Hi, Igor! On Jan 21, Igor Babaev wrote:

revno: 3177 revision-id: igor@askmonty.org-20130121084657-hvcval7hykojxpmg parent: sergii@pisem.net-20130109225151-z9e7gh8z5nl38dxl committer: Igor Babaev <igor@askmonty.org> branch nick: maria-5.1-bug56927 timestamp: Mon 2013-01-21 00:46:57 -0800 message: Fixed bug mdev-4063 (bug #56927). This bug could result in returning 0 for the expressions of the form <aggregate_function>(distinct field) when the system variable max_heap_table_size was set to a small enough number. It happened because the method Unique::walk() did not support the case when more than one pass was needed to merge the trees of distinct values saved in an external file.

Looks ok. Just one thought. You use merge_many_buff() if the memory is not enough for the number_of_buffpeks*size. But merge_many_buff() only reduces the number of buffpeks down to 15. What if max_heap_table_size < 15*size ? I'd suggest to either 1) issue an error or 2) merge completely, like in Unique::get(). Actually, 2) should be easy, you just pass the last argument to your new Unique::merge() instead of TRUE, like in your patch, change to merge( ..., max_in_memory_size > MERGEBUFF2*size); Regards, Sergei