Re: [m2e-dev] locking m2e minimal required eclipse platform version

[Anders Hammar <anders@xxxxxxxxxx>]

One problem I see with this is that to use any of the m2e release milestones, one would need to use the target yearly Eclipse milestone release as well. I can only speak for myself, but I would never use such an Eclipse platform for my development environment and thus I would then not be able to test the latest m2e release milestone in my dev environment.

Are you using the milestones of the target yearly release for the m2e development?

/Anders

On Fri, Mar 28, 2014 at 7:26 PM, Igor Fedorenko <igor@xxxxxxxxxxxxxx> wrote:

Some background. Historically, m2e always supported at least one older
release of Eclipse Platform. For example, m2e 1.4 released with Kepler
was also fully supported on Juno. So when Fred offered to release m2e
1.4.1 to provide m2e java 8 support on Kepler, I was not very supportive
to the idea (to put it mildly): extra release will mean extra support
surface for developers and extra confusion to the users, so why do it
when m2e 1.5 works on Kepler? After some rather interestingly worded
discussion, it turned out while m2e 1.5 itself works with Kepler just
fine, downstream m2e consumers, like m2e-wtp and jbosstools, do not
work, or at least it will be rather time consuming to assemble
everything required.

In other words, m2e project policy to support previous Eclipse Platform
version is not very useful in practice and pretty much amounts to wasted
development effort. (again, I wish we could collect actual usage stats,
but alas, Eclipse Foundation does not allow this).

So my proposal is to introduce minimal required Eclipse Platform version
and lock it to the target yearly release. For m2e 1.5 it will be Luna
(3.10 version of org.eclipse.osgi and org.eclipse.jdt bundles). For m2e
1.6 it will be the version(s) of these bundles shipped with Mars, and so
on. This will reduce development load somewhat and will also eliminate
any confusion about m2e and Eclipse Platform version compatibility.

Seems like a good deal to me, but I wonder if other developers and users
see any problems with this proposal?

--
Regards,
Igor
_______________________________________________
m2e-dev mailing list
m2e-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/m2e-dev