[mapserver-dev] Is WFS PROPERTYNAME overriding gml_include_items intentional?

Frank Warmerdam warmerdam at pobox.com
Sun Oct 10 15:38:32 EDT 2010

Yewondwossen Assefa wrote:
>  Hi Frank,
>  You are right, we should filter it through the 
> gml_include/exclude_items to make sure that protected fields are not 
> returned.
> I have added this and will correct it. [1]
> Another point related to this is that PROPERTYNAME is defined as being a 
> parameter that allows to enumerate " ..which of the non-mandatory 
> properties should be included in the response to a GetFeature 
> request...". This implies that some fields could be defined as mandatory 
> and some as optional. Although this might seem not that important for 
> most user, OGC Cite tests (for WFS 1.1) requires this support to pass a 
> series of test. I did note this requirement in a bug [2] and will 
> address it for 6.0. One idea would be to have the ability to define 
> which fileds are mandatory using item_[name]_mandatory = true and use 
> the propertyname parameter to augment the manadatory fileds with the 
> optional fields.  I believe if we process this at the WFS level and the 
> result is used to set gml_inlcude_items, then the different writers will 
> not be affected.


Sounds good.  I will note I have adapted my code to use msGMLGetItems()
rather than try to interprete the various gml related metadata myself so
as long as changes in permissions, etc end up effecting what
msGMLGetItems() returns then I'm good on my side too.

This is not true of WFS responses generated with the template mechanism
where the fields exposed ignore all these considerations and are
driven strictly by the template itself.   I shall attempt to clarify this
distinction in the RFC.

Best regards,
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent

More information about the mapserver-dev mailing list