Christopher Kings-Lynne wrote:
>>> need_paren = (PRETTY_PAREN(context) ?
>>> !IsA(op->rarg, RangeTblRef) : true);
>>
>>
>>
>> In a quick glance this code seems close to completely brain dead :-(
>> For one thing, why isn't it making separate determinations about whether
>> the left and right inputs of the UNION (resp INTERSECT or EXCEPT)
>> operator need to be parenthesized? After that maybe we could figure out
>> what the individual decisions need to be.
>
>
> So what are we going to do about it?
>
> Was it one of the pgAdmin guys who wrote it in the first place?
Yep, me. It was still on my radar to fix; not surprising, Tom was faster.
I'll have a look at the "braindead" issue.
Regards,