Re: [Geotools-devel] Prevent silent failures of online tests

[Jody G. <jod...@gm...>]

We need to talk to the project steering committee on this one (in 
tomorrows IRC meeting?) since we are talking about changing the build 
system.
I am way on a training course and cannot attend ... can you either 
attend; or we can start up a vote on email here.

The reason for the connection() not being considered fatal is because 
sometimes servers are down; in the past we had the build fail when ever 
some random server on the Internet is busted - which was not cool.

As I understand it you expecting people that run online tests to provide 
their own fixtures; and thus it there own problem if the server is down?

For me I am thinking about the various WMS Servers that we run against; 
these ones are not controlled by a fixture file (and I do not want the 
build broken if I can simply not connect).

So it looks like we are trying to handle two forms of online tests?
Jody


> Jody Garnett wrote:
>> Ben can you write up something sane and we can review the proposal in 
>> tomorrows IRC meeting? Either a Jira as you propose;
>
> Jody, I have created a patch and attached it to this Jira issue:
> http://jira.codehaus.org/browse/GEOT-1951
>
> Please let me know if this is enough.
>
>> or a policy change for the developers guide if you want all supported 
>> modules to behave in the same way.
>
> docs.codehaus.org is down for maintenance. I do not know the policy. I 
> just recommend that unit tests not swallow exceptions without good 
> reason.
>
>> Cory Horner started looking at this issue before he left Refractions; 
>> perhaps you can pick up where he left off?
>> My understanding is that the first connection failure should log the 
>> exception?
>
> At the moment, it just calls Throwable.printStackTrace(), which will 
> be lost in the storm.
>