|From:||Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>|
|To:||Jeff Davis <pgsql(at)j-davis(dot)com>|
|Subject:||Re: WIP: generalized index constraints|
|Views:||Raw Message | Whole Thread | Download mbox | Resend email|
Jeff Davis <pgsql(at)j-davis(dot)com> writes:
> On Sat, 2009-09-19 at 16:43 -0400, Tom Lane wrote:
>> I don't understand why this isn't handled exactly the way unique
>> constraints are done now. Frankly, the amount of added complexity you
>> propose below is enough to make me want to reject the patch forthwith;
>> given that it's going to be a relatively little-used feature, the bugs
>> are never going to be out of it completely if we do it like this.
> Unique constraints lock the index page while the insert is happening.
> How am I supposed to do that, when the conflicting values might be
> anywhere in the index (circles have no total order)?
Well, you can't do it *exactly* the same way btree does, but what
I would envision is first insert the index tuple and then do a
dirty-snapshot search for conflicting tuples. The interlock against
conflicting concurrent inserts doesn't need all this new infrastructure
you propose; just wait to see if conflicting transactions commit, same
as we do now. And I do maintain that that sort of code has a high risk
of undetected bugs.
regards, tom lane
|Next Message||Tom Lane||2009-09-19 22:35:45||Re: operator exclusion constraints [was: generalized index constraints]|
|Previous Message||Tom Lane||2009-09-19 21:54:43||Re: generic copy options|