(WIP)Test - Tyre Grip Tuning for the 500cc Varase....

[Hawk]

But If I personally was to put one thing as an absolute priority for fixing, it would be to stabilise GP Bikes while riding online from the "core.exe" crashes people experience. But that is a difficult one as I believe most "Core.exe's" come from poor circuit/track meshes(99.9% of which come from ripped circuits that are poorly cleaned up after the ripping process). I think people who are blaming GP Bikes for these core.exe's are probably doing Piboso and GP BIkes a huge injustice.

I fully agree with your priorities Hawk, but I wouldn't go as far as saying the "core.exe" problem comes from poor tracks.

Sure, a poor track doesn't help, but we know for sure it's not the only explanation. If it was, Piboso would surely had already told us so I hope.

Also, as a general rule, software should never ever crash in an unspecified manner. Admittedly, for stuff like GPB this is hard to enforce, but ...

MaX.

Hi Max.

Well.... After being able to create a core.exe at will when I identify artefacts in track meshes, and it can happen with the smallest of gaps in a mesh that unless you were zoomed right in on it you just wouldn't believe there was a gap there. I personally will be convinced that it is 99.9% track problems and 00.1% maybe a GP Bikes problem.
To my mind, we should try and sort out what we have the facts for first and then see if the core.exe problem still occurs, as I for one have not seen or heard of any other repeatable or demonstrable evidence that causes a core.exe.

I remember posting about this before for Piboso, and suggesting an error catching routine to allow GP Bikes to continue upon this error, but didn't get any reply. Surely this is possible, isn't it?


Hawk.

[HornetMaX]

Well, I've witnessed core.exe in the pits, in the garage (i.e. before going on track), exiting a race, loading a track (especially after having played on another track): these can't come from track issues, at least not from the kind you're talking about (gaps).

When you posted about the detection of gaps I said it was not sure Piboso could do it because he may be using ODE for collision detection. He only replied that he's using his own dedicated model for tire-track collision handling (which makes my comment irrelevant), but didn't really reply to your question.

MaX.

P.S
Talking about issues: in the italian championship we've seen many times (not always) the server skipping the detection of a lap (for no good reason), ending up in messing the lap count for the player (which by the ways confuses GPB and at the end of the race the player with skipped lap sees a white flag when he shouldn't, or two white flags, ot two checker flags).

And there's also the problem of the server sometimes detecting laps very late (like 30 secs after you cross the finish line).

And the very annoying thing of server going mad at race start and some bikes just jump starts with no input from their riders (I've seen this first person: I'm 100% sure I didn't apply any input, and my bike moved forward, along with 2 other bikes in the very same situation).

[Hawk]

Well, I've witnessed core.exe in the pits, in the garage (i.e. before going on track), exiting a race, loading a track (especially after having played on another track): these can't come from track issues, at least not from the kind you're talking about (gaps).

It is possible that preparation of the mesh models for rendering and or processing of the mesh model data is done as part of the connection process, or while your in the garage/pits(in the background) so that as soon as you click the "Go to Track" button most of the mesh transformations are complete, as a result if there are any split polygons or artefacts these can cause program crashes even before you enter the track; it is quite a real possibility.
I had an instance working on Monza were the spectator stands were a total mess with split polygons all over the place, GP Bikes would crash each time either before or soon after entering the pits/garage until I cleaned up the spectator stand mesh, then it worked fine(Took me a long time to track down the cause of this problem.). It would be interesting if Piboso could confirm any terrain track mesh model processing that might be going on in the background before the user actually enters the virtual circuit itself.

When you posted about the detection of gaps I said it was not sure Piboso could do it because he may be using ODE for collision detection. He only replied that he's using his own dedicated model for tire-track collision handling (which makes my comment irrelevant), but didn't really reply to your question.

MaX.

