[tarantool-patches] Re: [PATCH 4/5] sql: display error on FK creation and drop failure

n.pettik korablev at tarantool.org
Wed Jul 25 13:03:54 MSK 2018


>> Before insertion to _fk_constraint we must be sure that there in no
>> entry with given <name, child id>. Otherwise, insertion will fail and
>> 'duplicate key' will be shown. Such error message doesn't seem to be
>> informative enough, so lets manually iterate through whole space looking
>> for appropriate record.
> 
> 1. As I know, vdbe_emit_halt_with_presence_test do not iterate through
> the whole space. It uses an index to search for the record fast.

You are right, I’ve fixed commit message.

> 
>> The same is for dropping constraint, but here vice versa: we test
>> that _fk_contraint contains entry with given name and child id.
> 
> 2. Typo: _fk_contraint -> _fk_constraint.

Fixed.

> 
>> It is worth mentioning that during CREATE TABLE processing schema id
>> changes and check in OP_OpenRead opcode fails (which in turn shows that
>> pointer to space may expire). On the other hand, _fk_constraint space
>> itself remains immutable, so as a temporary workaround lets use flag
>> indicating pointer to system space passed to OP_OpenRead. It makes
>> possible to use pointer to space, even if schema has changed.
>> Closes #3271
>> ---
>>  src/box/errcode.h            |  2 ++
>>  src/box/sql/build.c          | 43 +++++++++++++++++++++++++++++++------------
>>  src/box/sql/sqliteInt.h      | 10 +++++++---
>>  src/box/sql/trigger.c        | 24 ++++++++++++++++--------
>>  src/box/sql/vdbe.c           |  3 ++-
>>  test/box/misc.result         |  2 ++
>>  test/sql-tap/alter2.test.lua | 25 ++++++++++++++++++++++++-
>>  7 files changed, 84 insertions(+), 25 deletions(-)
>> diff --git a/src/box/sql/build.c b/src/box/sql/build.c
>> index c2d3cd035..20ace09e4 100644
>> --- a/src/box/sql/build.c
>> +++ b/src/box/sql/build.c
>> @@ -1784,6 +1784,20 @@ vdbe_fkey_code_creation(struct Parse *parse_context, const struct fkey_def *fk)
>>  		sqlite3VdbeAddOp2(vdbe, OP_Integer, fk->parent_id,
>>  				  constr_tuple_reg + 2);
>>  	}
>> +	/*
>> +	 * Lets check that constraint with this name hasn't
>> +	 * been created before.
>> +	 */
>> +	const char *error_msg =
>> +		tt_sprintf(tnt_errcode_desc(ER_CONSTRAINT_EXISTS), name_copy);
>> +	if (vdbe_emit_halt_with_presence_test(parse_context,
>> +					      BOX_FK_CONSTRAINT_ID, 0,
>> +					      constr_tuple_reg, 2,
>> +					      ER_CONSTRAINT_EXISTS, error_msg,
>> +					      false, OP_NoConflict) != 0) {
>> +		free((void *) name_copy);
> 
> 3. Name_copy is allocated on db.

Fixed:

+++ b/src/box/sql/build.c
@@ -1603,7 +1603,7 @@ vdbe_emit_fkey_create(struct Parse *parse_context, const struct fkey_def *fk)
                                              constr_tuple_reg, 2,
                                              ER_CONSTRAINT_EXISTS, error_msg,
                                              false, OP_NoConflict) != 0) {
-               free((void *) name_copy);
+               sqlite3DbFree(parse_context->db, (void *) name_copy);

>> +		return;
>> +	}
>> diff --git a/src/box/sql/vdbe.c b/src/box/sql/vdbe.c
>> index b9723e2e7..0f227e637 100644
>> --- a/src/box/sql/vdbe.c
>> +++ b/src/box/sql/vdbe.c
>> @@ -3172,7 +3172,8 @@ case OP_OpenWrite:
>>  	 * during runtime.
>>  	 */
>>  	if (box_schema_version() != p->schema_ver &&
>> -	    (pOp->p5 & OPFLAG_FRESH_PTR) == 0) {
>> +	    (pOp->p5 & OPFLAG_FRESH_PTR) == 0 &&
>> +	    (pOp->p5 & OPFLAG_SYSTEMSP) == 0) {
> 
> 4. Why not p5 & (FRESH_PTR | SYSTEMSP) ?

Because I am not used to work with bit flags..
This is definitely proper way to do it. Anyway, FRESH_PTR flag has gone,
so now it is again simple (pOp->p5 & OPFLAG_SYSTEMSP) == 0 check.





More information about the Tarantool-patches mailing list