I'll try to add this to the dev section of our site this weekend.
On Tue, Jan 6, 2015 at 6:18 PM, Kristian Rosenvold
kristian.rosenv...@gmail.com wrote:
Conclusion: We're off to 1.6-land !
- Update to jdk 6.0 at will, but please be sure that we're not
leaving the last 1.5 version in a
Conclusion: We're off to 1.6-land !
- Update to jdk 6.0 at will, but please be sure that we're not
leaving the last 1.5 version in a regressed state.
- Minor version bump for jdk 1.6 upgrade, remember to tag jira version
as First 1.6 version (in manage versions)
- To achieve this it is probably
Go for Java 7! Or 8!
Gary
div Original message /divdivFrom: Tibor Digana
tibordig...@apache.org /divdivDate:01/01/2015 15:07 (GMT-05:00)
/divdivTo: dev@maven.apache.org /divdivCc: /divdivSubject: Re:
[DISCUSS] Move everything to 1.6, take 2 (was: Re: I can't make
a
JDK1.5 and 1.6 are unsupported anymore.
JDK 1.7 is still long alive and under maintenance.
The Java SE 7 API won't be taken back due to whatever JVM fault :))
The JDK 8 is alive too short.
I don't see any reason why the default Maven plugins have to go with awful
1.5 or 1.6.
We can freely
On Sun, Dec 28, 2014 at 9:04 PM, Robert Scholte rfscho...@apache.org
wrote:
Op Sun, 28 Dec 2014 19:37:47 +0100 schreef Kristian Rosenvold
kristian.rosenv...@gmail.com:
I'll sumarize what appears to be our consensus so far.
Update to jdk 6.0 at will, but please be sure that we're not
In my experience, there are significant API issues between 3.0 and
3.1. My particular obsession is with the plugin testing harness.
I've had several experiences of the following forn:
1: go to fix a problem in a plugin.
2: try to create an appropriately focussed test
3: try to set up the testing
personnally, I use invoker: no compatibility problems
Regards,
Hervé
Le mardi 30 décembre 2014 17:45:36 Benson Margulies a écrit :
In my experience, there are significant API issues between 3.0 and
3.1. My particular obsession is with the plugin testing harness.
I've had several
On Tue, Dec 30, 2014 at 6:36 PM, Hervé BOUTEMY herve.bout...@free.fr wrote:
personnally, I use invoker: no compatibility problems
I try to make unit tests when I can make unit tests. it's a religion,
or a disease. I agree that falling back to the invoker is the
fallback.
Regards,
Hervé
Le
I'll sumarize what appears to be our consensus so far.
Update to jdk 6.0 at will, but please be sure that we're not leaving
the last 1.5 version in a regressed state.
Version number indicates minimum maven version, so a simple JDK
upgrade only mandates a minor version update.
We are also in a
Op Sun, 28 Dec 2014 19:37:47 +0100 schreef Kristian Rosenvold
kristian.rosenv...@gmail.com:
I'll sumarize what appears to be our consensus so far.
Update to jdk 6.0 at will, but please be sure that we're not leaving
the last 1.5 version in a regressed state.
Version number indicates minimum
Le dimanche 28 décembre 2014 21:04:50 Robert Scholte a écrit :
Op Sun, 28 Dec 2014 19:37:47 +0100 schreef Kristian Rosenvold
kristian.rosenv...@gmail.com:
I'll sumarize what appears to be our consensus so far.
Update to jdk 6.0 at will, but please be sure that we're not leaving
the
Did we already cover what we want to keep supporting via Toolchains?
We would have to take some care in Surefire if we wanted to keep some
support for 1.6 when using toolchains or when allowing users to configure
a different JVM.
2014-12-25 15:57 GMT+01:00 Karl Heinz Marbaise
Hi Kristian,
I am +1 for any Release Manager wanting to up the minimum Java version
to 1.6 for any of our components, on one condition: if there are any
bugs fixed since the last release of the component, then please do a
final Java 5 compatible release of the component before moving it to
Java
It appears that IBM JDK6 is EOL september next year. People move at
different speeds :)
Kristian
2014-12-25 6:25 GMT+01:00 Gary Gregory garydgreg...@gmail.com:
+1
Gary
div Original message /divdivFrom: Benson Margulies
bimargul...@gmail.com /divdivDate:12/24/2014 17:08
... and when I say CodeHaus above, I mean Apache.
