[postgis-users] Inconsistent behavior with ST_Within across versions

[Imre Samu]

> I’ve updated my Trac ticket accordingly.

some other testing ideas *for the next week :*

1.) As I see the latest PostgreSQL patch release is available ( 2022-02-10 )
   and some corrupted index problem has been fixed in the 14.2:
https://www.postgresql.org/docs/release/14.2/

*"However, some bugs have been found that may have resulted in corrupted
indexes, as explained in the first two changelog entries. *
* If any of those cases apply to you, it's recommended to reindex
possibly-affected indexes after updating."*

   So, in theory, you should re-check your problem *the next week *- with
the newer "postgis/postgis" images  ( based on 14.2; 13.6; 12.10; 11.15 )

2.) You can add some Debian-based docker images to your test cases
    so we can check: it is a "musl" related problem or not.

3.) You can also test the  "14-master"  image ( near ~master postgis, geos,
gdal  .. )

https://github.com/postgis/docker-postgis/blob/master/14-master/Dockerfile#L183_L204

Kind Regards,
 Imre







Carsten Byrman <byrman+postgis at posteo.nl> ezt írta (időpont: 2022. febr.
11., P, 15:53):

> Thanks for your recommendations, Imre. I’ve updated my Trac ticket
> accordingly.
>
> Kind regards, Carsten
>
>
> On 11 Feb 2022, at 13:46, Imre Samu <pella.samu at gmail.com> wrote:
>
> Hi Carsten,
>
> > but am just not sure I chose the appropriate milestone/component/version
> > Could you please give me some guidance here?
>
> ( I am a docker-postgis contributor)
>
> I am so sorry,  but not "all" docker postgis/postgis image updating
> regularly ( with minor postgres,postgis) patches, bug fixes
> Only as in this repo:  https://github.com/postgis/docker-postgis
> (scheduled on every Monday )
>
> You can check the "age / created " of the local docker image
> $ docker images postgis/postgis
> postgis/postgis 11-2.5-alpine 10c7c19a572c 4 days ago 330MB
> postgis/postgis 11-3.2-alpine 0303403af49f 4 days ago 369MB
> postgis/postgis 11-3.1-alpine f8c3965c17b8 7 weeks ago 328MB <-- too old
> postgis/postgis 11-3.0-alpine 4605eb1b2223 14 months ago 289MB <-- extreme
> old
>
> So I am not recommending the current  -3.0; -3.1 docker images  ( or any
> images older than 1 month, like "12-2.5-alpine" )
> This is a technical limitation of the current docker image workflow. (
> probably it can be improved in the future )
>
> You can check the actual version with:
> SELECT version();
> SELECT PostGIS_Full_Version();
>
>
> The other important issue for debugging the alpine version ( 3.12 - 3.15 )
> and you can  check with "cat /etc/os-release"
>
> $ docker run -t --rm postgis/postgis:11-3.0-alpine sh -c "cat
> /etc/os-release | grep VERSION"
> VERSION_ID=3.12.2
> $ docker run -t --rm postgis/postgis:11-3.2-alpine sh -c "cat
> /etc/os-release | grep VERSION"
> VERSION_ID=3.15.0
>
>
> So my recommendations:
> -  please remove the "older docker" images from your test cases
> -  and add postgres / postgis / alpine versions
>
> It is an interesting issue;
>
> Regards,
>   Imre
>
>
> Carsten Byrman <byrman+postgis at posteo.nl> ezt írta (időpont: 2022. febr.
> 11., P, 7:41):
>
>> Hello,
>>
>> I noticed inconsistent behavior with ST_Within across different
>> PostgreSQL-PostGIS versions. Most notably, the outcome (not performance) of
>> my query is determined by the presence of an index, which strikes me as
>> odd. I filed my first bug report about 3 months ago and I expect no
>> solution any time soon, but am just not sure I chose the appropriate
>> milestone/component/version: https://trac.osgeo.org/postgis/ticket/5025.
>> Perhaps it’s not even a PostGIS but a PostgreSQL issue. Could you please
>> give me some guidance here?
>>
>> Thanks, Carsten
>> _______________________________________________
>> postgis-users mailing list
>> postgis-users at lists.osgeo.org
>> https://lists.osgeo.org/mailman/listinfo/postgis-users
>>
> _______________________________________________
> postgis-users mailing list
> postgis-users at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/postgis-users
>
>
> _______________________________________________
> postgis-users mailing list
> postgis-users at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/postgis-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20220211/d75467f0/attachment.html>