Re: new version of contrib-intarray - Mailing list pgsql-hackers

From Tom Lane
Subject Re: new version of contrib-intarray
Date
Msg-id 24301.980779701@sss.pgh.pa.us
Whole thread Raw
In response to Re: new version of contrib-intarray  (Oleg Bartunov <oleg@sai.msu.su>)
List pgsql-hackers
Oleg Bartunov <oleg@sai.msu.su> writes:
>  gist__int_ops    |       1007
>  gist__intbig_ops |       1007

> we want gist__int_ops to be default index opclass.
> If we delete gist__intbig_ops entry from opclass, then we couldn't use
> gist__intbig_ops !

Put in gist__intbig_ops with zero for the default type.  You should
never have more than one entry in pg_opclass claiming to be the default
for a given type OID.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Ungraceful handling of fatal flex errors
Next
From: Jan Wieck
Date:
Subject: Security hole in PL/pgSQL