Re: [Geoserver-devel] [Geotools-devel] Registered function support for JDBC databases

2008-07-22 Thread Rob.Atkinson
OK - I suppose thats a valid sanity check - not enough info to help formulate the request, since you dont even know what attributes of what feature types it applies to, but you can reject some dumb ones. Anyway, its such a trivial thing to advertise them we can do it now. I was just sceptical a

Re: [Geoserver-devel] [Geotools-devel] Registered function support for JDBC databases

2008-07-22 Thread Rob.Atkinson
> My initial feedback was more about how to advertise what functions are > available - allowing only these functions is possibly the higher priority. > so geoserver can report them as available etc. >There is one thing to check in the WFS specification; I am under the impression that server n

Re: [Geoserver-devel] [Geotools-devel] Registered function support for JDBC databases

2008-07-21 Thread Ben Caradoc-Davies
Jody Garnett wrote: > A couple things; executing any sql function is not a good thing from a > security standpoint (but you know this). Indeed. > You should be able to advertise > additional functions on a data store by datastore basis using the filter > capabilities data structure. You are th

Re: [Geoserver-devel] [Geotools-devel] Registered function support for JDBC databases

2008-07-21 Thread Jody Garnett
Rob Atkinson wrote: > hmm - IMHO not having to get a WFS change request is starting small. > Just accepting that we can get the functionality working and work out > how to advertise it later (i.e. there isnt really much point > advertising it if its feature specific) seems to me a much lower bar

Re: [Geoserver-devel] [Geotools-devel] Registered function support for JDBC databases

2008-07-21 Thread Rob Atkinson
hmm - IMHO not having to get a WFS change request is starting small. Just accepting that we can get the functionality working and work out how to advertise it later (i.e. there isnt really much point advertising it if its feature specific) seems to me a much lower bar. Getting the big picture righ

Re: [Geoserver-devel] [Geotools-devel] Registered function support for JDBC databases

2008-07-21 Thread Jody Garnett
[EMAIL PROTECTED] wrote: > Ahh! Herein lies a quandry. WFS doesnt allow per feature functions to be > advertised, and most interesting features have implicit or explicit > operations that would be nice to advertise per feature as functions. > Ideally, those functions are only allowed on those f

Re: [Geoserver-devel] [Geotools-devel] Registered function support for JDBC databases

2008-07-21 Thread Jody Garnett
Rob Atkinson wrote: > There is nothing intrinsically postgis about the solution, we envisage > it a JDBC capability. I agree :-) > A parameter with a sensible default could advertise the table or > function that would be interrogated to find functions intended to be > advertised. > > My could a

Re: [Geoserver-devel] [Geotools-devel] Registered function support for JDBC databases

2008-07-21 Thread Rob Atkinson
There is nothing intrinsically postgis about the solution, we envisage it a JDBC capability. A parameter with a sensible default could advertise the table or function that would be interrogated to find functions intended to be advertised. My could attempt to call a function getRegisteredFunctions

Re: [Geoserver-devel] [Geotools-devel] Registered function support for JDBC databases

2008-07-21 Thread Jody Garnett
A couple things; executing any sql function is not a good thing from a security standpoint (but you know this). You should be able to advertise additional functions on a data store by datastore basis using the filter capabilities data structure. You are the first person to want to do this so pl