On 2015/03/20 21:31, Bruce Momjian wrote:
> On Thu, Oct 23, 2014 at 06:29:07PM +0900, Etsuro Fujita wrote:
>> I don't think that the lock level mentioned in the following comment in
>> MergeAttributes() in tablecmds.c is right, since that that function has
>> opened the relation with ShareUpdateExclusiveLock, not with
>> AccessShareLock. Patch attached.
>>
>> 1749 /*
>> 1750 * Close the parent rel, but keep our AccessShareLock on it
>> until xact
>> 1751 * commit. That will prevent someone else from deleting or
>> ALTERing
>> 1752 * the parent before the child is committed.
>> 1753 */
>> 1754 heap_close(relation, NoLock);
>
> Agreed, patch applied. Thanks.
Thanks for picking this up!
Best regards,
Etsuro Fujita