Marcel
Thanks for the clarification and fix. That is an interesting observation
on 1.4 EOL and it's implication for support.
-- Rob
Marcel Offermans wrote:
On Oct 15, 2008, at 7:41 , Rob Walker wrote:
Noticed that when trying to build under 1.4, I got the following
compile error:
symbol : method signum (long)
location: class java.lang.Math
I think this method is from 1.5 onwards [...]
I know 1.4 is quite old, but we still try and keep compatibility to
it. I think for the above case we'll be fine in fact, since it occurs
in DM - which we don't currently use. We're also debating mandating a
later JDK version for our App too - so again, it may not be an issue.
It's actually in the optional shell command for the DM, however it was
my intention to keep that bundle 1.4 compatible, so I just committed a
small fix for that.
Just wondering if we have a "JDK level" policy in terms of what Felix
and bundles should be buildable/runnable against.
Just thought I'd query if we have an official policy on Felix.
We have a policy (although it might not be written down anywhere) to
keep the core framework compatible with basically 1.4 (or rather the
foundation profile). All other bundles can pretty much do what they like.
It is an interesting question though how long we, or rather the OSGi
Alliance, wants to maintain a specification that's based on a JDK
version that is no longer supported by Sun[1]: the actual End Of Life
for 1.4 is the 30th of this month. Of course, the spec is based on the
JDK spec, not implementation, but I guess it will become harder to
find platforms and tools that will actually still run 1.4 outside of
maybe the embedded space. Luckily they defined execution environments,
can be run on top of Java 5 (which EOL's a year from now) and 6 too.
Greetings, Marcel
[1] http://java.sun.com/products/archive/eol.policy.html
--
Ascert - Taking systems to the Edge
[EMAIL PROTECTED]
+44 (0)20 7488 3470
www.ascert.com
