[RVGL] Direction triggers bug

Report issues and bugs here.
User avatar
Kiwi
Posts: 473
Joined: Thu Dec 07, 2017 9:04 pm
Location: Austria

[RVGL] Direction triggers bug

Unread post by Kiwi » Tue Feb 04, 2020 7:11 am

During a multiplayer race in yesterdays Super-Pro session, Floxit had the issues, that directions triggers was not showing anymore during lap one at School's Out! (R).

After analysing, I am pretty sure he passed all direction triggers as intended.

The error happened between direction trigger 6 and 7.

Image

Floxit passed trigger 6 normally, and trigger 7 didn't showed up correctly. All the following triggers (8, 9, 10, 11,...) didn't showed up then also.

Here is the video where it happened (Timestamp 58m 3s): https://youtu.be/Cc2Qto5zHZc?t=58m3s

Any idea, who this can happen? Triggers 6 and 7 definitly reach the sides, floor, and ceiling of the track. According to the video, Floxit have definitly passed it. Can it have anything to do with Vsync, FPS, multiplayer lag, or even the "thickness" of the triggers?
User avatar
r6te
Posts: 31
Joined: Mon Dec 11, 2017 3:59 pm

Re: [RVGL] Direction triggers bug

Unread post by r6te » Tue Feb 04, 2020 12:06 pm

Hi Kiwi,

First of all, you've made an amazing looking track! I'm proud :D

What I think what happened is that Floxit passed the trigger in between the calculation of the frames. I assume at the start of every frame or physics frame it will see whether the car is inside a trigger field, if yes, display trigger. Making this field thicker increases the chance of a car being in it at the start of a frame. But I'm not sure though.
User avatar
Kiwi
Posts: 473
Joined: Thu Dec 07, 2017 9:04 pm
Location: Austria

Re: [RVGL] Direction triggers bug

Unread post by Kiwi » Tue Feb 04, 2020 1:52 pm

r6te wrote:
Tue Feb 04, 2020 12:06 pm
Hi Kiwi,

First of all, you've made an amazing looking track! I'm proud :D

What I think what happened is that Floxit passed the trigger in between the calculation of the frames. I assume at the start of every frame or physics frame it will see whether the car is inside a trigger field, if yes, display trigger. Making this field thicker increases the chance of a car being in it at the start of a frame. But I'm not sure though.
Hey Thomas, thanks a lot! :)

Hmm...this sounds plausible. Let's wait for Huki then. If this is really true, then I will have to release an update for the track, with thicker direction-triggers.
User avatar
607
Posts: 266
Joined: Thu Dec 07, 2017 5:49 pm

Re: [RVGL] Direction triggers bug

Unread post by 607 » Tue Feb 04, 2020 3:03 pm

Why anyway do subsequent triggers not appear when one is missed? This has probably been in the game since the beginning (it's why I could never manage to learn PetroVolt), but it doesn't make sense to me. I've never actually placed triggers in a track, though.
User avatar
hajducsekb
Posts: 28
Joined: Mon Feb 11, 2019 8:13 pm
Location: ~/
Contact:

Re: [RVGL] Direction triggers bug

Unread post by hajducsekb » Tue Feb 04, 2020 10:47 pm

Hey!

Managed to test this on different framerates, and R6TE was correct. On lower fps, there's a higher chance that you miss a trigger because of this, I believe. From my tests, it seems like the game's physics refresh rate is the same as the fps number. My video shows it well: Re: 607
One reason I can instantly think of is the fact that crossroads exist. There, the game wouldn't know which direction arrow to show, so it's much more hassle-free to have to go through each trigger than to somehow decide which one is the right one.
WORDART IS COOL
User avatar
Kiwi
Posts: 473
Joined: Thu Dec 07, 2017 9:04 pm
Location: Austria

Re: [RVGL] Direction triggers bug

Unread post by Kiwi » Wed Feb 05, 2020 5:55 am

Thanks a lot for testing, and doing this really nice and helpful video! I‘m shocked how often (and early) the problem occurs at a 20 fps framerate.

OK, it’s clear now, I have to make the direction arrow triggers thicker, and release another bugfix release then. :(

But to be 100% sure, I will wait for an „official“ answer by Huki as well, before I do this.
User avatar
r6te
Posts: 31
Joined: Mon Dec 11, 2017 3:59 pm

Re: [RVGL] Direction triggers bug

Unread post by r6te » Wed Feb 05, 2020 12:29 pm

hajducsekb wrote:
Tue Feb 04, 2020 10:47 pm
Hey!

Managed to test this on different framerates, and R6TE was correct. On lower fps, there's a higher chance that you miss a trigger because of this, I believe. From my tests, it seems like the game's physics refresh rate is the same as the fps number. My video shows it well: Re: 607
One reason I can instantly think of is the fact that crossroads exist. There, the game wouldn't know which direction arrow to show, so it's much more hassle-free to have to go through each trigger than to somehow decide which one is the right one.
Ah, thanks for the video. Great explanation :)

I do not necessarily think it's the physics that is running on a lower refresh rate (iirc it's running on 60+Hz, maybe even 120Hz), but rather the checking of the car's location within the trigger is running on the set refresh rate.
User avatar
607
Posts: 266
Joined: Thu Dec 07, 2017 5:49 pm

Re: [RVGL] Direction triggers bug

Unread post by 607 » Wed Feb 05, 2020 6:05 pm

hajducsekb wrote:
Tue Feb 04, 2020 10:47 pm
Re: 607
One reason I can instantly think of is the fact that crossroads exist. There, the game wouldn't know which direction arrow to show, so it's much more hassle-free to have to go through each trigger than to somehow decide which one is the right one.
Ah, that makes sense! Thanks.
User avatar
Huki
Developer
Posts: 279
Joined: Wed Dec 06, 2017 5:54 am
Discord: Huki#1668

Re: [RVGL] Direction triggers bug

Unread post by Huki » Thu Feb 06, 2020 2:04 pm

Kiwi wrote:
Wed Feb 05, 2020 5:55 am
Thanks a lot for testing, and doing this really nice and helpful video! I‘m shocked how often (and early) the problem occurs at a 20 fps framerate.

OK, it’s clear now, I have to make the direction arrow triggers thicker, and release another bugfix release then. :(

But to be 100% sure, I will wait for an „official“ answer by Huki as well, before I do this.
I can confirm this. Triggers are checked every frame. If the trigger box is too thin, the car could pass through it undetected.
Post Reply