[CITE-Forum] Solved (was: Re: Error in WMS 1.3 r3 test)
arnulf.christl at metaspatial.net
Sun Mar 31 10:36:24 EDT 2013
-----BEGIN PGP SIGNED MESSAGE-----
On 03/29/2013 06:56 PM, Andreas Schmitz wrote:
> Arnulf Christl wrote:
> Hi Arnulf,
>> I might be wrong but apparently in some queries the test engine
>> asks for a BBOX with zero extent. This produces an error on the
>> server side (in my opinion correctly) which is interpreted by the
>> test engine as a failure of the test in question.
>> URL generated by the test engine:
When the bbox is corrected (by adding 0.00001 deg to the ur
>> coordinate) it works fine:
> I'm unsure how that happens, but deegree for example always sends
> a service exception in case minx == maxx or miny == maxy, and we
> pass the tests. However, your service advertises the mentioned
> request box as the layer's bounding box, the scripts probably just
> try to request the layer with its advertised bounding box.
> Best regards, Andreas Schmitz
thanks for the quick answer. This is actually an interesting snag for
testing. As you correctly suggest the problem arises because of the
zero extent advertized in the capabilities of this layer - which is
used by the test engine without checking for viability (a dumb client
so to say).
If not manually specified MapServer will dynamically compute the BBOX
by querying the data in question. For the Bridge layer with a single
point the BBOX is (correctly) computed to have identical coordinates
for ll and ur - a zero extent. Which is probably correct even if not
very useful. So both MapServer and the test engine do the correct
thing but end up with a failure anyway.
My workaround is now to manually configure a BBOX which is has a width
and height larger zero. MapServer 6.2 passes the test without error now.
Thanks for the hint.
To avoid this from happening to other people we could simply add
another bridge to this CITE test data layer.
> _______________________________________________ CITE-Forum mailing
> list CITE-Forum at lists.opengeospatial.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with undefined - http://www.enigmail.net/
-----END PGP SIGNATURE-----
More information about the CITE-Forum