anti wall slide

Evil Inside · Post by **Evil Inside** » Wed Nov 03, 2004 11:35 pm

first i would like to say congratulations on fixing the chat exploit

well done everyone involved

second as I promised heres the simple mod that virtually removes all lag slide from the client code

file: gCycle.cpp
function: bool gCycle::TimestepCore(REAL currentTime){

remove or comment out these 5 lines

clamp(ts, -10, 10);
clamp(correctTimeSmooth, -100, 100);
clamp(correctDistSmooth, -100, 100);
clamp(correctSpeedSmooth, -100, 100);
clamp(speed, -1000, 1000);

Zapple · Post by **Zapple** » Thu Nov 04, 2004 12:08 am

Watch it be some other hack

just kidding

ishAdmin · Post by **ishAdmin** » Thu Nov 04, 2004 2:08 am

Welcome to the forum Mr. Inside. We appreciate you passing on this information.

I personally think you would be a good addition to the development team. I hope it's something you will consider.

Evil Inside · Post by **Evil Inside** » Thu Nov 04, 2004 3:58 am

i would be honnered to join your development team

philippeqc · Post by **philippeqc** » Thu Nov 04, 2004 10:12 am

Zapple wrote:Watch it be some other hack :lol:

just kidding

Not kidding about the following:

It looks ok to me. But there might be some unknow side effects. Because you know many ppl that are good to find bugs Zapple, you could be of great help here. Just send them the patch and ask them to try it. Then tell us if they find some new problems.

That would really help.

-ph

Post by **Tank Program** » Thu Nov 04, 2004 10:42 am

Well you see... If my theory is correct, this could be very bad... AFAIK lag slides occur when your cycle position on your client does not match up the position the server you're on has for it. So, your client gets updated data and slides you to the right place. Removing that code that does the sliding, fixing your position, leaves you displayed in the wrong place, free to hit walls you can't see...

Just a theory, but you know...

philippeqc · Post by **philippeqc** » Thu Nov 04, 2004 10:54 am

Tank Program wrote:Well you see... If my theory is correct, this could be very bad... AFAIK lag slides occur when your cycle position on your client does not match up the position the server you're on has for it. So, your client gets updated data and slides you to the right place. Removing that code that does the sliding, fixing your position, leaves you displayed in the wrong place, free to hit walls you can't see...

Just a theory, but you know...

Humm, yeah, something like that. I was about to comment that removing the sliding is not fixing the problem. If sliding was introduced, it was to address an issue. Fix that issue, and then we can remove sliding.

-ph

Post by **Tank Program** » Thu Nov 04, 2004 10:58 am

To fix the issue... We'd have to fix UDP packets from ever getting lost...

philippeqc · Post by **philippeqc** » Thu Nov 04, 2004 11:09 am

Tank Program wrote:To fix the issue... We'd have to fix UDP packets from ever getting lost...

No, just come come up with:
a) a good way to correct the game state when it gets re-sync.
b) better prediction (for me, sliding is when I turn and the server dissagree at the time when I did turn, and the local client has to change its state to reflect the server state. So either:
i) The server calculated some extra influence and I was moving faster on the server than on the client.
ii) The moment of turn is not decided by the time of the operation on the client, but when the server receive it. So its very susseptible to "lag" probs and need to be addressed.

You see, nothing that someone dedicated with a couple of free days(weeks) on his hands couldnt fix.

-ph

Evil Inside · Post by **Evil Inside** » Thu Nov 04, 2004 12:13 pm

no doubt they is a better way to fix this problem
this is just a quick fix and seems to work without any side effects

the cycle snaps into the correct position within about 1 frame with minimal slide instead of sliding slowly over several frames and you dont get the effect of sliding away from the correct position and then back again

please try it and give some feedback on what you think

Post by **Tank Program** » Thu Nov 04, 2004 2:34 pm

The snapping might be more annoying- but you didn't mention that before...

Evil Inside · Post by **Evil Inside** » Thu Nov 04, 2004 2:38 pm

try it tank i think you will like it

CharlieH · Post by **CharlieH** » Thu Nov 04, 2004 3:31 pm

the fact that people would invite someone who exploited a bug to spam and lag to their development team is absolutely rediculous. you are simply working from the fact that he knows at least some C++. he seems like a child and a dick in my eyes. by choosing to make the bug known by example for hours per day instead of simply posting it here in the forums, he shows his immaturity. just my opinion.

/dev/null · Post by **/dev/null** » Thu Nov 04, 2004 10:11 pm

They didnt invite him, ish said something, but im with you, someone who would blatantly exploit a bug to be an asshole instead of quickly releasing a fix for it should not be allowed to develop this project, the likelyhood they would slip in a backdoor or something is very high, and for any server admins watching, banning this fool would be a good idea.

ishAdmin · Post by **ishAdmin** » Thu Nov 04, 2004 11:56 pm

Hey guys, bridges are better built than burnt. If you have a problem with your neighbour, invite to a barbecue, and make friends. A friend won't complain about every little thing. Do you see where I'm going with this ...