Re: [CAPRA Development] Java version for building?
- From:
- David Winslow
- Date:
- 2010-07-08 @ 14:10
On 07/08/2010 09:52 AM, Gabriel Roldan wrote:
> On 7/8/10 10:47 AM, Andrea Aime wrote:
>> Gabriel Roldan ha scritto:
>>> On 7/8/10 8:14 AM, Andrea Aime wrote:
>>>> Hi,
>>>> noticed the geonode-geoserver-ext module uses a Java 6 only
>>>> constant, TimeUnit.MINUTES.
>>>>
>>>> Are the GeoNode Java parts supposed to be Java 6 only?
>>>> Just curious
>>>
>>> It's mostly that I didn't know TimeUnit.MINUTES were introduced at
>>> Java6. But, given Java5's end of service, I would actually recommend
>>> adopting Java6 for GeoNode?
>>>
>>> Concerns?
>>
>> Only that I asked already twice on geotools devel in the last six months
>> and the idea of switching was rejected twice in a row for EE reasons,
>> people reporting that Java 5 is still quite in use in the enterprise.
>>
>> Is GeoNode targeting traditional big enterprise deployments?
> If you mean deploying on an organizations rigid infrastructure using
> websphere 5 or such.... I don't know. I always thought GeoNode deploys
> as its own thing, controlling the jetty server it launches
> geoserver/geonetwork on.
>
> So at this stage getting rid of TimeUnit.MINUTES is not a big deal. We
> should know what the expected deployment scenarios are asap though,
> before more Java6'ish stuff gets in?
>
>> If so staying with Java 5 is safer.
> as said, that'd be trivial so far. Seb/Dwins?
>
> Thanks,
> Gabriel
>>
>> Cheers
>> Andrea
>
In GeoNode we have traditionally made this decision by asking "has
anyone asked for it?". All deployments thus far have been on dedicated
machines so Java versions haven't been restricted (and Java 5 is EOL for
crying out loud.)
So, we don't have any requirements right now that dictate Java 5
compatibility, but if it's easy and you two see some value in staying
compatible (for example, it makes it easier to move the process service
to GeoServer community modules later) then I will be easily swayed.
--
David Winslow
OpenGeo - http://opengeo.org/