Hi Josh,

I think that was pretty much the same way it worked with Flexmojos.

I do recall the problem then being that the server had to allocate a socket and 
to use that number in the compilation of the test-runner.
I always thought this was problematic. 

Do you have a way of calling the runner and passing along the port so the 
client will connect to dynamically allocated ports?

Chris


Am 01.06.20, 17:41 schrieb "Josh Tynjala" <[email protected]>:

    Sure, let's get this figured out!

    Some background: For the RoyaleUnit Ant task, I basically created a fork of
    the original FlexUnit Ant task. The task runs a server that listens for
    socket connections from Royale apps, and the apps report their test results
    over the socket. For the SWF side, I don't think that I needed to change
    anything. It just worked. For the JS side, I needed to add support for
    connections over WebSockets because raw sockets aren't supported in JS/HTML.

    --
    Josh Tynjala
    Bowler Hat LLC <https://bowlerhat.dev>


    On Sun, May 31, 2020 at 3:41 AM Christofer Dutz <[email protected]>
    wrote:

    > Starting a new thread about this … don’t enjoy too multi-topic monster
    > threads.
    >
    > Hi all and specially Josh.
    >
    > I was told that you added something RoyaleUint which should work the same
    > way as the old FlexUnit … great thing. This definitively was missing.
    > Currently it seems that these tests can be run in an Ant build but not in
    > a Maven build.
    >
    > We should change that ASAP :-)
    >
    > Having worked a lot on Flexmojos I think I have a fair idea of how we
    > could integrate this into a new Royale surefire runner.
    >
    > Would you: Josh be interested with working on this with me? I bet I will
    > have some questions.
    >
    > Chris
    >
    >
    >

Reply via email to