Fix joystick and gamepad library before jME 3.1 stable?

pspeed · October 23, 2015, 7:15am

I wish I could upvote this one more. It’s nice to know why so many people have problems with the joystick test lately.

zzuegg · October 23, 2015, 7:19am

What the… heck. Now that is super strange. Since i did not add the dll’s on my side manually, but i use the master from jme for testing purposes.

//Add:
@Tankwart i guess you are using 3.0 stable?

Tankwart · October 23, 2015, 7:29am

@zzuegg: Yes I’m using the 3.0 stable. But like I said a few postings before, if you need some results from the 3.1 alpha, I would download it and give it a try.

Here the complete informations which jme prints out when I choose “Help” → “About”

Product Version: jMonkeyEngine SDK 3.0
Java: 1.7.0_51; Java HotSpot(TM) 64-Bit Server VM 24.51-b03
Runtime: Java(TM) SE Runtime Environment 1.7.0_51-b13
System: Windows 8 version 6.2 running on amd64; Cp1252; de_DE (jmonkeyplatform)
User directory: C:\Users\Tankwart\AppData\Roaming\.jmonkeyplatform\3.0
Cache directory: C:\Users\Tankwart\AppData\Roaming\.jmonkeyplatform\3.0\var\cache

zzuegg · October 23, 2015, 7:30am

Since my xbox controller does also not work with the TestJoystick i first try to get my side working, but i come back to the offer

Tankwart · October 23, 2015, 7:41am

Wait… I think there is a little misunderstanding in the room! I do NOT need a joystick-control at the moment! I just wanted to deliver some results, cause it seems, that a lot of users have problems with using their sticks. I’m not able to fix these problems by myself, so I wanted to help out with results for them who are able to. That means that there is no need to fix one of my mentioned problems in the proceding thread, as long as they are unnessecary for the people who debug the “core”.

Long story short: I thought you wanted to overhaule the input-system with the goal, that everyone can use it with almost every joystick. So the misunderstanding was on my side… sorry for that!

pspeed · October 23, 2015, 7:46am

You can already use it with almost every joystick… even many of the xbox style controllers.

zzuegg · October 23, 2015, 7:48am

For that first it is to figure out what currently is wrong with the system.

There is some issue with rumblers. (Hard to fix, since it is usually a driver thing, or in case of linux a right management thing)
Some joysticks are reported to work with jinput, but not with jme.

Now second depends how people figured out that the joystick is not working. If they used the TestJoystick to validate than the issue is that the Test does not enable the JoyInputSystem

But it is interesting, that there where nearly no posts regaring joystick issues for a long time. And if they where, it was mostly mapping specific

One feature that i also would like is to be able to access the single rumblers. Since i can see usecases for that

Tankwart · October 23, 2015, 7:54am

As I said, I just wanted to help I thought that there is a problem somewhere in lwjgl or jme, which leads to not working joysticks.

Darkchaos · October 23, 2015, 8:40am

Infact you solved that problem with your investigation.

As for the dll’s: When you add the jar Files to the project jme will extract the dll/so/dylib Files for your OperatingSystem so you don’t need them but they are extracted once and then stay there.

Delete them and they will be reextracted.

For a git you want to add them to gitignore

Ogli · October 23, 2015, 11:27am

What about Yves initial problem? Has a gamepad that works with TestJoystick and jinput test, has a joystick that works with jinput but not TestJoystick. Also, I did never use TestChooser and it seems that Yves didn’t use TestChooser too.
Since Yves didn’t report back, I suppose he did not have time yet or did not understand how to set up and start zzueggs test project.

Rumblers #1: Yes, access to single rumblers is needed. As I said - just another method in the core, next to the setRumble method that is already there.

Rumblers #2: The usage of rumblers seems to be fundamentally wrong - either in jME or in jinput. I will investigate this, to see what’s wrong (sunday is still free for me).

@pspeed Do you know which version of jinput jME is using?

I’m interested what the jinput test does different - different jinput version, different Java version, any additional tricks, etc. - since it is an open source project, that should be possible. (Another thing for my sunday experiments).

Ogli · October 23, 2015, 11:38am

So… as zzuegg said here

So your test case was quite pointless, since your gaming devices already worked with jME.

It’s Yves who has a device that should be tested like that.

I don’t know why he did not post back (no time or doesn’t understand how to start zzueggs test?) Seems that we must wait until he reports back.

In the meantime I could have a look at why rumble is so terribly wrong and how to do it right.

Tankwart · October 23, 2015, 12:11pm

Sorry for delivering pointless results! As I said a few times before, I only wanted to help! I’m not this deep in the topic as I could find out if my tests are necessary or not. I just wanted to help… and have a little spare time to do so. Once more: Sorry! It won’t happen in the future anymore… I promise!

EDIT: If an admin is reading this: Feel free to remove my unnessecary postings!

