You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@isis.apache.org by Andi Huber <ah...@apache.org> on 2018/02/22 10:21:45 UTC
Testing seems broken!
Hi folks,
I've found a strange discrepancy between tests our surefire maven plugin
discovers compared to what eclipse discovers. Eclipse discovers more
tests, which is a rather *SEVERE* issue regarding our current surefire
setup!
I've opened an issue [1] to investigate this.
I already fixed all failing tests (as picked up by eclipse, but not with
surefire) within the 'core/metamodel' [2].
That lead to the discovery of 2 bugs [3][4] in 'master' and a
discrepancy between doc and code [5].
Cheers, Andi!
[1] https://issues.apache.org/jira/browse/ISIS-1869
[2] https://issues.apache.org/jira/browse/ISIS-1866
[3]
https://github.com/apache/isis/commit/dd9588d9e76b21dca07eb36956eca1f462d90fb7
[4]
https://github.com/apache/isis/commit/2d1c331d8dfe59381014ce5f07df0733c496020e
[5]
https://github.com/apache/isis/commit/d938728320057c01bb078fb0650cfb0a5bef4249
Current state of branches
Posted by Andi Huber <ah...@apache.org>.
Summarizing current state of branches:
== master (2.0.0-M1)
* simpleapp/helloworld do not compile on my eclipse checkout
* allows disabling of footer
* making progress with removing guava from API (applib)
* swagger-ui now packaged with metamodel (removed from example apps and artifacts)
* local-resource links fixed (by introducing LocalResourcePath value type), allows proper actions that link to swagger-ui or object's restful-API
* further refactored reusable code into Internal API (algorithms for byte array processing: base64 encoding + compression)
* pulled in all code necessary for JavaEE support from M2
* all JUnit Tests pass
== 2.0.0-M2
* has name 'dev/2.0.0-M2-cherrypicked' and is rebased on today's master (M1)
* simpleapp/helloworld do not compile on my eclipse checkout
Cheers,
Andi
Re: Rebasing again
Posted by Andi Huber <ah...@apache.org>.
When pulling in changes from M2 into M1, I needed to do some
hand-picking of code fragments. In the process we lost some original
commit history, replaced by consolidated commits with issue ISIS-1871 [1].
Instead of re-basing our latest M2 on top of the newly assembled M1, I
branched off on M1 fresh and cherry-picked commits only related to the
transition to DN-5.
The new branch for M2 is 'dev/2.0.0-M2-cherrypicked', which may be
renamed to its original after review.
Cheers,
Andi
[1] https://issues.apache.org/jira/browse/ISIS-1871
On 23.02.2018 10:03, Andi Huber wrote:
> I'm in the process of rebasing M2 on top of M1 (master) again, but
> before I do that, I'm pulling changes, that do not conflict with DN-4,
> from M2 into M1.
>
> Remember:
>
> * 2.0.0-M1 is a transition to JDK 8
> * 2.0.0-M2 is a transition to Datanucleus 5
>
> Cheers,
> Andi
>
Rebasing again
Posted by Andi Huber <ah...@apache.org>.
I'm in the process of rebasing M2 on top of M1 (master) again, but
before I do that, I'm pulling changes, that do not conflict with DN-4,
from M2 into M1.
Remember:
* 2.0.0-M1 is a transition to JDK 8
* 2.0.0-M2 is a transition to Datanucleus 5
Cheers,
Andi
Re: Testing seems broken!
Posted by Dan Haywood <da...@haywood-associates.co.uk>.
On Thu, 22 Feb 2018 at 13:53 Andi Huber <ah...@apache.org> wrote:
> Update:
>
> It seems for test-inclusion surefire defaults to a naming-convention
> [6], which we clearly do not always follow:
>
> * |Test*|
> * |*Test|
> * |*TestCase|
>
> However, we are able to override this behavior and include all tests
> regardless of naming. I've pushed the updated pom.xml to 'master'
> (2.0.0-M1). We have some tests to fix now, that went 'hidden' previously.
>
> This might be a candidate do be back-ported to 1.16.x.
>
>
I'm hoping there won't be any more 1.16.x or 1.17 releases, so let's not
bother backporting.
