[tarantool-patches] Re: [PATCH v1 1/1] sql: rework five syntax errors

["n.pettik" <korablev@tarantool.org>]

> On 21 Feb 2019, at 18:40, imeevma@tarantool.org wrote:
> 
> This patch reworks five syntax errors.

Why only five? There are 10 errors at all, so lets finish this part
adding all syntax errors.

> These SQL errors were
> converted into Tarantool errors.
> 
> Part of #3965
> ---
> https://github.com/tarantool/tarantool/issues/3965
> https://github.com/tarantool/tarantool/tree/imeevma/gh-3965-use-diag_set-to-describe-error
> 
> diff --git a/src/box/errcode.h b/src/box/errcode.h
> index 1523b19..1807151 100644
> --- a/src/box/errcode.h
> +++ b/src/box/errcode.h
> @@ -230,6 +230,11 @@ struct errcode_record {
> 	/*175 */_(ER_SQL_NO_SUCH_PRAGMA,	"Pragma '%s' does not exist") \
> 	/*176 */_(ER_SQL_CANT_RESOLVE_FIELD,	"Can’t resolve field '%s'") \
> 	/*177 */_(ER_INDEX_EXISTS_IN_SPACE,	"Index '%s' already exists in space '%s'") \
> +	/*178 */_(ER_SQL_SYNTAX,		"Syntax error in %s: %s") \
> +	/*179 */_(ER_SQL_STACK_OVERFLOW,	"Failed to parse SQL statement: parser stack limit reached") \
> +	/*180 */_(ER_SQL_SELECT_WILDCARD,	"Failed to expand '*' in SELECT statement without FROM clause") \
> +	/*181 */_(ER_SQL_STATEMENT_EMPTY,	"Failed to execute an empty SQL statement") \

I’ve already said that: why do we consider empty query as an error? Don’t understand.

The rest is ok as obvious.