| Bug #71679 | Avoid iterator creation when invoking statement interceptors in MysqlIO. | ||
|---|---|---|---|
| Submitted: | 12 Feb 2014 7:47 | Modified: | 10 Mar 2014 16:28 |
| Reporter: | Andrej Golovnin (OCA) | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | Connector / J | Severity: | S5 (Performance) |
| Version: | 5.1.29 | OS: | Any |
| Assigned to: | Filipe Silva | CPU Architecture: | Any |
| Tags: | interceptors, jdbc, MysqlIO | ||
[12 Feb 2014 7:48]
Andrej Golovnin
Patch for this issue.
Attachment: MysqlIO.patch (application/octet-stream, text), 1.78 KiB.
[12 Feb 2014 12:23]
Filipe Silva
Hi Andrej, Thank you for this bug report. Verified by code review. Thanks for noticing the duplicated line too.
[18 Feb 2014 13:48]
Andrej Golovnin
Patch for this issue. (*) I confirm the code being submitted is offered under the terms of the OCA, and that I am authorized to contribute it.
Contribution: MysqlIO.patch (application/octet-stream, text), 1.78 KiB.
[21 Feb 2014 10:13]
Alexander Soklakov
Thanks Andrej, accepted.
[10 Mar 2014 16:28]
Daniel So
Added the following entry into the Connector/J 5.1.30 changelog: "Avoided the use of an iterator over the list of statement interceptors in the methods MysqlIO.invokeStatementInterceptorsPost() and MysqlIO.invokeStatementInterceptorsPre(), so that Connector/J does not increase the stack size unnecessarily."
Description: In the methods MysqlIO#invokeStatementInterceptorsPost and MysqlIO#invokeStatementInterceptorsPre an iterator is used to loop over the list of statement interceptors. As far as I can tell, the list of the statement interceptors is always an ArrayList. So using the old-school for-loop should be the preferable way there to loop over the list. Even if the Hotspot would allocate the iterator on the thread's stack (and he does that), I think MySQL Connector/J should still avoid creation of the iterator, because the size of the thread stack is limited and I would prefer to see that objects from my application are allocated on the stack and not those from MySQL Connector/J. One more thing. I won't submit a patch for this. But could you please review the method ConnectionImpl#isSameResource(Connection), lines 4030-4031: if (!nullSafeCompare(otherOrigDatabase, this.origDatabaseToConnectTo)) { directCompare = false; directCompare = false; I think the assignment "directCompare = false;" in the line 4030 can be removed. :-) How to repeat: See my patch. Suggested fix: Use the old-school for-loop to iterate over the list.