Onwards and upwards!
Cheers
Dan
> Cheers, Andi!
>
> [6]
>
> https://stackoverflow.com/questions/6178583/maven-does-not-find-junit-tests-to-run
>
> On 22.02.2018 11:21, Andi Huber wrote:
> > Hi folks,
> >
> > I've found a strange discrepancy between tests our surefire maven plugin
> > discovers compared to what eclipse discovers. Eclipse discovers more
> > tests, which is a rather *SEVERE* issue regarding our current surefire
> > setup!
> >
> > I've opened an issue [1] to investigate this.
> >
> > I already fixed all failing tests (as picked up by eclipse, but not with
> > surefire) within the 'core/metamodel' [2].
> >
> > That lead to the discovery of 2 bugs [3][4] in 'master' and a
> > discrepancy between doc and code [5].
> >
> > Cheers, Andi!
> >
> > [1] https://issues.apache.org/jira/browse/ISIS-1869
> >
> > [2] https://issues.apache.org/jira/browse/ISIS-1866
> >
> > [3]
> >
> https://github.com/apache/isis/commit/dd9588d9e76b21dca07eb36956eca1f462d90fb7
> >
> > [4]
> >
> https://github.com/apache/isis/commit/2d1c331d8dfe59381014ce5f07df0733c496020e
> >
> > [5]
> >
> https://github.com/apache/isis/commit/d938728320057c01bb078fb0650cfb0a5bef4249
> >
> >
> >
>
>
Re: Testing seems broken!
Posted by Andi Huber <ah...@apache.org>.
Update:
It seems for test-inclusion surefire defaults to a naming-convention
[6], which we clearly do not always follow:
* |Test*|
* |*Test|
* |*TestCase|
However, we are able to override this behavior and include all tests
regardless of naming. I've pushed the updated pom.xml to 'master'
(2.0.0-M1). We have some tests to fix now, that went 'hidden' previously.
This might be a candidate do be back-ported to 1.16.x.
Cheers, Andi!
[6]
https://stackoverflow.com/questions/6178583/maven-does-not-find-junit-tests-to-run
On 22.02.2018 11:21, Andi Huber wrote:
> Hi folks,
>
> I've found a strange discrepancy between tests our surefire maven plugin
> discovers compared to what eclipse discovers. Eclipse discovers more
> tests, which is a rather *SEVERE* issue regarding our current surefire
> setup!
>
> I've opened an issue [1] to investigate this.
>
> I already fixed all failing tests (as picked up by eclipse, but not with
> surefire) within the 'core/metamodel' [2].
>
> That lead to the discovery of 2 bugs [3][4] in 'master' and a
> discrepancy between doc and code [5].
>
> Cheers, Andi!
>
> [1] https://issues.apache.org/jira/browse/ISIS-1869
>
> [2] https://issues.apache.org/jira/browse/ISIS-1866
>
> [3]
> https://github.com/apache/isis/commit/dd9588d9e76b21dca07eb36956eca1f462d90fb7
>
> [4]
> https://github.com/apache/isis/commit/2d1c331d8dfe59381014ce5f07df0733c496020e
>
> [5]
> https://github.com/apache/isis/commit/d938728320057c01bb078fb0650cfb0a5bef4249
>
>
>
Re: Active branches rebased
Posted by Dan Haywood <da...@haywood-associates.co.uk>.
Thanks, Andi, appreciate the work you're doing paying back some of our
technical debt!
Cheers,
Dan.
On Thu, 22 Feb 2018, 19:54 Andi Huber, <ah...@apache.org> wrote:
> Hi folks,
>
> I've again re-based '2.0.0-M2' on top of today's 'master' (=2.0.0-M1)
> and branch 'ISIS-1779-jax-rs-2' re-based on top of '2.0.0-M2'.
>
> Cheers, Andi!
>
Active branches rebased
Posted by Andi Huber <ah...@apache.org>.
Hi folks,
I've again re-based '2.0.0-M2' on top of today's 'master' (=2.0.0-M1)
and branch 'ISIS-1779-jax-rs-2' re-based on top of '2.0.0-M2'.
Cheers, Andi!