WFS 1.1 timeStamp parameter

[assefa]

OCG cite tests requires to have the optional timeStamp parameter (plus the numerOfFeatures that we already generate). We could generate it using the current time.

[assefa]

committed in r9789.

Sorry for not giving enough time for discussing things but CITE test is such that sometime, not fixing an issue can prevent from going further in the test.

In this particular test, timeStamp (which is optional in the specs) is tested. If you guys see a need to control the output of the timeStamp using a metadata, please let me know and I will add it.

Note that I will update the msautotest at the end of the process.

[tomkralidis]

1./ This is an optional attribute. Why does CITE require this? 2./ How would we msautotest this, seeing that a current time would be generated

[assefa]

yes Tom you are right, we can't really test it in msautotest (I kind of missed that :)

I am not sure why CITE tests requires it but they do a xpath test to see if a numberOfFeatures and the timeStamp is there. If not it fails the test.

2 things here:

• either we control the output using something the wfs_generate_timestamp so it is controllable and we can set it for CITE and would be turned off by default
• or we just ignore it and fail the CITE test

I am more in favor of being able to control it but don't mind failing the test either since It is I beleive a CITE problem more than a MapServer problem.

[tomkralidis]

Thanks for the info. I would suggest bringing this up to CITE and ignore the failure. If timestamp is intended to be required, then this should become a change request (by whoever wants it to be required) into OGC to update OGC:WFS schemas.

Else, the CITE tests should be changed to _not_ look for /@timestamp.

[assefa]

Re-read again the specs, here is the xsd:

<xsd:element name="FeatureCollection?" type="wfs:FeatureCollectionType" substitutionGroup="gml:_FeatureCollection"/> <xsd:complexType name="FeatureCollectionType?"> <xsd:complexContent> <xsd:extension base="gml:AbstractFeatureCollectionType"> <xsd:attribute name="lockId" type="xsd:string" use="optional"/> <xsd:attribute name="timeStamp" type="xsd:dateTime" use="optional"/> <xsd:attribute name="numberOfFeatures" type="xsd:nonNegativeInteger" use="optional"/> </xsd:extension> </xsd:complexContent> </xsd:complexType>

But there is a clause to it in section 9.3 (document reference OGC 04-094 WF 1.1.0):

" ...If, however, the value of the resultType attribute is specified as hits, a web feature service must generate a <wfs:FeatureCollection> element with no content (i.e. empty) but must populate the values of the timeStamp attribute and the numberOfFeatures attribute. .."

Thus, we need to generate the timeStamp for resultType=hits. This would certainly prevent us from adding a test case for hits in msautotest.

Sorry for misleading you guys with incomplete info.

Should I make the correction to only generate the timeStamp for hits?

Thanks

[tomkralidis]

Replying to assefa:

Re-read again the specs, here is the xsd: <xsd:element name="FeatureCollection?" type="wfs:FeatureCollectionType" substitutionGroup="gml:_FeatureCollection"/> <xsd:complexType name="FeatureCollectionType?"> <xsd:complexContent> <xsd:extension base="gml:AbstractFeatureCollectionType"> <xsd:attribute name="lockId" type="xsd:string" use="optional"/> <xsd:attribute name="timeStamp" type="xsd:dateTime" use="optional"/> <xsd:attribute name="numberOfFeatures" type="xsd:nonNegativeInteger" use="optional"/> </xsd:extension> </xsd:complexContent> </xsd:complexType> But there is a clause to it in section 9.3 (document reference OGC 04-094 WF 1.1.0): " ...If, however, the value of the resultType attribute is specified as hits, a web feature service must generate a <wfs:FeatureCollection> element with no content (i.e. empty) but must populate the values of the timeStamp attribute and the numberOfFeatures attribute. .." Thus, we need to generate the timeStamp for resultType=hits. This would certainly prevent us from adding a test case for hits in msautotest. Sorry for misleading you guys with incomplete info. Should I make the correction to only generate the timeStamp for hits? Thanks

Thanks for the info. Yes you're right. Only generate timeStamp if resultType=hits.

[assefa]

ok thanks. committed in r9793

[assefa]

It was fixed.