Jody Garnett ha scritto:
> Summary:
> - http://jira.codehaus.org/browse/GEO-141 Formally track this little 
> adventure of ours (so Martin does not kick me)
> - http://jira.codehaus.org/browse/GEOT-1916 Update the Hints to mention 
> the new factories
> 
> Andrea Aime wrote:
>>> Difficulty is that this either needs to be a structures exception 
>>> message (indicating say the xpath to the problem and kind of problem) 
>>> or we should  provides a utility method to do this kind of check.
>> I was leaning towards a structured exception message.  Like, I'd 
>> prefer to keep all the validation logic in one place,  and possibly in 
>> an obvious one, splitting it around does not look nice to me.
>> IllegalAttributeException is already structured, it has the descriptor 
>> of the failing attribute and the value that  does not conform to it, 
>> too bad it's not exposing them, but  it's just a matter of adding a 
>> couple of getters.
> Okay - so can you decisive and direct me on this one?

Sure, I'll follow you up on GEO-141

>> If the strict option is meant to help debug only, what about a system 
>> variable to enable it?
> Yes; the idea is to swap out factories - see hint below

So you can have a system wide hint, right?

>>>> It seems to me this kind of change would require at least a one day 
>>>> sprint... what do we do in the meantime?
> If you want to call a sprint on this stuff I will join you; we are too 
> close to our deadline on the 15th here to do much. Perhaps we can chat 
> online and sort this out?

Eh yeah, I already exhausted my mandate (1 week of work) so I'll
only be able to work on this during the weekend.
Actually being able to use a factory is sort of a new requirement
that only collaterally touches the feature model speedup. For the
15th deadline I can switch features to use explicit validation,
changing all datastores to allow the usage of a hint
is a much bigger task.

>>> Backing up:
>>> - Our goal was to make the FeatureTypes static utility method 
>>> available for people to find; we have now done that correct?
>> More or less, we do if we provide a validate() method that throws an 
>> exception giving people the full meal, otherwise we have to
>> keep that Types utility class around, have isValid() call it,  and 
>> then have the user call it again if he wants to know what is  not 
>> valid, bleck...
> Okay so you want to change the method; I can do that now. Please review 
> when you wake up.
>> System wide == configurable? Where is that located? 
> When I last looked it was a Hint; let me check ... okay we got a 
> FEATURE_FACTORY but no FEATURE_FACTORY hint yet; something to fix... 
> http://jira.codehaus.org/browse/GEOT-1916

Cool

>> Or is it just a hard coded default in the builder?
> No the builder should either:
> - be created with a factory
> - or look up the default in the usual manner

Agreed

>>> Aside: The idea of remembering the factory seems like a little bit of 
>>> gain; while I had similar thoughts they come from a different problem 
>>> When working with Shapefiles only a limited set of content is 
>>> supported (max string length etc...). Having 
>>> ShapefileDataStore.getFeatureTypeFactory() return a factory that 
>>> knows about these limitations is a sensible way to go...
>> Uuh, not sure... the ideas of factories is to put the user in control 
>> not the lower levels?
> Yeah you see the conflict eh? (oh wait that was a Canadian moment).

Uh? What's Canadian about it? :)

> I agree that the idea of factories is to put the user in control; we 
> still may need something like "capabilities" to let people know what 
> data formats are supported ... something for another day?

Yeah, most probably? ;)
Cheers
Andrea

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to