Andrea Aime wrote:
> Ok ok... having that visitor everywhere filters are build or consumed
> is going to be very ugly, but if that's "the way", so be it...
Thanks for understanding Andrea; I mostly want the ability to represent
exactly what the user asked for.
Having a sanity check before you en
Jody Garnett ha scritto:
> Andrea Aime wrote:
>> Validation? Warnings? Sorry?
>>
> I thought that was what you were wanting; apparently i was wrong.
>> A filter like "INCLUDE and INCLUDE and EXCLUDE" is perfectly valid,
>> it's just redundant. I was speaking about simplification, not validation.
Andrea Aime wrote:
> Validation? Warnings? Sorry?
>
I thought that was what you were wanting; apparently i was wrong.
> A filter like "INCLUDE and INCLUDE and EXCLUDE" is perfectly valid,
> it's just redundant. I was speaking about simplification, not validation.
>
Okay for simplification I w
Jody Garnett ha scritto:
> I understand the motivation; and I don't want to inflict logic
> expectations on factory writers if we can find another way. Can we make
> a FilterFactoryWrapper that performs validation. The code can find
> problems like this; and depending on a "strict" flag produce
Andrea Aime wrote:
> Another solution would be to introduce a smarter builder
> for filters, but fact is, most of the code is using FilterFactory
> and it's simple enough that most people won't look for
> the builder even if there was one.
>
Depends if the build adds any value; as we move into f
Hi,
today I fixed a bug in Oracle related to encoding
Filter.INCLUDE/Filter.EXCLUDE. The bug was there, no excuse,
but the filter that was going to be encoded was like:
INCLUDE AND INCLUDE AND geom bbox (...)
which is kind of silly. One workaround for this would be to
have datastores encoding fi