Re: [BUGS] incorrect index behaviour with rtree on box values

2005-02-14 Thread Bruce Momjian
This has been saved for the 8.1 release: http://momjian.postgresql.org/cgi-bin/pgpatches2 --- Andrew - Supernews wrote: > On 2005-01-25, Tom Lane <[EMAIL PROTECTED]> wrote: > > Andrew - Supernews <[EMAIL PROTECTED]>

Re: [BUGS] incorrect index behaviour with rtree on box values

2005-01-26 Thread Andrew - Supernews
On 2005-01-25, Tom Lane <[EMAIL PROTECTED]> wrote: > Andrew - Supernews <[EMAIL PROTECTED]> writes: >> The problem is that the semantics of the &< and &> operators for the box >> type are not what rtree needs for the "OverLeft" and "OverRight" slots of >> the operator class. > > This was observed n

Re: [BUGS] incorrect index behaviour with rtree on box values

2005-01-24 Thread Tom Lane
Andrew - Supernews <[EMAIL PROTECTED]> writes: > The problem is that the semantics of the &< and &> operators for the box > type are not what rtree needs for the "OverLeft" and "OverRight" slots of > the operator class. This was observed nearly a year ago, see this thread: http://archives.postgres

[BUGS] incorrect index behaviour with rtree on box values

2005-01-24 Thread Andrew - Supernews
Testcase: create table boxtest (a box); create index boxtest_idx on boxtest using rtree (a); create function gen_data() returns void as ' begin for i in 1..200 loop insert into boxtest values (box(point((i*2-1)::float,0),point((i*2)::float,1))); end loop; return; end;' language pl