Debbie failing PostGIS trunk regress tickets on ST_GeneratePoints test on PostgreSQL 9.6

[robe]

I thought I complained about this issue already, but can't find the ticket.

Debbie is failing (seems only on 9.6 runs)

 swapordinates .. ok 
 summary .. ok 
 temporal .. ok 
 tickets .. failed (diff expected obtained: /var/lib/jenkins/workspace/postgis/tmp/2_3_pg9.6w64/test_87_diff)
-----------------------------------------------------------------------------
--- tickets_expected	2016-02-26 14:25:42.278086250 +0000
+++ /var/lib/jenkins/workspace/postgis/tmp/2_3_pg9.6w64/test_87_out	2016-03-11 02:26:21.953618580 +0000
@@ -280,8 +280,8 @@
 #3375|GEOMETRYCOLLECTION(POINT(0 -7))
 #3399|1|1
 #3399|10|10
-#3399|1000|1000
 #3399|100|100
+#3399|1000|1000
 ERROR:  invalid KML representation
 #3437a|5
 #3437b|5

Even though it shows as 2.3 only issue, I think that's because I'm testing 9.6 only on 2.3 (others seem to be passing fine). I had the same issue with this test on one run on my computer against 9.5/ 2.2

So I think the main issue is this test is not deterministic because it lacks an order by so since debbie always tests against 9.6 dev change - its possible something in 9.6 changed in planner to make this always fail.

[robe]

okay this is only 2.3 since it's ticket on ST_GeneratePoints

WITH g as (
select 'POLYGON((1 0, 0 1, 1 2, 2 1, 1 0))'::geometry as geom
),
n as (
select n from unnest(ARRAY[-1,0,1,10,100,1000]) n
),
pts as (
  select n,(st_dump(st_generatepoints(geom, n))).geom from g,n 
)
select '#3399' as t, n, count(*) from
g, pts
where st_contains(g.geom, pts.geom)
group by n;

[robe]

(In [14779]) Force order in ST_GeneratePoints so results are deterministic on all platforms Closes #3502