#5271 closed defect (duplicate)

garden test crash on ST_SnapToGrid

Reported by: robe Owned by: robe
Priority: blocker Milestone: PostGIS 3.1.8
Component: QA/buildbots Version: master
Keywords: Cc:

Description

I thought this might be the geos, but don't recall GEOS crashing on ST_SnapToGrid

PostgreSQL 15.0 on armv7l-unknown-linux-gnueabihf, compiled by gcc (Raspbian 10.2.1-6+rpi1) 10.2.1 20210110, 32-bit
22:21:46   Postgis 3.4.0dev - (0415ccb) - 2022-10-27 02:10:28
22:21:46   scripts 3.4.0dev 0415ccb
22:21:46   raster scripts 3.4.0dev 0415ccb
22:21:46   GEOS: 3.11.1dev-CAPI-1.17.1
22:21:46   PROJ: 7.2.1
22:21:46   GDAL: GDAL 3.2.2, released 2021/03/05

berrie is crashing on this garden test, which of course happened when I put back all the tests

SELECT ST_SnapToGrid(foo1.the_geom, 20.1, 20.1, 20.1, 20.1)  As result
                                                        FROM (
                (SELECT ST_GeomFromText('MULTIPOLYGON (((9 9, 9 1, 1 1, 2 4, 7 7, 9 9)), EMPTY)', 4326) As the_geom ) ) As foo1 LIMIT 10;

I just confirmed my dev is crashes too on this test

with my geos 11 (dated in June). Could be both our GEOS are old.

I'll switch to her geos 3.12 which should be much newer to see if it resolves the issue.

Change History (3)

comment:1 by Regina Obe <lr@…>, 21 months ago

In 030dfee/git:

Switch berrie to GEOS 3.12.0. References #5271

comment:2 by robe, 21 months ago

Milestone: PostGIS GEOSPostGIS 3.1.8
Summary: berrie garden test crash on ST_SnapToGridgarden test crash on ST_SnapToGrid

I just upgraded my dev to latest GEOS 3.11 (dated yesterday from winnie) and I still get a crash, though I didn't recompile.

At a glance hmm maybe this is a PostGIS crash. SnaptoGrid doesn't look like it uses geos.

comment:3 by robe, 21 months ago

Resolution: duplicate
Status: newclosed

Oops this is a duplicate of #5241.

I had it remarked out of the garden tests, and I guess I didn't fix this one yet and thought I did.

Note: See TracTickets for help on using tickets.