Sure, give it a shot!
> If memory serves me correctly, we have released the examples as an artifact
Would not be a good idea. They contain some binaries which are not meant for
distribution (licenses).
They are also way too big to deploy them all.
Bundling them in a source zip is fine though.
> I _think_ I have something
> where a release:prepare/perform should work. Any objections to me taking a
> shot at it?
Sure, go on.
I just want to avoid doing all that stuff manually.
If it doesn't work out we can still try to extract the examples into an own
project.
LieGrue,
strub
> Am 02.07.2018 um 15:02 schrieb Jonathan Gallimore
> <jonathan.gallim...@gmail.com>:
>
> If memory serves me correctly, we have released the examples as an artifact
> in the past, albeit as one large zip with the code in, as opposed to
> individual artifacts. After some playing around, I _think_ I have something
> where a release:prepare/perform should work. Any objections to me taking a
> shot at it?
>
> Jon
>
> On Thu, Jun 21, 2018 at 7:38 AM, Mark Struberg <strub...@yahoo.de.invalid>
> wrote:
>
>> yes, the local build on my 2600X takes 42 minutes including all the Bean
>> Validation and CDI TCKs.
>> But the full EE6 TCK is run on literally 100s of spot instances (donated
>> by Tomitribe afaik).
>> The full TCK is only available under NDA so only a handful people have
>> access to it otoh :(
>>
>> This is what I strongly assume to change with JakartaEE!
>> After that we will be able to openly give access to the TCK results as
>> well.
>>
>> LieGrue,
>> strub
>>
>>
>>
>>> Am 21.06.2018 um 08:22 schrieb Gurkan Erdogdu <cgerdo...@gmail.com>:
>>>
>>> Aha 3 h is ok, assumed 13 h :) Actually I was running the TCK in my
>> laptop
>>> less than 3h
>>>
>>> On Thu, Jun 21, 2018 at 9:20 AM, Romain Manni-Bucau <
>> rmannibu...@gmail.com>
>>> wrote:
>>>
>>>> Last time i checked it was 3h of tck - assuming you have the setup -
>> and 3h
>>>> of tomee build (all-adapters profile to ensure we cover all tomee
>> flavors
>>>> and not only one as in the default build).
>>>>
>>>> However 3h of tck is due to a highly parallel execution (kudo David) and
>>>> will not be that few without the full infra. So if you want to check
>> some
>>>> coverage it easily takes that much or more yes.
>>>>
>>>> Le jeu. 21 juin 2018 07:30, Gurkan Erdogdu <cgerdo...@gmail.com> a
>> écrit :
>>>>
>>>>> Romain, do you mean that each release running with TCK takes 13h?
>>>>>
>>>>> On Wed, Jun 20, 2018 at 9:33 AM, Romain Manni-Bucau <
>>>> rmannibu...@gmail.com
>>>>>>
>>>>> wrote:
>>>>>
>>>>>> Le mer. 20 juin 2018 à 07:59, Mark Struberg <strub...@yahoo.de.invalid
>>>>>
>>>>> a
>>>>>> écrit :
>>>>>>
>>>>>>> A build should work out of the box and not require 2 weeks of first
>>>>>> fixing
>>>>>>> sporadically broken unit tests.
>>>>>>>
>>>>>>
>>>>>> A long time ago we agree the CI was the platform of truth. Indeed the
>>>>> fixes
>>>>>> are welcomed but we cant run the 13h of build locally each time (yes
>>>> TCK
>>>>>> are part of our project quality and therefore should be considered
>> part
>>>>> of
>>>>>> our build ;))
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> And a release usually should be
>>>>>>>
>>>>>>> $> mvn release:prepare
>>>>>>> $> mvn release:perform
>>>>>>>
>>>>>>
>>>>>> Doesn't fit tomee cause of the lifecycle of these two tasks which are
>>>>>> design but centralized sources (not git which allows to bypass part of
>>>>> the
>>>>>> loonngg steps by design).
>>>>>> Feel free to update the doc to the actual procedure if the existing
>>>> pages
>>>>>> are not that great but these steps are not the real issue we hit. The
>>>> one
>>>>>> we face is generally the review since we have a lot (did I say a lot?
>>>> ;))
>>>>>> of artifacts and this is why it got tooled in the "tools" repo (sub
>> svn
>>>>>> tree).
>>>>>>
>>>>>> Feel free to ping me if you need help to push the artifacts, can have
>> a
>>>>>> spot in the day today for that (likely beginning of the afternoon,
>>>> after
>>>>> it
>>>>>> will be hard for me).
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> that's it.