Fair?
;)
2014-12-25 13:11 GMT+01:00 Lennart Jörelid lennart.jore...@gmail.com:
Quite true.
:)
But this opens another interesting discussion.
Do we move the codehaus products with the slowest of the major JDK release
cycles (i.e. to match
Quite true.
:)
But this opens another interesting discussion.
Do we move the codehaus products with the slowest of the major JDK release
cycles (i.e. to match the IBM JDK release cycle in this case)?
Or with the Oracle JDK's release cycles?
There may not be much difference in the mechanics of
+1 for moving to at least 1.6 or even 1.7. While 1.8 would be the release
with more interesting features, I think requiring this would be too early.
Regards
Mirko
--
Sent from my mobile
On Dec 25, 2014 1:12 PM, Lennart Jörelid lennart.jore...@gmail.com
wrote:
Quite true.
:)
But this opens
Hi,
let me summarize things a little bit:
Last time discussed this we established a consensus to establish 3.0.5
(maybe 3.0.6) as a minimum baseline for the 3.x range of plugins.
http://www.mail-archive.com/dev@maven.apache.org/msg102539.html
that was not three months ago...so the line to
Oops. Snappy contains 1.6 java bytecode, which breaks the build on maven
plugins. We need to upgrade to 1.6; I'm taking this to the mailing list :)
Last time discussed this we established a consensus to establish 3.0.5
(maybe 3.0.6) as a minimum baseline for the 3.x range of plugins.
This 3.0.X
We already discussed this so many times
But seriously with 2015 coming really soon I believe it's time.
Finally so many years after java 1.5 EOL! :-)
--
Olivier
On 25 Dec 2014 00:20, Kristian Rosenvold krosenv...@apache.org wrote:
Oops. Snappy contains 1.6 java bytecode, which breaks the
+1
if someone really wants to stay with JDK 5, just don't update plugins to
latest and greatest
and IMHO, if we need to maintain Maven 3.0.x in parallel from 3.2.x, that's
not because of the JDK prerequisite: that's because there are problems to
upgrade some plugins because of Aether change
+1
(Hoping we can get up to 1.7 soon too)
On Wednesday, 24 December 2014, Kristian Rosenvold krosenv...@apache.org
wrote:
Oops. Snappy contains 1.6 java bytecode, which breaks the build on maven
plugins. We need to upgrade to 1.6; I'm taking this to the mailing list :)
Last time discussed
+1, would also make testing with JDK9 easier, although I've already found
a good solution for that.
Robert
Op Wed, 24 Dec 2014 14:20:06 +0100 schreef Kristian Rosenvold
krosenv...@apache.org:
Oops. Snappy contains 1.6 java bytecode, which breaks the build on
maven plugins. We need to
First: +1 for 1.6 minimum.
Second: I feel we need to take a more strategic look at java in general and
plugin mechanics dependencies in particular. 1.6 is deprecated since a
few years - and while its bytecode runs fine on a JDK 8 runtime, any
implicit dependencies and internal reflection magic
+1.
jdk 1.6 is EOL-ed for some time (Feb 2013) already and even 1.7 will be
EOL-ed in April 2015..
I would suggest moving straight to 1.7 but I guess that's been already
discussed.
Milos
On Thu, Dec 25, 2014 at 7:54 AM, Robert Scholte rfscho...@apache.org
wrote:
+1, would also make testing
Here's what I don't understand. I can see why people need to keep
building apps that run on antediluvian version. I can't see why it's
such a problem for a tool, such as Maven, to require 1.7. Who are we
accomodating by the current policy, or even the 1.6 plan?
Meanwhile, it seems to me that we
I assume that anyone wishing for 1.7 will also accept 1.6.
I would really just like to establish a consensus that we're leaving
1.5 in favour of 1.6. We have a certain tradition for being last to
leave jdk versions and I don't really mind this. It *does* become a
problem when it makes practical
+1
Gary
div Original message /divdivFrom: Benson Margulies
bimargul...@gmail.com /divdivDate:12/24/2014 17:08 (GMT-05:00)
/divdivTo: Maven Developers List dev@maven.apache.org /divdivCc:
/divdivSubject: Re: [DISCUSS] Move everything to 1.6, take 2 (was: Re: I
can't make a
28 matches
Mail list logo
