Possible undocumented physics change

  • If anyone in the dev team is able to respond on this, it would be great.

    It hasn't made it's way into a stable build yet, so I'm not sure if I should be worried. But after loading my project into r166, physics has been acting weirdly. Specifically, revolute joints. I use it to make a chain in my game that moves with the mouse, but as of r166 it's just been pieces of the joints flying all over the screen.

    I've looked through all the changelogs and found nothing. In addition to that, after loading the game into r166 and saving, and then reverting the version back to r164.3, the bug persists. So I'm thinking perhaps it's a change in the physics engine that's been bundled with the game?

    I've tested this on a light weight version of my game's feature as well.

  • Try Construct 3

    Develop games in your browser. Powerful, performant & highly capable.

    Try Now Construct 3 users don't see these ads
  • after loading the game into r166 and saving, and then reverting the version back to r164.3, the bug persists.

    So the project works fine in r164, you open it in r166, change nothing, just save in new version and it stops working? Then you edit savedWithRelease tag in c3proj file back to r164, open in r164 and it doesn't work correctly?

    Have you tried also changing Physics behavior version to r164.3 in the same file? (I don't know if this field if actually used)

    So I'm thinking perhaps it's a change in the physics engine that's been bundled with the game?

    I'm pretty sure this is not the case.

    By the way, why r166? The latest beta is r170.2

  • So the project works fine in r164, you open it in r166, change nothing, just save in new version and it stops working? Then you edit savedWithRelease tag in c3proj file back to r164, open in r164 and it doesn't work correctly?

    Yes, it works fine in r164 or below, and then load it into r166 and above and it stops working correctly, save it, and I edit the savedwithrelease tag back to r164 and it doesn't work correctly.

    Have you tried also changing Physics behavior version to r164.3 in the same file? (I don't know if this field if actually used)

    Yeah, tried this

    By the way, why r166? The latest beta is r170.2

    Problem starts from r166 and above, it's the same way for r170, not tested with r170.2

  • I still don't understand why it doesn't work when you revert back to 164. Behavior code is definitely not bundled with the project.

    But if you are sure that there is a problem with physics, I suggest making a small project to demonstrate the issue, save it in versions 164 and 166 and submit bug report.

  • It's going to take a lot of work to do that because I had to strip down my project to just the cursor/revolute joint part which took me hours on a large project once for the r166+ demo, and then I stopped for the r165 and below version because I have to restrip it down from a unbugged copy again, to post this incase a staff is able to indentify the bug without me doing it.

    I was hoping atleast. I will do it if I have to. I do think they might fix it before a stable version if they are aware of it so I can move to the next stable version smoothly!

  • Still curious why the same project doesn't work when you roll back to r164. It doesn't make sense.

    Also, Physics seems to me an overkill for this task. You can make a similar effect with Pin (rope style) behavior. It may not be as bouncy, but will be better for performance.

  • I checked our code logs and there have been no substantive changes to the Physics behavior this year. This means it's pretty unlikely there really was any breaking change in the Physics behavior, so I would suspect that's not actually the cause and you've mis-attributed something else. Additionally if you roll back to a previous release and it still doesn't work, that is a strong sign that there was not actually a change in Construct itself; it could well be a change you made to the project and mis-attributed to Construct.

    This is very common - the bug report guidelines are there to help you get to the bottom of the real issue and provide all the information we need to actually help, so I can only advise that you follow those.

  • Thank you for taking the time to look in more detail Ashley.

    I understand that, it's very logical, it's most likely not the physics behavior then, but at the same time, I can assure you all I'm doing is loading the same project into two different versions without making any changes to the code. I'm current talking to Tom in an email conversation, so ill liasse with Tom/Support to get a copy to somebody on your team to try it out, and I'll make a bug report with the details.

    I have found a weird work around that if I change the speed of physics calculations to match the fps of the game, it works normally, but that single event uses up some amount of CPU.

Jump to:
Active Users
There are 1 visitors browsing this topic (0 users and 1 guests)