#5391 closed defect (fixed)
Offset curve regression and oriented envelop on debbie, winnie, and berries
Reported by: | robe | Owned by: | strk |
---|---|---|---|
Priority: | high | Milestone: | PostGIS 3.2.6 |
Component: | postgis | Version: | 3.3.x |
Keywords: | Cc: |
Description
I'm guessing offset curve behavior changed again.
Berries and Debbie all which are testing against geos branch are showing this regression
Offset curve
10:40:37 regress/core/offsetcurve .. failed (diff expected obtained: /var/lib/jenkins/workspace/postgis/tmp/3_4_pg15w64/test_92_diff) 10:40:37 ----------------------------------------------------------------------------- 10:40:37 --- ./regress/core/offsetcurve_expected 2022-08-22 20:53:14.411688691 +0000 10:40:37 +++ /var/lib/jenkins/workspace/postgis/tmp/3_4_pg15w64/test_92_out 2023-05-29 14:40:37.403659293 +0000 10:40:37 @@ -6,7 +6,7 @@ 10:40:37 t4|SRID=42;LINESTRING(0 10,10 10) 10:40:37 t5|SRID=42;LINESTRING(0 -10,10 -10,12 -10,14 -9,16 -8,17 -7,18 -6,19 -4,20 -2,20 0,20 10) 10:40:37 t5b|SRID=42;LINESTRING EMPTY 10:40:37 -t6|SRID=42;LINESTRING(0 -10,10 -10,17 -7,20 0,20 10) 10:40:37 +t6|SRID=42;LINESTRING(0 -10,10 -10,12 -10,14 -9,16 -8,17 -7,18 -6,19 -4,20 -2,20 0,20 10) 10:40:37 t7|SRID=42;LINESTRING(0 -10,10 -10,20 0,20 10) 10:40:37 t8|SRID=42;LINESTRING(0 -10,20 -10,20 10) 10:40:37 t10|SRID=42;LINESTRING(0 2,7 2,3 9) 10:40:37 ----------------------------------------------------------------------------- 10:40:37 regress/core/relatematch .. ok in 59 ms
oriented envelop
10:40:44 regress/core/oriented_envelope .. failed (diff expected obtained: /var/lib/jenkins/workspace/postgis/tmp/3_4_pg15w64/test_110_diff) 10:40:44 ----------------------------------------------------------------------------- 10:40:44 --- ./regress/core/oriented_envelope_expected 2022-08-22 20:53:14.411688691 +0000 10:40:44 +++ /var/lib/jenkins/workspace/postgis/tmp/3_4_pg15w64/test_110_out 2023-05-29 14:40:44.815634699 +0000 10:40:44 @@ -3,4 +3,4 @@ 10:40:44 t3|t 10:40:44 t4|t 10:40:44 t5|t 10:40:44 -t6|t 10:40:44 +t6|f
Change History (14)
comment:1 by , 18 months ago
Owner: | changed from | to
---|
comment:2 by , 18 months ago
Owner: | changed from | to
---|
comment:3 by , 18 months ago
comment:5 by , 18 months ago
The culprit for the change in ST_RotatedEnvelope must be this: https://github.com/libgeos/geos/commit/2efd5ed41009149db31b896cea98aacf765b004c
I've pushed a commit to show the obtained result, for a start
comment:6 by , 18 months ago
On my system the reduced precision used by the test refactoring works also with newer GEOS. I'll wait to see a green from debbie before closing this ticket
comment:7 by , 18 months ago
Also let me know if it's ok to push to the stable-3.3 branch now so that I can backport the fixes
comment:10 by , 18 months ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
Got a green: https://debbie.postgis.net/view/PostGIS/job/PostGIS_trunk/ Closing
comment:11 by , 17 months ago
Milestone: | PostGIS 3.4.0 → PostGIS 3.2.6 |
---|---|
Resolution: | fixed |
Status: | closed → reopened |
I think this needs to be backported to 3.2. Before i was testing winnie on 3.11.0, and flipped her to testing GEOS 3.11 branch, she is failing on offset curve
PostgreSQL 14.8, compiled by Visual C++ build 1914, 64-bit Postgis 3.2.6dev - (3.2.5-4-g59a8f9d63) - 2023-06-20 04:56:23 scripts 3.2.6dev 3.2.5-4-g59a8f9d63 raster scripts 3.2.6dev 3.2.5-4-g59a8f9d63 GEOS: 3.11.3dev-CAPI-1.17.3 PROJ: 7.2.1 SFCGAL: 1.4.1 GDAL: GDAL 3.4.3, released 2022/04/22 ./regress/core/offsetcurve .. failed (diff expected obtained: /projects/postgis/tmp/3.2.6dev_pg14_geos3.11_gdal3.4.3w64/test_89_diff) ----------------------------------------------------------------------------- --- ./regress/core/offsetcurve_expected 2023-06-19 18:45:49.789926800 -0400 +++ /projects/postgis/tmp/3.2.6dev_pg14_geos3.11_gdal3.4.3w64/test_89_out 2023-06-20 01:16:29.225477900 -0400 @@ -6,12 +6,12 @@ t4|SRID=42;LINESTRING(0 10,10 10) t5|SRID=42;LINESTRING(0 -10,10 -10,12 -10,14 -9,16 -8,17 -7,18 -6,19 -4,20 -2,20 0,20 10) t5b|SRID=42;LINESTRING EMPTY -t6|SRID=42;LINESTRING(0 -10,10 -10,17 -7,20 0,20 10) +t6|SRID=42;LINESTRING(0 -10,10 -10,12 -10,14 -9,16 -8,17 -7,18 -6,19 -4,20 -2,20 0,20 10) t7|SRID=42;LINESTRING(0 -10,10 -10,20 0,20 10) t8|SRID=42;LINESTRING(0 -10,20 -10,20 10) t10|SRID=42;LINESTRING(0 2,7 2,3 9) t10b|SRID=42;LINESTRING(0 2,7 2,3 9) t11|LINESTRING(37.6 39.2,39.2 36.6,42 35.8,43 34.8,46.4 33.6,48.2 30,48.8 30,50.6 33.8,56.6 35) t12|LINESTRING(34.4 36.8,36.8 33.4,40 32.2,41 31.2,43.6 30.4,45.8 26,51.2 26,53.4 30.2,57.4 31) -t13|LINESTRING(-2 0,-2 22,12 22,12 8,2 8) +t13|LINESTRING(-2 0,-2 10,-2 22,12 22,12 8,2 8) #2508|t
comment:12 by , 17 months ago
Summary: | Offset curve regression and oriented envelop on debbie and berries → Offset curve regression and oriented envelop on debbie, winnie, and berries |
---|
Commit [7be99c62a1dd9ee272a87d673960e84822a72a66/git] should fix the offsetcurve test, will look at oriented envelope now (had no idea it was implemented in terms of offsetcurve)