* Andrew Overholt <overholt@xxxxxxxxxx> [2005-01-26 09:47]:
>
> I'm hoping to go back to a 3.1 milestone build before FC4.
Looking at the schedules for Eclipse 3.1 [1] and FC4 [2], I see the
following:
Eclipse 3.1 M6 (stable build, API freeze): 2005-04-01
FC4 Absolute devel freeze: 2005-05-02
Eclipse 3.1 M7 (stable build, development freeze, etc.): 2005-05-13
FC4 Release open, announced: 2005-06-16
I'd like to ship a 3.1 M6 (or some stable incremental build between then
and FC4 freeze) with FC4 and then update to M7 and 3.1 final as they come
out.
Choices:
a) ship as above
b) ship 3.0.x
3.1 Pros:
- between now and then we get to spend lots of time testing and getting
things pushed upstream
- this is where all development is happening
- awesome new plugins like Jeff Pound's ongoing work on the Bugzilla
plug-in are dependent upon 3.1 (without a tonne of back-porting)
- making RPMs will be much simpler than how we're doing the 3.0.1 now
- upstream supports many more architectures than they did with 3.0.1
3.1 Cons:
- lots of stuff that may not work with libgcj4/gij4
- a lot less tested than 3.0.x
3.0.x Pros:
- well-tested, established, stable code
- already have RPMs and know they build okay with gcj4
- will be the latest upstream _released_ version
3.0.x Cons:
- old code where new features are not being added
- will need back-porting if we want new things
- build is a bit of a mess (especially for arches that were unsupported
upstream)
I'd like to go with 3.1, but we'll need to do lots of testing and
bug-fixing. The number of gcj/libgcj hackers capable of fixing bugs isn't
that high and even getting things down to test cases is not an easy task.
I'd like to hear people's opinions.
Thanks,
Andrew
[1]
http://fedora.redhat.com/participate/schedule/
[2]
http://eclipse.org/eclipse/development/eclipse_project_plan_3_1.html
