Matthew Gabeler-Lee wrote:
> Okay, now I understand the problem better.
>
> How about, as an intermediate solution, a list of 'sane' locales in which
> the optimization applied to the C/POSIX locale works? It should be fairly
> straight forward to write a test program that will test a locale and see if
> any of the common problems apply. And I expect that several locales can be
> added straight away as they are known to obey simple sorting rules.
>
> Alternately, though I don't think this would be a very good idea, a SET
> option along the lines of enable_dangerous_like_optimization or such that
> would (of course) default to off.
I think we do know the C local is OK for this. It is just the other
ones we don't know about, and we don't even know. Is there some test we
can write for this?
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073