Re: unicode upper/lower functions

Bruce Momjian wrote:
> John Hansen wrote:
> > ... Except,.. It was never decided if the 'C' locale
> optimisations was
> > going to be removed if/when implementing ICU.
>
> Uh, why would we remove it?  Oh, meaning if the locale is C
> we bypass locale lookups?  I think we will have to see what
> performance we have with things.

Uhh, not quite: If locale is 'C' the current assumption is 7-bit ASCII
for upper/lower/initcap.
ICU is capable of properky doing upper/lower/initcap except for the
cases described in this (obsolete) patch.

>
> > Tho I think the  conclusion was a postgresql.conf parameter to
> > enable/disable the optimisations.
> > Either way, this code is now obsolete.
>
> Thanks.
>
> --
>   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
>
>