The point is my webservice client is part of my app and then need app config. The design cant change cause of tests. I can isolate it and mock ut through cdi but using config source is nicer Le 25 sept. 2013 20:39, "John D. Ament" <john.d.am...@gmail.com> a écrit :
> Yeah... the target path of the deployment isn't available at > deployment creation. It's only available after. > > When I was doing some webservice testing, i simply instantiated using > the URL param, not injection of the webservice (I honestly find > webservice injection to be a bit difficult since endpoints will be > different in environments). > > > > On Wed, Sep 25, 2013 at 2:35 PM, Jason Porter <lightguard...@gmail.com> > wrote: > > Ah, okay. Now I see. > > > > > > On Wed, Sep 25, 2013 at 12:12 PM, Romain Manni-Bucau > > <rmannibu...@gmail.com>wrote: > > > >> Yep but the app doesnt know it and arquillian doesnt have it in > packaging > >> phase (@deployment) > >> Le 25 sept. 2013 19:51, "Jason Porter" <lightguard...@gmail.com> a > écrit : > >> > >> > In that particular example, in the test, Arquillian knows the URL of > the > >> > server, so the port should already be there, right? Maybe I'm missing > >> > something. > >> > > >> > > >> > On Wed, Sep 25, 2013 at 10:51 AM, Romain Manni-Bucau > >> > <rmannibu...@gmail.com>wrote: > >> > > >> > > It would be a contextual config so in the case of arquillian you'd > set > >> it > >> > > in the beginning of your test method. > >> > > > >> > > The point is not if it works but if we can/should support it. > >> > > > >> > > typically how to configure a webservice client url when the port is > >> > random? > >> > > > >> > > *Romain Manni-Bucau* > >> > > *Twitter: @rmannibucau <https://twitter.com/rmannibucau>* > >> > > *Blog: **http://rmannibucau.wordpress.com/*< > >> > > http://rmannibucau.wordpress.com/> > >> > > *LinkedIn: **http://fr.linkedin.com/in/rmannibucau* > >> > > *Github: https://github.com/rmannibucau* > >> > > > >> > > > >> > > > >> > > 2013/9/25 Jason Porter <lightguard...@gmail.com> > >> > > > >> > > > I'm not sure what good a ThreadLocal is going to give you. Unless > >> > you're > >> > > > using @InSequence in your tests you're not guaranteed when the > tests > >> > will > >> > > > run and if that ThreadLocal variable will be set. Simply having > >> > > Arquillian > >> > > > inject the URL should be fine. Also if depending on the forking > >> > parameter > >> > > > with JUnit it may not work anyway. > >> > > > > >> > > > > >> > > > On Wed, Sep 25, 2013 at 5:01 AM, Romain Manni-Bucau > >> > > > <rmannibu...@gmail.com>wrote: > >> > > > > >> > > > > Hi, > >> > > > > > >> > > > > I have the following use case: a config is dynamic (typically > the > >> url > >> > > of > >> > > > > the server using arquillian - @ArquillianResource URL url). I > need > >> > this > >> > > > url > >> > > > > in a config. In prod i use apache-deltaspike.properties or a > custom > >> > > > > ConfigSource. I see an easy solution being a ThreadLocal (or a > >> global > >> > > > Map) > >> > > > > backing a TestConfigSource. > >> > > > > > >> > > > > The question now: do we provide a default impl answering this > need? > >> > > > (maybe > >> > > > > an in memory configuration == map/properties updatable through a > >> > static > >> > > > > method) > >> > > > > > >> > > > > wdyt? > >> > > > > > >> > > > > > >> > > > > *Romain Manni-Bucau* > >> > > > > *Twitter: @rmannibucau <https://twitter.com/rmannibucau>* > >> > > > > *Blog: **http://rmannibucau.wordpress.com/*< > >> > > > > http://rmannibucau.wordpress.com/> > >> > > > > *LinkedIn: **http://fr.linkedin.com/in/rmannibucau* > >> > > > > *Github: https://github.com/rmannibucau* > >> > > > > > >> > > > > >> > > > > >> > > > > >> > > > -- > >> > > > Jason Porter > >> > > > http://en.gravatar.com/lightguardjp > >> > > > > >> > > > >> > > >> > > >> > > >> > -- > >> > Jason Porter > >> > http://en.gravatar.com/lightguardjp > >> > > >> > > > > > > > > -- > > Jason Porter > > http://en.gravatar.com/lightguardjp >
