Lars Heill wrote, On 12/ 2/09 05:56 PM: > Lars Heill wrote, On 12/ 1/09 03:03 PM: >> sunanda menon wrote, On 12/ 1/09 08:16 AM: >>> Darryl Gove wrote: >>>> sunanda menon wrote: >>>>> Pete Dennis wrote: >>>>>> So the change looks okay to me in that you are just reducing the >>>>>> optimization levels. >>>>>> >>>>>> I would like to see follow up bugs logged so that the reason why >>>>>> the optimization breaks things can be investigated - is it a MySQL >>>>>> coding bug or a compiler bug. >>>>>> >>>>> The mysql bug is logged at http://bugs.mysql.com/bug.php?id=49091. >>>>> >>>>> Is that what you are looking for or something else? >>>> >>>> That reports the problem with mysql, and suggests it's a compiler >>>> issue. Is it filed somewhere as a compiler issue? This is the first >>>> I've heard of it, and it's unlikely to get fixed if we don't know >>>> more about it. >>>> >>> The bug is verified by the mysql dev community and they will take it >>> further appropriately. >>> I do not handle that part,so can't speak much on that. >> >> There is communication with the Sun Studio compiler team now, yes. > > Compiler bug is 6906104. > > Workaround using "#pragma opt 2 (prev_record_reads)" being tried out.
This worked, but instead we got other test failures. And two more compiler bugs filed. The combined bug picture has led me to file a placeholder and tracking bug CR 6907393 for lowering the optimization level for MySQL 5.0 and 5.1 in OpenSolaris pending further investigations. -- Lars