Yes, I remember this. But as you say, Piboso never actually reply'd to my question, only reply'd to your comment. He never does reply to me for some reason, never has and I'm not expecting it now either. I can't say it maybe because he has never seen my posts because with the evidence of your post above he obviously has and does, well at least some of them. I think I must unintentionally just rub him up the wrong way. But despite many asking questions he only seems to answer certain people or total newbies having probs with getting started, you just happen to be one of lucky few.  Is he French? That would explain a lot(No offence to the French of course. Love your cheese! ). Probably put my foot in my mouth again now. Hehe. 

P.S
Talking about issues: in the italian championship we've seen many times (not always) the server skipping the detection of a lap (for no good reason), ending up in messing the lap count for the player (which by the ways confuses GPB and at the end of the race the player with skipped lap sees a white flag when he shouldn't, or two white flags, ot two checker flags).

And there's also the problem of the server sometimes detecting laps very late (like 30 secs after you cross the finish line).

And the very annoying thing of server going mad at race start and some bikes just jump starts with no input from their riders (I've seen this first person: I'm 100% sure I didn't apply any input, and my bike moved forward, along with 2 other bikes in the very same situation).

All the above, and I'm certainly no expert in this, this is just me reasoning this out off the top of my head(so to speak), but all the above could be down to poor telecommunications connection(temp not in sync or incomplete data received) in combination with poor network coding to deal with the bad data situation, or one or either of both.

Most jump starts seem to be caused by riders parking their bikes right against or too near the forward white line of their start box; I'm not saying that is definitely the case with your examples as I haven't seen the replays, but in my experience it is always best to park your bike at least half way down your start box to avoid a jump start. Since I started doing this I haven't had a jump start at all; after all, your not going to lose a race because of a couple of feet difference from the start. But especially when you get a lot of riders on the grid there does seem to be an inclination for lag from the server to the client or vice versa.

As I said, I'm no expert at all in network/server coding, so take what I said as a pinch of salt(so to speak). 


Hawk.

[HornetMaX]

Yes, I remember this. But as you say, Piboso never actually reply'd to my question, only reply'd to your comment. He never does reply to me for some reason, never has and I'm not expecting it now either. I can't say it maybe because he has never seen my posts because with the evidence of your post above he obviously has and does, well at least some of them. I think I must unintentionally just rub him up the wrong way. But despite many asking questions he only seems to answer certain people or total newbies having probs with getting started, you just happen to be one of lucky few.  Is he French? That would explain a lot(No offence to the French of course. Love your cheese! ). Probably put my foot in my mouth again now. Hehe. 

I've never really understood how he selects what to reply to. I have a large share of totally ignored posts, many simply asking for useful info that would probably take him 10sec to provide.
It seems pretty random, I just live with that.

MaX.

P.S.
Just noticed another weird thing on monzabikes v1.1: sometimes the last split is taken 1.5-2 sec before the correct position, ending up in a very nice 3rd sector and a very bad 4th sector when in fact you just did a lap like many others.

[rodney007]

see at 0:57

https://www.youtube.com/watch?v=7X08zELY-lo

This is a combination of slightly standing the bike up to give more grip and allow for more throttle.

I have had a read through this thread and am guessing this is what you guys are trying to achieve?

If so I think I have got it... I have only changed the rear suspension settings in the .cfg

[Hawk]

see at 0:57

https://www.youtube.com/watch?v=7X08zELY-lo

This is a combination of slightly standing the bike up to give more grip and allow for more throttle.

I have had a read through this thread and am guessing this is what you guys are trying to achieve?

If so I think I have got it... I have only changed the rear suspension settings in the .cfg

Hi Rodney.

We were discussing trying to achieve a realistic power slide through a corner(though the topic did slip occasionally ); but I cannot see a power slide @ 0:57 in your video above? Is this what you guessed we were talking about here?

BTW... Very nice video mate! 

Hawk.

[rodney007]

Update: Saturday 14th June 2014 @ 00:36 hrs

that the 500cc GP 2 stroke bikes should be able to rear wheel power steer around the corners using the power of the engine through good throttle control.

First post*