← Back to team overview

launchpad-dev team mailing list archive

Re: Stop the line: Test failures in *stable*

 

On Fri, Oct 8, 2010 at 11:00 AM, Michael Hudson
<michael.hudson@xxxxxxxxxxxxx> wrote:
> On Thu, 7 Oct 2010 21:06:30 -0500, Edwin Grubbs <edwin.grubbs@xxxxxxxxxxxxx> wrote:
>> On Thu, Oct 7, 2010 at 7:40 PM, Michael Hudson
>> <michael.hudson@xxxxxxxxxxxxx> wrote:
>> > On Thu, 7 Oct 2010 22:53:58 +0100, Graham Binns <graham@xxxxxxxxxxxxx> wrote:
>> >> On Thursday, October 7, 2010, Michael Hudson
>> >> <michael.hudson@xxxxxxxxxxxxx> wrote:
>> >> > Back in the SQLObject days we had a hack that would add "ORDER BY
>> >> > random()" to any query that didn't have an ORDER BY already.  Do we
>> >> > still have that?  Although in this case it seems we had an ORDER BY,
>> >> > just not a sufficienly discriminating one.  Could you add ", random()"
>> >> > to any query that does have an ORDER BY?
>> >> >
>> >>
>> >> Wouldn't that just break everything that relied on a specific ordering?
>> >
>> > Not if the ordering was already specific enough to be unambiguous, and
>> > if it's not specific enough to be unambiguous, it's a timebomb.
>> >
>> > Cheers,
>> > mwh
>>
>>
>> I think the confusion here is that Graham thinks you are suggesting replacing
>> "ORDER BY name" with "ORDER BY random()", but you are actually suggesting
>> that we append it, so "ORDER BY name" becomes "ORDER BY name, random()".
>
> Yes, exactly.  Sorry if I was confusing!

It seems to have disappeared. We still have the
randomise_select_results config item, and a descriptor listed in
sqlbase.__all__ without an implementation. I suspect it moved to Storm
core and was later removed from there?

-- 
Stuart Bishop <stuart@xxxxxxxxxxxxxxxx>
http://www.stuartbishop.net/



References