def *space_def)
> +				  struct space_def *space_def,
> +				  uint64_t *column_mask)
>  {
>  	struct Parse parser;
>  	sql_parser_create(&parser, sql_get(), default_flags);
>  	parser.parse_only = true;
> -	sql_resolve_self_reference(&parser, space_def, NC_IsCheck, expr, NULL);
> +	sql_resolve_self_reference(&parser, space_def, NC_IsCheck, expr, NULL,
> +				   column_mask);

2. Next to last argument 'expr_list' is always NULL. Please, drop it.

>  	int rc = parser.is_aborted ? -1 : 0;
>  	sql_parser_destroy(&parser);
>  	return rc;
> @@ -98,10 +105,24 @@ ck_constraint_program_compile(struct ck_constraint_def *ck_constraint_def,
>  	 * bind tuple fields there before execution.
>  	 */
>  	uint32_t field_count = space_def->field_count;
> +	/*
> +	 * Use column mask to prepare binding only for
> +	 * referenced fields.
> +	 */
>  	int bind_tuple_reg = sqlGetTempRange(&parser, field_count);
> -	for (uint32_t i = 0; i < field_count; i++) {
> +	struct bit_iterator it;
> +	bit_iterator_init(&it, &column_mask, sizeof(uint64_t), true);
> +	size_t used_fieldno = bit_iterator_next(&it);
> +	for (; used_fieldno < SIZE_MAX; used_fieldno = bit_iterator_next(&it)) {
>  		sqlVdbeAddOp2(v, OP_Variable, ++parser.nVar,
> -			      bind_tuple_reg + i);
> +			      bind_tuple_reg + used_fieldno);
> +	}
> +	if (column_mask_fieldno_is_set(column_mask, 63)) {

3. Please, do not hardcode that value. Add a one-line function to
column_mask.h for that. On other places too.

> +		used_fieldno = 64;
> +		for (; used_fieldno < (size_t)field_count; used_fieldno++) {
> +			sqlVdbeAddOp2(v, OP_Variable, ++parser.nVar,
> +				      bind_tuple_reg + used_fieldno);
> +		}
>  	}
>  	/* Generate ck constraint test code. */
>  	vdbe_emit_ck_constraint(&parser, expr, ck_constraint_def->name,

  reply	other threads:[~2019-05-19 16:02 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-05-16 13:56 [tarantool-patches] [PATCH v4 0/4] box: run checks on insertions in LUA spaces Kirill Shcherbatov
2019-05-16 13:56 ` [tarantool-patches] [PATCH v4 1/4] schema: add new system space for CHECK constraints Kirill Shcherbatov
2019-05-19 16:01   ` [tarantool-patches] " Vladislav Shpilevoy
2019-05-23 10:32     ` Kirill Shcherbatov
2019-05-26 12:03       ` Vladislav Shpilevoy
2019-05-31 13:45         ` Kirill Shcherbatov
2019-05-16 13:56 ` [tarantool-patches] [PATCH v4 2/4] box: run check constraint tests on space alter Kirill Shcherbatov
2019-05-19 16:02   ` [tarantool-patches] " Vladislav Shpilevoy
2019-05-23 10:37     ` Kirill Shcherbatov
2019-05-16 13:56 ` [tarantool-patches] [PATCH v4 3/4] box: introduce column_mask for ck constraint Kirill Shcherbatov
2019-05-19 16:02   ` Vladislav Shpilevoy [this message]
2019-05-23 10:38     ` [tarantool-patches] " Kirill Shcherbatov
2019-05-26 12:03     ` Vladislav Shpilevoy
2019-05-31 13:45       ` Kirill Shcherbatov
2019-05-16 13:56 ` [tarantool-patches] [PATCH v4 4/4] box: user-friendly interface to manage ck constraints Kirill Shcherbatov
2019-05-19 16:02   ` [tarantool-patches] " Vladislav Shpilevoy
2019-05-23 10:41     ` Kirill Shcherbatov
2019-05-26 12:04       ` Vladislav Shpilevoy
2019-05-31 13:45         ` Kirill Shcherbatov
2019-06-03 21:15           ` Vladislav Shpilevoy

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=148cfacb-757b-0c4a-7b7d-efaa18c6ffac@tarantool.org \
    --to=v.shpilevoy@tarantool.org \
    --cc=kshcherbatov@tarantool.org \
    --cc=tarantool-patches@freelists.org \
    --subject='[tarantool-patches] Re: [PATCH v4 3/4] box: introduce column_mask for ck constraint' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox