Concave Hull Hard failing on debbie and winnie's 32bit

[robe]

This might be working fine on trunk because we took out the concave_hull_hard tests.

But both debbie and winnie seem to be failing.

I wouldn't be so alarmed if it was just debbie, as she's still suffering from memory issues I'm trying to troubleshoot.

What concerns me is it's winnie's 32-bit tests failing and I think she's the only one that tests 32-bit

On winnie getting:

 concave_hull_hard .. failed (diff expected obtained: /projects/postgis/tmp/2.5.3dev_pg10_geos3.7.1_gdal2.2.4w32/test_35_diff)
-----------------------------------------------------------------------------
--- concave_hull_hard_expected	2018-10-02 01:02:30.672320700 -0400
+++ /projects/postgis/tmp/2.5.3dev_pg10_geos3.7.1_gdal2.2.4w32/test_35_out	2019-03-29 14:41:32.506425100 -0400
@@ -2,4 +2,3 @@
 ST_ConcaveHull Lines 0.80|t|t
 ST_ConcaveHull Lines 0.80 holes|t|t
 #3638|t|t
-#3697|t|t

debbie just hits the roof on the concavehull_hard

 concave_hull_hard .. failed (diff expected obtained: /var/lib/jenkins/workspace/postgis/tmp/2_5_pg9.6w64/test_35_diff)
-----------------------------------------------------------------------------
--- concave_hull_hard_expected	2019-04-08 06:26:53.706696460 +0000
+++ /var/lib/jenkins/workspace/postgis/tmp/2_5_pg9.6w64/test_35_out	2019-04-08 14:35:01.904674217 +0000
@@ -2,4 +2,8 @@
 ST_ConcaveHull Lines 0.80|t|t
 ST_ConcaveHull Lines 0.80 holes|t|t
 #3638|t|t
-#3697|t|t
+server closed the connection unexpectedly
+	This probably means the server terminated abnormally
+	before or while processing the request.
+no connection to the server
+connection to server was lost
-----------------------------------------------------------------------------
 ctors .. failed (diff expected obtained: /var/lib/jenkins/workspace/postgis/tmp/2_5_pg9.6w64/test_36_diff)
-----------------------------------------------------------------------------
--- ctors_expected	2019-04-08 06:26:53.706696460 +0000
+++ /var/lib/jenkins/workspace/postgis/tmp/2_5_pg9.6w64/test_36_out	2019-04-08 14:35:01.964674245 +0000
@@ -1,11 +1 @@
-32749
-ERROR:  Operation on mixed SRID geometries
-SRID=3;LINESTRING(0 0,1 1)
-ERROR:  Operation on mixed SRID geometries
-ST_MakeLine1|LINESTRING(0 0,1 1,10 0)
-ST_MakeLine_agg1|LINESTRING(0 0,1 1,10 0,20 20,40 4,40 4,40 5,40 5,40 6,40 6,40 7,40 8)
-ST_MakeLine_agg2|LINESTRING(0 0,1 0,1 0)
-BOX(0 0,1 1)
-ERROR:  Operation on mixed SRID geometries
-BOX3D(0 0 0,1 1 0)
-ERROR:  Operation on mixed SRID geometries
+psql: FATAL:  the database system is in recovery mode

and has trouble recovering. I had to revert debbie back to a 3/11 backup to try to debug her memory issues, so don't know when that started with her (sometime after 3/11).

On winnie, issue seems to have started after r17367

Commit 5de526daad5fa0a9c8c86814f5303f8a45de6d55 by rmrodriguez

Fix postgis_type_name with (GEOMETRYM,3)
Patch by Matt Bretl
References #4361
 git-svn-id: http://svn.osgeo.org/postgis/branches/2.5@17367
b70326c6-7e19-0410-871a-916f4a2858ee

[robe]

well second run of winnie's just reran not giving regress error anymore on concave_hard. So not sure what caused that discrepancy. debbie is still having issues with anything involving lots of memory so that might be nothing.