Justin Deoliveira ha scritto:
> On 10-05-28 10:25 AM, Andrea Aime wrote:
>> Andrea Aime ha scritto:
>>> The idea of grafting a jdbc specific bit in the FeatureTypeInfo
>>> interface does not really make me jump of joy. Alternatives?
>>> Extending the brief map converter to support complex objects as
>>> values I guess, which kind of seems to invalidating the reason
>>> to have a brief converter in the first place, isn't it?
>> Ah, I see now that MetadataMap has a get(key, Class target)
>> method using converters. So having a converter between String and
>> VirtualTable would work.
>> However it would be nice if the persister went also the other
>> way, and used converters instead of toString()... or used
>> MetadataMap.get(key, String.class) which might have a special
>> handling of the string case (toString() if the converters
>> don't return anything else)
>>
>> Sounds better?
> 
> Yeah. We could also try to make the MetadataMap smarter and instead of 
> just blindly trying to convert to a string for every object only convert 
> for "primitives". But if you want to go the converter route that sounds 
> good too.

The downside of using a converter will be that you'll get a property
like thing inside the xml value, which won't look too pretty. Something
like:
<entry key="JDBC_VIRTUAL_TABLE">
query=select a, b, the_geom from ...
geom.the_geom=point,4326
id=a,b
</entry>

I mean, it's valid, but looks out of place ;-)
(well, assuming the xml output and reading preserves newlines, those
are important in a property file format).
Or I can just get a bit creative and make a converter generating
a xml-like dump using DOM manipulation

Cheers
Andrea


-- 
Andrea Aime
OpenGeo - http://opengeo.org
Expert service straight from the developers.

------------------------------------------------------------------------------

_______________________________________________
Geoserver-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

Reply via email to