[Geoserver-devel] What's the status of wfs 1.1.0 cite tests?

Hi,
I've tried to run the wfs 1.1.0 cite test to check whether Martin
modification affected it in any way (trying to get him the get
go for the merge back on 2.4.x) and I was greeted by 25 failures
much of which I've never seen.

As quick summary of the failures:
* some tests apparently assume the geometric properties of
   primitiveGeoFeature are required? I have a test asking
   only for the date property and the validation of the
   result fails because they are not there?
* there is a test using "sf:measurand"... new feature type?
* failures for tests using urn:x-ogc:def:crs:FOO:2006.10:9999,
   and the test is actually expecting a failure (I think so
   at least) but the exception code is not the one expected
   (it should have been "InvalidParameterValue", it's
   "NoApplicableCode" instead).
* there are quite a bit of failures to to OOM... hummm...
   do we have a memory leak lurking there? I've run GeoServer
   with default JVM params, and for the standard cite test
   64MB are plenty of memory imho?
* there is one test searching for a feature named "cite.gmlsf0-f503"
   but there is no one in the dataset?

I've also found the axis flipping issue, but I tracked it
down in an error of mine: when flipping the data for postgis
I forgot to flip the coordinates of one feature, that's why
it was failing. The gml definition of that feauture is:

   <sf:featureMember>
     <sf:PrimitiveGeoFeature gml:id="f015">
       <gml:name codeSpace="http://cite.opengeospatial.org/gmlsf&quot;&gt;name\-f015&lt;/gml:name&gt;
       <sf:pointProperty>
         <gml:Point srsName="urn:x-ogc:def:crs:EPSG:6.11.2:4326">
           <gml:pos>34.94 -10.52</gml:pos>
         </gml:Point>
       </sf:pointProperty>
       <sf:intProperty>-900</sf:intProperty>
       <sf:measurand>-2.4</sf:measurand>
       <sf:decimalProperty>7.90</sf:decimalProperty>
     </sf:PrimitiveGeoFeature>
   </sf:featureMember>

So the insert statement for postgis should have been:

INSERT INTO "PrimitiveGeoFeature" VALUES (NULL, 'name-f015', NULL, geometryfromtext('POINT(-10.52 34.94)',4326), NULL, -900, NULL, 2.4, NULL, NULL, 7.90, 'f015');

instead I had the coordinates in the same order as the GML.

Curious, is there any way to retrieve the official wfs 1.1.0 test?
I've noticed a "2007-RC1" tag in the tests repo, but that does not
sound very promising (the official ones should be early 2007 and
final, not RC, no?). I have the impression someone pointed me
at some zips some time ago, but cannot find them anymore...

Cheers
Andrea

Hi Andrea,

I am not surprised there are failures. The engine from the trunk of the
ogc repo contains a lot more tests than the online engine at
cite.opengeospatial.org does.

I have successfully passed all tests with h2, i have not tried with
postgis yet. If you send me the logs i can probably help you recognize
the issues.

Andrea Aime wrote:

Hi,
I've tried to run the wfs 1.1.0 cite test to check whether Martin
modification affected it in any way (trying to get him the get
go for the merge back on 2.4.x) and I was greeted by 25 failures
much of which I've never seen.

As quick summary of the failures:
* some tests apparently assume the geometric properties of
   primitiveGeoFeature are required? I have a test asking
   only for the date property and the validation of the
   result fails because they are not there?

strange... they should not be since many of them are null. Is it a
validation error? sometimes they can be misleading.

* there is a test using "sf:measurand"... new feature type?

? not sure what you mean? measurand is a property of PrimitiveGeoFeature.

* failures for tests using urn:x-ogc:def:crs:FOO:2006.10:9999,
   and the test is actually expecting a failure (I think so
   at least) but the exception code is not the one expected
   (it should have been "InvalidParameterValue", it's
   "NoApplicableCode" instead).

Strainge... the same exception code should be reported regardless of the
backend.

* there are quite a bit of failures to to OOM... hummm...
   do we have a memory leak lurking there? I've run GeoServer
   with default JVM params, and for the standard cite test
   64MB are plenty of memory imho?

I have not run into any memory issues in any of my runs against h2..
could the memory leak be in postgis?

* there is one test searching for a feature named "cite.gmlsf0-f503"
   but there is no one in the dataset?

Ahhh... i think i know what might be happening and why you are getting
failures. They have added some data to the tests suite as well as tests.
So the script that loads the database needs to be updated.

I've also found the axis flipping issue, but I tracked it
down in an error of mine: when flipping the data for postgis
I forgot to flip the coordinates of one feature, that's why
it was failing. The gml definition of that feauture is:

   <sf:featureMember>
     <sf:PrimitiveGeoFeature gml:id="f015">
       <gml:name
codeSpace="http://cite.opengeospatial.org/gmlsf&quot;&gt;name\-f015&lt;/gml:name&gt;
       <sf:pointProperty>
         <gml:Point srsName="urn:x-ogc:def:crs:EPSG:6.11.2:4326">
           <gml:pos>34.94 -10.52</gml:pos>
         </gml:Point>
       </sf:pointProperty>
       <sf:intProperty>-900</sf:intProperty>
       <sf:measurand>-2.4</sf:measurand>
       <sf:decimalProperty>7.90</sf:decimalProperty>
     </sf:PrimitiveGeoFeature>
   </sf:featureMember>

So the insert statement for postgis should have been:

INSERT INTO "PrimitiveGeoFeature" VALUES (NULL, 'name-f015', NULL,
geometryfromtext('POINT(-10.52 34.94)',4326), NULL, -900, NULL, 2.4,
NULL, NULL, 7.90, 'f015');

instead I had the coordinates in the same order as the GML.

Curious, is there any way to retrieve the official wfs 1.1.0 test?
I've noticed a "2007-RC1" tag in the tests repo, but that does not
sound very promising (the official ones should be early 2007 and
final, not RC, no?). I have the impression someone pointed me
at some zips some time ago, but cannot find them anymore...

Not sure... i think the 2007-RC1 tag might be it... We should ask on the
cite list.

Cheers
Andrea

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

!DSPAM:4007,479f0c10284901431913854!

--
Justin Deoliveira
The Open Planning Project
http://topp.openplans.org