EDIT of EDIT: My first edit sounds a little bit offended. But I never was offended by Oglis comment! I just got the Thread wrong and I hate it myself, if I have to read 500 OTs before the thread comes back to its topic. Everythings ok

david_bernard_31 · October 23, 2015, 12:14pm

In my project I override the version of jinput to fix some issues on linux

github.com

davidB/jme3_skel/blob/master/build.gradle#L40


    compile "org.jmonkeyengine:jme3-core:${v_jme}"
    compile "org.slf4j:slf4j-api:${v_slf4j}"
    compile "com.jme3x:jfx:1.+"
    compile 'io.reactivex:rxjava:1.1.1'
    compile 'com.github.davidb:jme3_ext:0.5.0:all'
    runtime "org.jmonkeyengine:jme3-lwjgl:${v_jme}"
    runtime("org.jmonkeyengine:jme3-desktop:${v_jme}"){
        exclude module: "jinput"
    }
    runtime 'net.java.jinput:jinput:2.0.6'
    runtime "org.jmonkeyengine:jme3-jogg:${v_jme}"
    //runtime "org.slf4j:slf4j-jdk14:${v_slf4j}"
    compile "org.slf4j:jul-to-slf4j:${v_slf4j}"
    runtime 'ch.qos.logback:logback-classic:1.1.2'
    testCompile 'junit:junit:4.12'
    testCompile 'org.mockito:mockito-core:1.8.5'
}
sourceSets {
    main {
        resources {

I’ll also take a look when back to home. and I’ll provide an updated version of my test app (where I display control).

Ogli · October 23, 2015, 12:16pm

Ehm … it’s okay. I think we all appreciate your enthusiasm.
No bad emotions connected when saying “quite pointless”.

I think you helped to improve zzueggs test software.

And the OT you posted sounds very interesting too… really would like to make a light gun game with jME now! :chimpanzee_closedlaugh:

pspeed · October 23, 2015, 12:34pm

I have no idea. I think we might have inherited it with lwjgl but it was already in long before I messed with the joystick stuff and rewrote parts of it.

Ogli · October 23, 2015, 12:53pm

LOL, another one of those. :chimpanzee_closedlaugh:
If anyone messes things up then it’s probably me (might just put a warning sign in my code).

I thought that you as a core dev might know whether it’s always the latest version and if not, which max version the jME will use.

Idea: it might be interesting to give the jME 3.1 version of the jinput library to the jinput test application and see if this Hotas works or not. Then we know for sure if it’s because of a different version or not.

Tankwart · October 23, 2015, 1:19pm

Attention, noob is talking again I just searched for the jinput.jar and opened it with 7zip. In this package there is a class named Version.class . It isn’t completly readable, but it shows up a version-number which is “2.0.0-b01”. Maybe it helps you. I’m sure that it would be possible to get the Version within java, cause there seems to be a getter called getVersion().

While googeling I found out, that there are a few never versions available. 2.0.6 seems to be the latest. But don’t nail me on that

jakebriggs · October 23, 2015, 7:34pm

That explains a lot, I struggled with this for a while!

Yves_Tanas · October 24, 2015, 1:18pm

I’m back in. Had some trouble with my PC, but i am back in running and overwhelmed, what has been done here.

Okt 24, 2015 3:23:53 PM net.java.games.input.DefaultControllerEnvironment getControllers
WARNUNG: Found unknown Windows version: Windows 10
Okt 24, 2015 3:23:54 PM net.java.games.input.DefaultControllerEnvironment getControllers
WARNUNG: Attempting to use default windows plug-in.
Okt 24, 2015 3:23:54 PM net.java.games.input.ControllerEnvironment log
INFORMATION: Failed to initialize device USB Keyboard because of: java.io.IOException: Failed to acquire device (8007001e)

T.Flight Hotas X
18
Stick
Unknown
0
Taste 0
Taste 1
Taste 2
Taste 3
Taste 4
Taste 5
Taste 6
Taste 7
Taste 8
Taste 9
Taste 10
Taste 11
Mehrwegeschalter
X-Achse
Y-Achse
Schieberegler
Z-Rotation
Z-Achse
BUILD SUCCESSFUL (total time: 1 second)

Yves_Tanas · October 24, 2015, 1:31pm

I worked on my own spaceshooter for a long time, first steering just by keys, then moving on to a gamepad x-box-style and finally went over to that hotas.
It cost 50€ so it’s quite cheap but also very robust, good quality, good wages.
I just ran over that problem, because i could not just switch the joysticks.
I still can use my gamepad normally but the hotas is out of order.
Otherway the JInput Test works fine with it and i can use it really good in unity engine for the same purpose.
I went over there, tried some things, to look, if this would be a better solution.
Maybe it is, but i have made much work in here i would like to build out more than really begin again in unity. Even if it really looks damn cute in unity.