Bug #40120 Possible infinite loop in remove_dup_with_compare()
Submitted: 17 Oct 2008 18:35 Modified: 10 Feb 2018 16:52
Reporter: Ingo Strüwing Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Server: General Severity:S3 (Non-critical)
Version:4.1, 5.0, 5.1, 6.0 OS:Any
Assigned to: CPU Architecture:Any

[17 Oct 2008 18:35] Ingo Strüwing 
Description:
This code snippet from sql/sql_select.cc:remove_dup_with_compare() shows that we could possibly run into an infinite loop:

  file->ha_rnd_init(1);
  error=file->rnd_next(record);
  for (;;)
  {
    if (thd->killed)
    {
      thd->send_kill_message();
      error=0;
      goto err;
    }
    if (error)
    {
      if (error == HA_ERR_RECORD_DELETED)
	continue;

If the first record is a deleted one...

However, this may be unused code. It seems like duplicate removal is done by unique indexes nowadays.

How to repeat:
I was not able to create a test case to exploit this problem.
But it is quite obvious from code reading.

Suggested fix:
Either read the next record before continuing,
or eliminate the unused code altogether.
[21 Oct 2008 6:08] Sveta Smirnova 
Thank you for the report.

Verified as described.
[10 Feb 2018 16:52] Roy Lyseng 
Posted by developer:
 
Fixed in version 5.5 and up.