New To The Forum? Click Here To Read The How To Guide. -- Developers Click Here.

Issue with version 1.27 and SDK v1.26

bknothbknoth Posts: 3
NerveGear
Our systems auto-updated to version 1.27 last night. Calls to ovr_GetInputState() now produce values for m_inputState.IndexTrigger[0] that are always 0 no matter how the trigger is positioned. We're using version 1.26 for the SDK (I haven't seen a newer version). Has anyone else seen this problem?

Thanks,
Bruce

Comments

  • ramezanifardramezanifard Posts: 1
    NerveGear
    Hi Bruce, 
    the same problem here. I can't read the buttons anymore. The only thing that still works is "inputState.Touches". It shows what button is touched, but is super sensitive to touch.  
  • bknothbknoth Posts: 3
    NerveGear
    Luckily, we had a computer that had not auto-upgraded. We copied the Oculus program files folders from that machine (1.26) to a machine that had auto-upgraded to 1.27 and the latter machine worked again. We had a critical demo to pull off and that was our work-around.
  • bknothbknoth Posts: 3
    NerveGear
    Please post if it does resolve the problem. I'm not well-versed in Oculus development and I'd appreciate knowing if this problem's been solved. Thanks.
  • MikeFMikeF Posts: 423
    Nexus 6

    Others who have been developing for Oculus devices- has this been a common occurrence in the past? Automated updates breaking your production environment? Definitely makes me hesitant to continue supporting this platform if it isn't going to be reliable. 
    Big time, we've had several important demo's derailed because of a last minute update forced on us.
    When things work it's great but the uncertainty of when an update hits and what it might break is a big problem
  • NinjaGaijinNinjaGaijin Posts: 66 Oculus Staff
    Hello everyone,

    Our engineers are aware of the problem and are currently looking into a solution.

    In the meantime, a workround is to press the Oculus button to enter the Universal Menu, then press it again to return to the game. The buttons should work then.
    Please tag me, @NinjaGaijin, in your forum post/response if you need immediate assistance or want additional support/context on an issue you are having with our software or integrations.

    Having an issue with our platform, services, or integrations? Try using our new bug tool to report this and receive emailed updates as we proceed to address it internally.
  • kojackkojack Posts: 4,570 Volunteer Moderator
    Just thought I'd join in, the same issue has stopped my AutoOculusTouch project from working. It uses ovrInit_Invisible as it's intended to run without the headset worn at all. But all button and trigger values are zero unless rendering is performed and the headset is being worn. Wearing (or triggering the sensor) without rendering or rendering without wearing both result in input stopping (tracking continues to work though).


  • CloudhopperCloudhopper Posts: 5
    NerveGear
    I work with Bruce who initiated this thread.  We have just built two new rigs with fresh software installs.  I attempted the workaround posted by NinjaGaijin, without success.  I have copied the non-updated Program Files/Oculus directory from a working system, and turned off automatic updates, but this solution will not work with our customers.  Please @NinjaGaijin, this has been an issue since June 19. 
  • imperativityimperativity Posts: 3,465 Oculus Staff
    @Cloudhopper

    We are still pursuing a fix internally at this time. Sorry for the long delay here. This regression straddles a number of build process and source control changes that are making it difficult to bisect. 
    Please tag me @imperativity in your forum post/response if you need immediate assistance or want additional support or context on an issue you are having with our software or integrations.

    Having an issue with our platform, services, or integrations?

    Try using our new bug tool to report this and receive emailed updates as we proceed to address it internally.
  • YvesMartelYvesMartel Posts: 6
    NerveGear
    Any news on a fix for this problem?
  • YvesMartelYvesMartel Posts: 6
    NerveGear

    I am running SDK V1.29 and that problem is still there.

    My application has 2 VR modes, when I switch modes, I first get out of VR and get back in with a new ovr_Initialize and a new session (ovr_Create).  To get the ovr_GetInputState to work again in the new session, just having the headset on is not enough. You then need to remove the headset and put it back on again!  This make for awkward demos when I want to show both modes to a client...

    The Oculus button work around does work, but after all, it is a work around...  A real solution would be better.


  • imperativityimperativity Posts: 3,465 Oculus Staff
    @YvesMartel

    Yes, this issue is still present at this time and is being internally addressed at this time. 
    Please tag me @imperativity in your forum post/response if you need immediate assistance or want additional support or context on an issue you are having with our software or integrations.

    Having an issue with our platform, services, or integrations?

    Try using our new bug tool to report this and receive emailed updates as we proceed to address it internally.
  • CloudhopperCloudhopper Posts: 5
    NerveGear
    @imperativity We are preparing a build for a customer to be delivered later this week.  Has there been any progress on enabling successful reads of the Touch trigger state from the SDK?  I have been unable to update any of our development systems since June, and we are still running 1.26 which is the last system that worked.  How is it that many systems (like UE4) are able to read the triggers, but the SDK can't?
  • zakk102zakk102 Posts: 1
    NerveGear
    @imperativity

    This bug is very critical to our application. Do you have any schedule to fix this bug?
  • imperativityimperativity Posts: 3,465 Oculus Staff
    @zakk102

    There is no ETA on a fix for this.
    Please tag me @imperativity in your forum post/response if you need immediate assistance or want additional support or context on an issue you are having with our software or integrations.

    Having an issue with our platform, services, or integrations?

    Try using our new bug tool to report this and receive emailed updates as we proceed to address it internally.
  • CloudhopperCloudhopper Posts: 5
    NerveGear
    edited November 7
    @imperativity
    @NinjaGaijin

    The Sample software provided with the SDK "t.Controllers"  illustrates reading buttons from the touch.  I have modified this code to read the IndexTrigger, and it works.  There are many examples (all of them) of applications that are able to utilize the trigger., yet this issue, if fundamental, would be a blocker for all applications utilizing the Touch controllers.  The workaround of pressing the menu button twice does nothing to correct the errant behavior in our application.  Why is this not shown as a know bug in the current SDK release?  Is it really just a handful of folks that use the SDK interface to access the hardware?  How does Unity or UE4 get the correct value?

  • CloudhopperCloudhopper Posts: 5
    NerveGear
    @imperativity
    @NinjaGaijin

    The t.Controllers sample stops updating the trigger unless the function  "API.ovr_CommitTextureSwapChain.Ptr"  in file OVR_CAPIShim.c  is called.  In our application, the VR display is rendered in a separate application, so we need to read the state of the buttons and triggers without a graphics window associated with the application.  This worked fine up until 1.27.
  • CloudhopperCloudhopper Posts: 5
    NerveGear
    @imperativity
    @NinjaGaijin

    The t.Controllers sample stops updating the trigger unless the function  "API.ovr_CommitTextureSwapChain.Ptr"  in file OVR_CAPIShim.c  is called.  In our application, the VR display is rendered in a separate application, so we need to read the state of the buttons and triggers without a graphics window associated with the application.  This worked fine up until 1.27.
Sign In or Register to comment.