you can vote if you like :-)
good job to everyone, looking forward to seing 3.4!
On Aug 7, 2012, at 10:41 PM, Matt Wringe <mwringe(a)redhat.com> wrote:
On Mon, 2012-08-06 at 17:12 +0200, Julien Viet wrote:
> On Aug 6, 2012, at 5:01 PM, Matt Wringe <mwringe(a)redhat.com> wrote:
>
>> On Sat, 2012-08-04 at 10:47 +0200, Julien Viet wrote:
>>>
>>> Julien Viet
>>>
julienviet.com
>>>
>>>
>>> On Saturday, August 4, 2012 at 4:01 AM, Matt Wringe wrote:
>>>
>>>> On Sat, 2012-08-04 at 00:11 +0200, Julien Viet wrote:
>>>>> do you mean that bolek tag 3.4.0.M01 is not correct and we need to do
it after your PR ?
>>>>
>>>>
>>>> The device property based redirects currently don't work properly
(well,
>>>> not at all right now), user-agent redirects do work properly which are
>>>> more important. Not sure if this is really important enough to do
>>>> another tag.
>>>>
>>>>
>>>
>>>
>>> ok
>>>>
>>>> I would have done a PR by now, but my tests are failing due to issues
>>>> with the kernelbootstrap is not properly getting disposed and eating all
>>>> the memory (also, why are we still using such an old version of junit
>>>> without support for @beforeclass and @afterclass?)
>>>>
>>>>
>>
>> Is there anything special I need to do to release the resources used by
>> the kernelbootstrap? Calling bootstrap.dispose does not release the
>> resources and eventually leads to the heap space running out of memory.
>
> the kernel bootstrap is a nice facade in front of a pile of crap : had to use some
hack to shut down things.
>
> can you identify the leak ?
>
>
>>
>> Since I can't use @beforeclass or @afterclass (to only create the
>> kernelbootstrap before the suite and dispose of it after). I will have
>> to recreate the bootstrap for each test and since the bootstrap doesn't
>> release its resources, it cause the running out of memory before the
>> tests are completed.
>
> can you identify the leak with JProfiler or GC verbosity ?
A quick look at the heap dump shows that its most likely a problem with
jcr/jboss cache not being released (org.jboss.cache.lock.MVCCLockManager
specifically). But I am not really sure if I am looking at the results
correctly in the eclipse memory analyzer.
I updated the tests to get around this issue by only having junit call
one method for the tests instead of multiple methods (so I can create
the bootstrap at the start of this method and dispose of it at the end,
and not have to do this for each test method).
> actually I whish we can upgrade to JUnit 4, as we could use @RunWith in some places
(looking at Byteman).
>
> however those classes are used by some UXP team (at least social) as basis for their
unit testing, so we should ensure that stuff will remain as compatible as possible.
>
>
>>
>>> because it was working fine until now and it could take time to migrate unit
tests.
>>>
>>>>
>>>>> Julien Viet
>>>>>
julienviet.com (
http://julienviet.com)
>>>>>
>>>>>
>>>>> On Friday, August 3, 2012 at 11:39 PM, Matt Wringe wrote:
>>>>>
>>>>>> On Thu, 2012-08-02 at 20:39 +0200, Boleslaw Dawidowicz wrote:
>>>>>>> Here is tag:
>>>>>>>
https://github.com/gatein/gatein-portal/tree/3.4.0.M01
>>>>>>>
>>>>>>> Here are download links:
>>>>>>>
http://www.jboss.org/gatein/downloads/gateinportal
>>>>>>>
>>>>>>> Here is documentation:
>>>>>>>
http://www.jboss.org/gatein/documentation/gateinportal
>>>>>>>
>>>>>>> as always I would be grateful if everyone try out links above
and report problems.
>>>>>>>
>>>>>>> At this point we need blog posts from Matt about mobile and
from Julien about Javascript improvements. It doesn't make sense to announce software
on Friday so we will probably push the news on Monday. Julien will handle it.
>>>>>>
>>>>>> Just a quick update: I have a blog post and docs mostly ready for
the
>>>>>> mobile stuff. I unfortunately ran into a couple of minor issues
with the
>>>>>> mobile stuff while writing the docs (double checking that
everything was
>>>>>> working as expected). Will post the docs/blog and do a pull
request
>>>>>> either later tonight or tomorrow with the fixes and updated
tests.
>>>>>>
>>>>>>> I will be mostly disconnected next two weeks. In case we need
to repackage things again Trong has all needed permissions and proper knowledge. Thomas
can help with Magnolia on
jboss.org (
http://jboss.org).
>>>>>>>
>>>>>>> Thanks everyone for hard work and help with this release!
>>>>>>>
>>>>>>> Bolek
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> gatein-dev mailing list
>>>>>>> gatein-dev(a)lists.jboss.org
(mailto:gatein-dev@lists.jboss.org)
>>>>>>>
https://lists.jboss.org/mailman/listinfo/gatein-dev
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> gatein-dev mailing list
>>>>>> gatein-dev(a)lists.jboss.org (mailto:gatein-dev@lists.jboss.org)
>>>>>>
https://lists.jboss.org/mailman/listinfo/gatein-dev
>>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>