Commands out of sync; you can't run this command now about gomysql

[Philio]

This is strange, I've never had any issue like this from a simple select. Can you provide the table structures please?

[ziutek]

There is table structure with modified field names (security reasons):

CREATE TABLE `SomeTable` (
  `id` int(2) unsigned NOT NULL auto_increment,
  `s1` varchar(80) NOT NULL default '',
  `s2` varchar(50) NOT NULL default '',
  `s3` varchar(24) NOT NULL,
  `s4` varchar(20) NOT NULL default '',
  `s5` varchar(9) default NULL,
  `s6` varchar(5) default NULL,
  `e1` enum('T','N') NOT NULL default 'T',
  `s7` varchar(32) default NULL,
  `d1` date NOT NULL default '0000-00-00',
  `s8` varchar(80) default NULL,
  `s9` varchar(80) default NULL,
  `s10` varchar(16) NOT NULL default '',
  `d2` date default NULL,
  `d3` date default NULL,
  `d4` date default NULL,
  `d5` date default NULL,
  `d6` date default NULL,
  `e2` enum('T','N','K') NOT NULL default 'T',
  `d7` datetime NOT NULL default '0000-00-00 00:00:00',
  `s11` varchar(24) NOT NULL default '',
  PRIMARY KEY  (`id`)
) ENGINE=MyISAM AUTO_INCREMENT=2339 DEFAULT CHARSET=latin2

My MySQL version is: 5.0.51a-24+lenny4-log

I get no errors when select is limited to <300 entries.

[Philio]

Thanks, will check this out when I get a chance!

[ziutek]

I tried thoj's Go-MySQL-Client-Library. His quickly written code contains a lot of hacks, and poor error checking but after fix to Reader.Read usage it works without this issue (without my fix it doesn't work for long results).

[Philio]

This should be resolved in the latest release.

[thomaslee]

I spoke to ziutek briefly about this issue on the go-nuts mailing list. He had this to say about the issue:

"I realized during writing my package that this bug appears when text result row has NULL value in its first column. GoMySQL treat such Row Data Packet as OK packet and goes into the forest - Polish idiom ;). But this bug can be avoided by using prepared statements."

So this bug should manifest itself when the first column in a result set is NULL. Haven't attempted to reproduce this yet.

[thomaslee]

Will open a separate issue if this can be reproduced.

[Philio]

I've reopened :)

[ziutek]

I made a mistake in the sentence: "bug appears when text result row has NULL value in its first column". I meant an empty string, not a NULL value. NULL in first column gives 251 value in first byte after packet header, and it isn't problem. Empty string gives 0 value in first byte after header and it is treated by GoMySQL as mark of OK packet.

[Philio]

Ah right, yer that makes sense, as there is no check in place to make sure that the OK packet is the first packet in the set. Of course OK should only occur if it is the first packet!

[Philio]

Fixed, 0.2.5