>>>>>>>
>>>>>>> Everything else is not really user friendly and will make it harder
>>>> for
>>>>>>> any new committer to get on board.
>>>>>>> I get the argument test coverage. But some of these examples still
>>>> use
>>>>>> the
>>>>>>> javaee6 apis, etc...
>>>>>>>
>>>>>>> LieGrue,
>>>>>>> strub
>>>>>>>
>>>>>>>> Am 20.06.2018 um 06:48 schrieb Romain Manni-Bucau <
>>>>>> rmannibu...@gmail.com
>>>>>>>> :
>>>>>>>>
>>>>>>>> Dropping the example will require to move their tests in the main
>>>>> chain
>>>>>>>> since they are part of our coverage.
>>>>>>>>
>>>>>>>> Also note you probably dont want to use release plugin cause
>>>> running
>>>>>> the
>>>>>>>> test (with -Pall-adapters if you respect the plugin philosophy) is
>>>>>> quite
>>>>>>>> long (should be ~3h x2). Just tag and deploy manually from a green
>>>>>> build
>>>>>>> on
>>>>>>>> the CI, this is more reliable for tomee.
>>>>>>>>
>>>>>>>> Le mer. 20 juin 2018 00:14, Jonathan Gallimore <
>>>>>>> jonathan.gallim...@gmail.com>
>>>>>>>> a écrit :
>>>>>>>>
>>>>>>>>> Sure, I'll take a look tomorrow morning.
>>>>>>>>>
>>>>>>>>> On Tue, 19 Jun 2018, 21:57 Mark Struberg,
>>>> <strub...@yahoo.de.invalid
>>>>>>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> it's license is ALv2, so not a biggie.
>>>>>>>>>> Do you put it onto your list, Jon?
>>>>>>>>>> txs and LieGrue,strub
>>>>>>>>>>
>>>>>>>>>> On Tuesday, 19 June 2018, 22:11:23 CEST, Jonathan Gallimore <
>>>>>>>>>> jonathan.gallim...@gmail.com> wrote:
>>>>>>>>>>
>>>>>>>>>> Definitely a left over, they should be self contained.
>>>>>>>>>>
>>>>>>>>>> Jon
>>>>>>>>>>
>>>>>>>>>> On Tue, 19 Jun 2018, 21:04 Mark Struberg,
>>>>> <strub...@yahoo.de.invalid
>>>>>>>
>>>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>>> while fixing the versions of almost all our examples I found the
>>>>>>>>>>> following parent pom
>>>>>>>>>>> <parent>
>>>>>>>>>>> <groupId>org.tomitribe</groupId>
>>>>>>>>>>> <artifactId>oss-parent</artifactId>
>>>>>>>>>>> <version>2</version>
>>>>>>>>>>> </parent>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> do we really want this?I thought our examples should be
>>>>>>> self-contained,
>>>>>>>>>>> isn't?
>>>>>>>>>>> Guess that's just an oversight and a leftover from a code
>>>>> donation?
>>>>>>>>>>> txs and LieGrue,strub
>>>>>>>>>>>
>>>>>>>>>>> On Tuesday, 19 June 2018, 21:49:40 CEST, Mark Struberg
>>>>>>>>>>> <strub...@yahoo.de.INVALID> wrote:
>>>>>>>>>>>
>>>>>>>>>>> And blowing up badly :(
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> [ERROR] Failed to execute goal
>>>>>>>>>>> org.apache.maven.plugins:maven-release-plugin:2.5.3:prepare
>>>>>>>>> (default-cli)
>>>>>>>>>>> on project tomee-project: The version could not be updated:
>>>>>>>>>>> ${tomee.version} -> [Help 1]
>>>>>>>>>>> org.apache.maven.lifecycle.LifecycleExecutionException: Failed
>>>> to
>>>>>>>>> execute
>>>>>>>>>>> goal org.apache.maven.plugins:maven-release-plugin:2.5.3:
>>>> prepare
>>>>>>>>>>> (default-cli) on project tomee-project: The version could not be
>>>>>>>>> updated:
>>>>>>>>>>> ${tomee.version}
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> On Tuesday, 19 June 2018, 21:15:06 CEST, Mark Struberg
>>>>>>>>>>> <strub...@yahoo.de.INVALID> wrote:
>>>>>>>>>>>
>>>>>>>>>>> Rolling a 7.0.5 release now.
>>>>>>>>>>>
>>>>>>>>>>> Doing this on my linux box, so I hope I didn't forget to setup
>>>>>>>>> anything.
>>>>>>>>>>> LieGrue,strub
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>
>>
