Fix fetching last insert ID for sequences on SQL Server 8000 #2634
Conversation
| return parent::lastInsertId($name); | ||
| } | ||
|
|
||
| $sql = 'SELECT CONVERT(VARCHAR(MAX), current_value) FROM sys.sequences WHERE name = ?'; |
There was a problem hiding this comment.
Redundant? You mean I should inject it into prepare() directly?
| @@ -40,6 +40,22 @@ public function __construct($dsn, $user = null, $password = null, array $options | |||
| /** | |||
| * {@inheritDoc} | |||
| */ | |||
| public function lastInsertId($name = null) | |||
| { | |||
| if (null === $name) { | |||
There was a problem hiding this comment.
I don't think so. What should we expect in this case?
There was a problem hiding this comment.
That's why I asked. Probably needs a new issue though, where we validate that the passed in value is not empty or such.
There was a problem hiding this comment.
@Ocramius the problem is the interface does not specify such case. Neither does PDO's interface. No exception, not non-string return value specified.
There was a problem hiding this comment.
Sure, not happy with the situation either but that's what it is right now :(
| $stmt = $this->prepare($sql); | ||
| $stmt->execute(array($name)); | ||
|
|
||
| return $stmt->fetchColumn(); |
There was a problem hiding this comment.
Can current_value ever be NULL?
There was a problem hiding this comment.
Given to the documentation no.
3e79791 to
5533854
Compare
|
Btw the Oracle test failure is obviously unrelated. That test in question seems fragile. Think I have encountered that one on Oracle sometime before. Need to have a look at that. |
|
Test failure fixed in #2635 |
…fix' into 2.5 Backport #2634 into 2.5.*
|
Backported to |
Fixes fetching last insert ID for sequences on both
sqlsrvandpdo_sqlsrvdrivers. The implementation onsqlsrvwas wrong andpdo_sqlsrvdoes not support this out of the box.The issue was revealed in #2617. See: