Hervé and I discussed on irc earlier today.
My suggestion - I was going to write up tomorrow - is that we drop 7 for
We should focus 3.6.x on making the codebase Java 8 (lambdas, replace
File____Stream with the non-file handle leak version, adopt Path, etc) with
a view to encourage contributions... and best support for Java 9 (such that
if you run j9 we recommend 3.6.x)
We can keep 3.5.x alive for security issues and S1/S2 backports for the
next 6 months or so. Close all other lines.
75% of people were fine with requiring Java 8
Also remember that Java 7 ships with out of date TLS root CA certs and iirc
does not support SNI, so the reality is you shouldn't be running Maven with
Java 7 unless you are within a secure intranet with a corp proxy that can
perform the TLS verification for you.
(Now openjdk 7 is a different story, but if we say some subset of 7 then
people will assume oracle 7 also)
On Sun 10 Sep 2017 at 17:52, Arnaud Héritier <[hidden email]> wrote:
> But on twitter thus probably far from being a good representation of our
> users ...
> Le dim. 10 sept. 2017 à 18:45, Gary Gregory <[hidden email]> a
> écrit :
> > On Sep 10, 2017 10:33, "Robert Scholte" <[hidden email]> wrote:
> > So what would be the conclusion?
> > 35% want to keep Java7 as JRE for Maven for a shorter or longer period?
> > IMHO that's a lot
> > but less than the 65% who do not, an overwhelming majority.
> > Gary
> > Robert
> > On Sun, 10 Sep 2017 11:55:20 +0200, Stephen Connolly <
> > [hidden email]> wrote:
> > So poll results:
> > >
> > > 493 votes cast
> > >
> > > 25% want Java 7,8&9 for Maven 3.6.x
> > > 65% want Java 8&9 for Maven 3.6.x
> > > 10% want Java 8&9 for Maven 3.6.x and 6 months of backporting to 3.5.x
> > >
> > > On Sat 9 Sep 2017 at 11:50, Stephen Connolly <
> > > [hidden email]> wrote:
> > >
> > > https://twitter.com/asfmavenproject/status/906451059966693376
> > >> --
> > >> Sent from my phone
> > >>
> > >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [hidden email]
> > For additional commands, e-mail: [hidden email]
> Arnaud Héritier
> Mail/GTalk: aheritier AT gmail DOT com
> Twitter/Skype : aheritier
Sent from my phone
Am 2017-09-10 um 19:07 schrieb Stephen Connolly:
> Hervé and I discussed on irc earlier today.
> My suggestion - I was going to write up tomorrow - is that we drop 7 for
> We should focus 3.6.x on making the codebase Java 8 (lambdas, replace
> File____Stream with the non-file handle leak version, adopt Path, etc) with
> a view to encourage contributions... and best support for Java 9 (such that
> if you run j9 we recommend 3.6.x)
This will take months to properly rewrite Maven core. I'd rather see
this for Maven 4 and continue to polish Maven with 3.5.x, 3.6.x, etc.
Unless no one will do this within a short timeframe, I see no benefit here.
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]
In reply to this post by stephenconnolly
Only lurking those days, but chiming in quickly here :-).
2017-09-10 20:54 GMT+02:00 Robert Scholte <[hidden email]>:
> On Sun, 10 Sep 2017 20:21:11 +0200, Stephen Connolly <
> [hidden email]> wrote:
> On Sun 10 Sep 2017 at 19:04, Tibor Digana <[hidden email]> wrote:
>> Hi All,
>>> Are we facing new API regarding networking and security useful in Java 8?
>>> When I first saw these options I asked myself what benefit would have the
>>> User and Jenkins from Java 8.
>>> And second question was whether we would be so flexible to rewrite the
>>> and use Lambda fully anywhere in the code.
>> There is the social aspect. If you are a potential new contributor to
>> and you look at our heavy Java 1.3 convention codebase (ok, I'm being a
>> demagogue, it's had a bit updated to 5.0) are you going to be encouraged
>> step forward?
>> How can you make small improvements and demonstrate you are a safe pair of
>> hands to gain the commit bit?
>> Now if we have the opertunity to make lots of tidy up and you can show you
>> are a safe pair of hands, retaining binary compatibility with older
>> plugins, making the code more readable, finding file handle leaks, etc...
>> well now you have a welcome path to demonstrate your skills.... while
>> gaining familiarity with the codebase so that when we turn around to start
>> on Maven 5.0.x you can join in the fun.
>> Now that is not a technical argument, but we are a community first... so
>> maybe the technical arguments are not so right to push!
> bq. but we are a community first
> What's the definition of community: the large user group or the few we're
> hoping to find who can help working on Maven?
> And if we're focusing on the latter, shouldn't the twitter question be:
> I want to become a Maven developer
>  right now!
>  only if Maven requires Java 8 ( so I can use lambda's, etc.)
This is definitely one the things we put forward for moving Jenkins to Java
8 recently , keeping contribution appealing.
Because, well, that's great the Maven community would try to help large
companies not be bothered by any kind of upgrade, like Free Enterprise
software support :-).
But I think this is a dead-end, those kind of teams/companies won't upgrade
Looking at the stats, I would think there's actually not much debate:
people not wanting to upgrade, well, won't upgrade... And proportion is not
What is the actual issue anyway? They will keep using 3.5.x for 3 years? So
what, I bet they already do. No, wait, they're not, they're using 2.2.1.
I just checked, the company I left 1 year ago is still running 3.2.1, and I
don't expect them to upgrade anytime soon...
Maven backward compatibility is great, so IMO don't bother spending time
The Maven dev team should IMO concentrate of making Maven great, not
supporting a big variety of versions/flavours developers will hate the
codebase each time they remember "urgh, I can't use this!".
The less combinatorials, the better for anyone working on her/his free time
My 2 cents
|Free forum by Nabble||Edit this page|