Commands out of sync

[lemartin]

Hello,

when running the following code I get this error.

PHP Fatal error:  Uncaught exception 'Doctrine\DBAL\Driver\Mysqli\MysqliException' with message 'Commands out of sync; you can't run this command now'

$stmt1 = $connection->prepare("SELECT * FROM information_schema.TABLES");
$stmt1->execute();
$stmt1->closeCursor();

$stmt1->execute();

$stmt2 = $connection->prepare("SELECT * FROM information_schema.VIEWS");
$stmt2->execute();

If $stmt1 is only executed once, the error does not occur. My guess is, that is is related to $this->_stmt->store_result(), which is only called the first time when MysqliStatement::execute is executed, because the second time $this->_columnNames is initialised (see. Doctrine\DBAL\Driver\Mysqli\MysqliStatement).

In commits 10a96cc and d00511c it was changed when store_result() is called. Before these commits, it was called when $this->_columnNames was initialised, now it is only called when $this->_columnNames is being initialised.

The main question is, was this intentionally and, is there a way to circumvent the problem?

No, closing the $stmt1 after the second execute is not really an option.

We are using doctrine/dbal 2.5.4 on PHP 5.4.45 with MySQL 5.5.

[deeky666]

I don't quite get what your example is trying to achieve? Why the second execute()? What is it supposed to do? AFAIK we do not support executing a prepared statement more than once. I am not even sure all PHP drivers support this.

[kimhemsoe]

PDO supports it seems http://php.net/manual/en/pdostatement.closecursor.php

I haven't looked if would be possible for mysqli driver to support it. How close do we follow PDO in these cases?

[deeky666]

@kimhemsoe we (should) support what it is defined in our interface contracts (as much as possible). Looks like we adopted the contract from PDO here, so basically it means we should support it where possible. I would suggest adding a failing test case for this is DBAL first and then start to fix the implementations.

[kimhemsoe]

👍

[morozov]

I would suggest adding a failing test case for this is DBAL first and then start to fix the implementations.

@deeky666 please see #2487.

[morozov]

Must have been fixed by #2546.

[github-actions]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.