Re: [HACKERS] What is "index returned tuples in wrong order" for recheck supposed to guard against?

2017-02-02 Thread Kyotaro HORIGUCHI
I forgot to mention this.. At Thu, 02 Feb 2017 21:11:16 +0900 (Tokyo Standard Time), Kyotaro HORIGUCHI wrote in <20170202.26.49572782.horiguchi.kyot...@lab.ntt.co.jp> > Hello, I digged this topic covered by a spiderweb.. > > # PostGIS requires so many

Re: [HACKERS] What is "index returned tuples in wrong order" for recheck supposed to guard against?

2017-02-02 Thread Kyotaro HORIGUCHI
Hello, I digged this topic covered by a spiderweb.. # PostGIS requires so many libraries installed :( FIY, for the given test case, the following query hits the bug. | SELECT edge.geom AS geom | FROM (SELECT * FROM knn_recheck_geom) AS edge | ORDER BY

Re: [HACKERS] What is "index returned tuples in wrong order" for recheck supposed to guard against?

2017-01-03 Thread Robert Haas
On Tue, Jan 3, 2017 at 12:36 AM, Regina Obe wrote: >> cmp would return 0 if the estimated distance returned by the index AM were >> greater than the actual distance. >> The estimated distance can be less than the actual distance, but it isn't >> allowed to be more. See

Re: [HACKERS] What is "index returned tuples in wrong order" for recheck supposed to guard against?

2017-01-02 Thread Regina Obe
>> If things are out of order, why isn't just going to was_exact = false >> good enough? >> >> I'm not sure if the mistake is in our PostGIS code or something in >> PostgreSQL recheck logic. >> If I change the elog(ERROR ...) to a elog(NOTICE, the answers are >> correct and sort order is

Re: [HACKERS] What is "index returned tuples in wrong order" for recheck supposed to guard against?

2017-01-02 Thread Robert Haas
On Fri, Dec 30, 2016 at 12:51 AM, Regina Obe wrote: > I've been trying to troubleshoot the cause of this PostGIS recheck bug we > have reported by two people so far. The last test was a nice simple > repeatable one that triggered the issue: > >

[HACKERS] What is "index returned tuples in wrong order" for recheck supposed to guard against?

2016-12-29 Thread Regina Obe
I've been trying to troubleshoot the cause of this PostGIS recheck bug we have reported by two people so far. The last test was a nice simple repeatable one that triggered the issue: https://trac.osgeo.org/postgis/ticket/3418 from what I have seen this only affects cases where we are doing a