[Crayder]

Quote:

Originally Posted by DRIFT_HUNTER I actually think that people wont share files because there is no need to...It will take much longer to download those files than to extract them...

Not exactly, my extractor takes ~2 minutes to get through the files and delete those not in the array of files (i made an array of the files pottus listed).

I could downloads 4 gbs in less than 10 seconds, it merely depends on the uploaders speed.

I use fiber optics, if you don't know the capability of light speed internet do some research :P.

[DRIFT_HUNTER]

Quote:

Originally Posted by Crayder Not exactly, my extractor takes ~2 minutes to get through the files and delete those not in the array of files (i made an array of the files pottus listed). I could downloads 4 gbs in less than 10 seconds, it merely depends on the uploaders speed. I use fiber optics, if you don't know the capability of light speed internet do some research :P.

Well if you have optic links than you are the lucky one...

Anyway, you dont need to extract all files, just a few that you need.

[Crayder]

Quote:

Originally Posted by DRIFT_HUNTER Well if you have optic links than you are the lucky one... Anyway, you dont need to extract all files, just a few that you need.

My method requires first extracting all the files to the extraction folder, then deleting the ones that are not in the array i described above, the slow part is searching the imgs and taking out each file.

[Arastair]

That's too great

[Mauzen]

Quote:

Originally Posted by Crayder My method requires first extracting all the files to the extraction folder, then deleting the ones that are not in the array i described above, the slow part is searching the imgs and taking out each file.

Thats a bit over the top then. My batch method just takes 2-3 seconds for everything. Isnt there a way to extract single files only from the img with your code?

[AroseKhanNiazi]

i didn't really get what this does ?

[Crayder]

Quote:

Originally Posted by AroseKhanNaizi i didn't really get what this does ?

Basically, mapandreas for every direction. Or, 3D mapandreas. To be even more basic, mapandreas times 3. Or the most basic thing to call it, the best plugin of 2015.

Ok, so mapandreas, you input x and y, you get the highest z point available. ColAndreas, you input to 3D space points, it returns the point where the line ends. Watch the video, I'm pretty sure the two points in space Pottus used were the camera position and the front vectors. The virtual 'ray' was casted along the line, it returned where it ended in the world. Now, imagine the possibilities.

[Admigo]

Looks amazing. I will try this out for sure. Thanks Pottus and Clan.

[vermaritt]

In my opinion:

• High number of object will slow down the processing for each object if the processing is not multithreaded or parallel computed.

• This may happen if pottus unlocks JernejL's trick of extraction of object data.

• Same reason, my first point.

[Pottus]

1.) It really doesn't really matter Bullet uses an very efficient method to do this.

2.) The collision binary would have to be rebuilt this is automatic and not an issue. Reading the data directly is possible however there needs to be work done before that data can be used. Primarily the work is creating the lists of collision object positions. Given that it takes time to create the collision objects in Bullet building the binary will save a lot time.

3.) 6000x6000 is only 36000000 raycasts the amount of objects does not matter this would take approximately 4 minutes.

It really depends on the server ColAndreas is much slower on Windows for me but our dedicated server can do 160000 raycasts per second.

[vermaritt]

To Pottus, on what ram and processor your computer is running ?

[Pottus]

My computer ColAndreas runs significantly slower, our dedicated linux server it runs much faster keep in mind my computer is getting older now too it's a first generation i5 (Lynnfield) 2.4ghz

[Crayder]

Quote:

Originally Posted by Pottus My computer ColAndreas runs significantly slower, our dedicated linux server it runs much faster keep in mind my computer is getting older now too it's a first generation i5 (Lynnfield) 2.4ghz

My home computer is also i5 2.4ghz. But the times still sounds great since I only need to update it every once in a while. I plan on making a height map that will make the highest point of my map alone red, lowest agree, and everything in between will be gradient to yellow from above and below. I also plan on making a map like the default SA map, but for this I need to make a list of which modelids will make what color on the generated map (this won't be hard) and a way to return the model of the collision detected (CA_RayCastLine returns the modelid right?).

[Pottus]

You can certainly generate custom height maps very easily.

[n0minal]

Chris finished the extractor tool?

[Crayder]

Quote:

Originally Posted by n0minal Chris finished the extractor tool?

This^ and not begging, but at least some progress report please? I've been waiting for this for my upcoming project...

[vermaritt]

So, it will not support physics simulation ?

[Pottus]

The problem is stepping the simulation I was playing around with this but the issue is every time you step the simulation it takes time to complete and actually freezes the SA-MP server I am sure there is a way to fix this problem however.

The simulation also needs to return the results instantly to create a set of frames that can be played with MoveObject() (This won't work properly with player objects). The alternative would be to modify the physics include to use raycasting. A big problem is trying to sync a physics simulation to MoveObject() that isn't going to be pretty and at best shoddy so you would need to calculate your frames before actually playing them.

[vermaritt]

Well, for a while just ignore all physics simulations rather than collision simulation. It is possible via two ways:

First, by using bullet physics.(It will create a major problem as we cannot rebuild the game again with bullet)

Second, by using pawn. Probably making an updater and then casting a ray from the center of a moving object to it velocity updated position, if the ray detects collision between the points then we can determine the point of collision and thus we can resolve the collision.

[Crayder]

Quote:

Originally Posted by vermaritt Probably making an updater and then casting a ray from the center of a moving object to it velocity updated position, if the ray detects collision between the points then we can determine the point of collision and thus we can resolve the collision.

Actually you would not just be able to do it with one ray, you'd have to use like one for each axis of the object. Not every surface is level