Re: [Geotools-devel] Small fix on xml-xsd, but wondering about the possibility of similar errors

[Justin D. <jde...@op...>]

Hi Andrea,

Good catch. This is something i came across more than a couple of times.
 If you look at a lot of the binding testing cases ( like in gml module
) all of them usually test this case.

However, there are lots that don't still. So i had the idea of having
the code generator automatically generate binding unit tests as well.
The test created would fail out of the box until the implementor
actually implemented the getTarget() method.

We could also just generate a single test case that tested all of them
in batch as you suggest.

-Justin

Andrea Aime wrote:
> Hi Justin (hi all),
> I stumbled on an xml-xsd issue today: XSLongTypeBinidng was returning
> a null type instead of Long.class.
> The type is needed for the parser to work, and if not declared, it'll
> just break.
> 
> I committed the obvius fix, but I'm wondering if some unit
> test can be designed to mass test all the bindings for this kind of
> error. Do all the binding need to return a type? I can't remember.
> 
> Btw, if you're wondering how did I hit XSLongTypeBinding, it's by using
> this GML geometry:
> 
> <gml:Point srsName="http://www.opengis.net/gml/srs/epsg.xml#26713">
>           <gml:coordinates decimal="." cs="," ts="
> ">604000,4930000</gml:coordinates>
>         </gml:Point>
> 
> which has integral coordinates (not uncommon in projected srs)
> 
> Cheers
> Andrea
> 
> !DSPAM:1004,45f59872183011804284693!
> 


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