Something is really messed up with the gserialized or my install of it

[robe]

Paul,

Can you look at these and confirm you have the same issues. I'm not sure if I somehow mixed up my installs or these are true issues. I did these tests before moving forward with the speed compares you discussed. I then stopped when I started seeing these bizarre differences between the 2.

To compile — I first enabled gserialized in gserialize.h then did my usual install on a parallel postgresql 9.0 windows 32 bit.

Then I repeated the same exercise with default gserialized disabled.

1) Looks like we are missing all those over left, over right operators Mark whined about. Is that intentional (since my raster install complains when installing it from the gserialized batch of scripts (postgis.sql, rtpostgis.sql)).

2) ST_IsClosed crashes with LINE Empty (again works or at least doesn't crash with gserialized disabled.

—- gserialized ST_IsClosed — SELECT ST_IsClosed(ST_GeomFromText('LINESTRING EMPTY',4326));

—server crashes—

3) This one I wanted to simplify to make it easier to inspect. I can't figoure what is wrong with this. I tried to dissect it, but I can't rebuild the geometry from the ST_AsEWKT of the outputs — complains about polygons not well formed (the error you get when you have non-closed polygons) or something

— create table

SELECT ST_Collect(geom) As geom
INTO test_collgeomm
FROM
(VALUES ( ST_GeomFromEWKT('SRID=4326;MULTIPOLYGONM(((-71.0821 42.3036 2,-71.0822 42.3036 3,-71.082 42.3038 2,-71.0819 42.3037 2,-71.0821 42.3036 2)))') ),
	( ST_GeomFromEWKT('SRID=4326;POLYGONM((-71.1261 42.2703 1,-71.1257 42.2703 1,-71.1257 42.2701 2,-71.126 42.2701 1,-71.1261 42.2702 1,-71.1261 42.2703 1))') )
		)	As g(geom) CROSS JOIN generate_series(1,3) As i
		GROUP BY i;

		-- run query
SELECT ST_DFullyWithin(f1.geom, f2.geom,100)
FROM test_collgeomm As f1 CROSS JOIN test_collgeomm As f2;

— get error — ERROR: Return size (160) not equal to expected size (152)!

[pramsey]

If you're missing the operators then something is wrong in your build/install/setup, because they are most definitely there.

The isclosed issue is a real crasher, I will work on that.

And the third one is also a problem, worthy and worrisome.

[pramsey]

IsClosed crasher is gone at r6542

[pramsey]

And the third one is actually a manifestation of a bug in convex hull

select st_convexhull(ST_GeomFromEWKT('SRID=4326;POLYGONM((-71.1261 42.2703 1,-71.1257 42.2703 1,-71.1257 42.2701 2,-71.126 42.2701 1,-71.1261 42.2702 1,-71.1261 42.2703 1))'));

which I have reported at #741. I'm marking this ticket closed now while you debug your install! (Down with issue splurging!)