openwebbeans-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Karl Kildén <karl.kil...@gmail.com>
Subject Re: Nullpointer in WebContextsService
Date Wed, 02 Mar 2016 16:09:14 GMT
I will register the ticket now but I won't have time to fix it until
tomorrow. Here's a preview
https://github.com/karlkilden/openwebbeans/commit/0b39aba0594dec50ef96d9e403787c3ac5032cc9?diff=unified

I am not sure what style you want on the patch. I broke out some methods
but maybe you prefer as small change as possible?

I will also include tests unless it's complicated for some reason

Cheers

On 2 March 2016 at 14:53, Romain Manni-Bucau <rmannibucau@gmail.com> wrote:

> wonder if throwing an exception context not active instead of logging
> passes the tcks, if so can be enough.
>
>
> Romain Manni-Bucau
> @rmannibucau <https://twitter.com/rmannibucau> |  Blog
> <http://rmannibucau.wordpress.com> | Github
> <https://github.com/rmannibucau> | LinkedIn
> <https://www.linkedin.com/in/rmannibucau> | Tomitriber
> <http://www.tomitribe.com>
>
> 2016-03-02 14:29 GMT+01:00 Thomas Andraschko <andraschko.thomas@gmail.com>
> :
>
>> Thats true Romain...
>>
>> We have some code where we need to check if we are running in an http
>> request and not e.g. in a background worker or on startup, to determine if
>> we can take some settings from the user session.
>>
>> As i said, it's no problem if you are checking RequestScoped &&
>> SessionScoped. RequestScoepd is not active and therefor the lazyStartSessionContext
>> will not happen.
>>
>> so +0, it was just an idea :)
>>
>> 2016-03-02 14:20 GMT+01:00 Romain Manni-Bucau <rmannibucau@gmail.com>:
>>
>>> About the log: originally it was not there and it is a nightmare to
>>> understand OWB is not well setup in your container. The log shouldn't
>>> happear - as much as you shouldn't even try to use session scope in a
>>> thread where it is not available theorically. A consensus can be to log it
>>> once but in "big" (like taking 20 lines of logs just to say something is
>>> not correct). A step further can be to return a fake context never active
>>> in such cases.
>>>
>>> Wydt?
>>>
>>>
>>> Romain Manni-Bucau
>>> @rmannibucau <https://twitter.com/rmannibucau> |  Blog
>>> <http://rmannibucau.wordpress.com> | Github
>>> <https://github.com/rmannibucau> | LinkedIn
>>> <https://www.linkedin.com/in/rmannibucau> | Tomitriber
>>> <http://www.tomitribe.com>
>>>
>>> 2016-03-02 13:56 GMT+01:00 Thomas Andraschko <
>>> andraschko.thomas@gmail.com>:
>>>
>>>> The method in deltaspike is ok, it just checks if the context is active
>>>> in the current thread.
>>>>
>>>> +1 for jira and a patch.
>>>>
>>>> 2016-03-02 13:47 GMT+01:00 Karl Kildén <karl.kilden@gmail.com>:
>>>>
>>>>> Right, Deltaspike hides the ugly that you call a boolean method but
>>>>> instead of false you get ContextNotActiveException. One of the more
>>>>> mystical parts of the specification if you ask me. Maybe you are refereeing
>>>>> to how the BeanManager should behave if the thread is not managed?
>>>>>
>>>>> It is simply a little bit over my head but at least nothing in the
>>>>> javadoc for ContextUtils in deltaspike forbids usage from unmanaged threads
>>>>> and frankly if that was a requirement it seems like a poor idea to offer
>>>>> such static methods in the first place since managed threads have
>>>>> injectionpoints.
>>>>>
>>>>> Regarding a JIRA och a patch: I will take care of it.
>>>>>
>>>>> Thanks for helping me understanding this better
>>>>>
>>>>> cheers / Kalle
>>>>>
>>>>
>>>>
>>>
>>
>

Mime
View raw message