[geotk] NullPointerException with WFSFeatureStore

[Martin Desruisseaux]

Le 10/12/14 17:05, emmanuel.blondel1 at gmail.com a écrit :
> Ok for a snapshot, i will argue here on the fact we should use first a snapshot until M3 is available, it shouldn't be a problem.
Thanks

> Some questions:
> - Is it reasonable to make this timeout hardcoded in geotk? Isn't there a way to specify this timeout threshold as client param..?
Yes it should be a parameter. I'm not familiar with the way
WFSFeatureStore uses optional parameters. If I find an obvious place
where to put it, I will do. Otherwise a static constant and a "TODO"
note gives an opportunity to revisit this question by someone more
familiar. There is a danger that a "TODO" note is forgot for a long
time, but it should be no longer than until the code is ported to SIS
(since we try to resolve pending issue in this process).

> - if it is only a time 'warning', ie not an exception, why should i have a null getcapabilities? I would expect a result anyway.
I agree, this is why I proposed alternatives in my previous email. I
would prefer an exception to be thrown, and Johann seems to agree. Would
it be okay for you?

>  - in case it's'considered as failure, i should have a timeout exception (but it supposes that i can specify the timeout, otherwise it would be very restrictive and will not work for large GetCapabilities docs)
Yes I agree for the TimeoutException. I will try to find a place for the
timeout parameter.

    Martin