Forums
Regressions - Printable Version

+- Forums (https://www.vdrift.net/Forum)
+-- Forum: Project (https://www.vdrift.net/Forum/forumdisplay.php?fid=4)
+--- Forum: Development (https://www.vdrift.net/Forum/forumdisplay.php?fid=9)
+--- Thread: Regressions (/showthread.php?tid=1427)

Pages: 1 2


Regressions - NaN - 01-20-2011

It's been like a year since I've started messing with the wheel simulation to advance its accuracy. Unfortunately the changes turned out to be very fragile and caused serious regressions(rollover, curbs, bumps) which I haven't been able to resolve completely yet. Running the trunk unstable for the whole time has been rather ignorant from my side. I'd like to apologize for this.

I will revert the relevant code to the last year state. The car behavior should match(be close to) the 2009 release.

I will revert some parts of the GUI too to fix the draw order regression introduced by me, will try to keep the translation support though.

I encourage everyone to post regressions not listed above so that I can address them in a timely manner, just for the case there is still a desire to push a release soon.


- thelusiv - 01-20-2011

I think this is a good idea, but I think you should keep working on what you were working on in a branch, unless you just want to start over or think it's not worth trying again. Take some time, I have gotten incredibly busy and haven't had time to work on release in the last few weeks (this has probably been obvious to everyone, and sorry for not posting something earlier).


- fudje - 01-20-2011

Confusedhock: No I think that's a bad idea Confusedhock:


- portets - 01-20-2011

:?

Sad


- charlieg - 01-21-2011

My 0.02c here:
http://vdrift.net/Forum/viewtopic.php?p=12058#12058


- NaN - 01-21-2011

Quote:I will revert some parts of the GUI too to fix the draw order regression introduced by me
The GUI code looks OK. The text and the 2D GUI elements are drawn in separate passes due to different shaders used. This also means we can not sort them before drawing. The text being drawn after the 2D elements will always be on top of them. Not sure how to fix this for the in-game GUI.


f1-02 lost power - alex25 - 01-21-2011

r3022. on the straight after l'eau rouge at spa the f1 reaches only 130 mph (previously i could get it to 190+). also, if i get stuck in the sand traps at suzuka, i can't get out. this is probably related to the recent changes/reverts. any ideas?

--alex--


- NaN - 01-21-2011

I've kept the new rolling resistance code. It calculates rolling resistance proportional to tire normal force. I'll review the code and tire configs asap.


- joevenzon - 01-21-2011

NaN Wrote:The text and the 2D GUI elements are drawn in separate passes due to different shaders used. This also means we can not sort them before drawing. The text being drawn after the 2D elements will always be on top of them. Not sure how to fix this for the in-game GUI.

Indeed, this has been a problem for a long time. I can add another pass if we need to layer images on top of text, or come up with another solution if we need to do a lot of image/text overlapping. Where are you currently seeing a problem? The pause menu is the only one I'm aware of.


- NaN - 01-22-2011

Quote:this is probably related to the recent changes/reverts. any ideas
The new rolling resistance implementation would need track surface parameter adjustments. I'll revert it.

Quote:I can add another pass if we need to layer images on top of text, or come up with another solution if we need to do a lot of image/text overlapping. Where are you currently seeing a problem? The pause menu is the only one I'm aware of.
It's only the pause menu. It's just that haven't noticed it before.


- charlieg - 01-22-2011

NaN Wrote:The new rolling resistance implementation would need track surface parameter adjustments. I'll revert it.
So your plan is to just back out everything you have done? That's not much of a plan. Why not compile a list of 1. bugs to fix and 2. changes that need to be made to tracks. Then see if there's a middle ground rather than running away from the problem.


- NaN - 01-22-2011

Thanks for your input charlieg.

I am aware of the concept of a version control system. The use of the trunk as the main development branch was intentional, to speed up development by having a second pair of eyes onto changes and getting immediate feedback. It worked quite well for some parts but totally backfired for the simulation code.

This is not about "running away", it's about people asking for a stable trunk for some time now. The safest way to achieve this is to go back to the last known stable state.


- alex25 - 01-22-2011

NaN Wrote:
Quote:this is probably related to the recent changes/reverts. any ideas
The new rolling resistance implementation would need track surface parameter adjustments. I'll revert it.
there is no rolling resistance on the track itself so it's probably something else as well.

--alex--


- charlieg - 01-22-2011

NaN Wrote:Thanks for your input charlieg.

I am aware of the concept of a version control system.

...

This is not about "running away", it's about people asking for a stable trunk for some time now. The safest way to achieve this is to go back to the last known stable state.

So where's the branch then?

Or are you really jacking it all in?


- joevenzon - 01-23-2011

charlieg Wrote:Or are you really jacking it all in?

What's that mean?