[QUOTE=Grayman;43743346]Fixed an out of memory crash. Should fix maps with lots of ragdolls.
[url=https://dl.dropboxusercontent.com/u/5619434/Development/muyphysics.zip]Download[/url][/QUOTE]
So, what is the plus for us downloading this now. What does it add for us, and how does it help you?
[QUOTE=pilot;43744809]So, what is the plus for us downloading this now. What does it add for us, and how does it help you?[/QUOTE]
You literally quoted the answer to your question. It fixes a crash on some maps.
[QUOTE=Grayman;43747971]You literally quoted the answer to your question. It fixes a crash on some maps.[/QUOTE]
Sorry I really phrased that wrong ;-; . I meant downloading this whole project in general. What does it add for us, and how does it help you.
That's still a pretty dumb question but.
He's only putting up download links and keep us up to date with this because it gives us a chance to give him input and suggestions, along with just playing around with this our selves.
[QUOTE=pilot;43750102]Sorry I really phrased that wrong ;-; . I meant downloading this whole project in general. What does it add for us, and how does it help you.[/QUOTE]
This project aims to replace valve's closed source physics engine with an open source physics engine that aims to be better than valve's.
Since valve's physics is closed source, even to licensees, developers can't correct any of the problems it has (like a bunch of crashes, at one point jeeps on linux caused a crash and it wasn't fixed for months I believe, and atm ragdolls sometimes crash ttt servers).
In addition, they can't add any new physics features, they can only use the ones supported by valve's physics.
Also, this project keeps me from being bored all the time since I have something to do.
[QUOTE=Grayman;43751420]This project aims to replace valve's closed source physics engine with an open source physics engine that aims to be better than valve's.
Since valve's physics is closed source, even to licensees, developers can't correct any of the problems it has (like a bunch of crashes, at one point jeeps on linux caused a crash and it wasn't fixed for months I believe, and atm ragdolls sometimes crash ttt servers).
In addition, they can't add any new physics features, they can only use the ones supported by valve's physics.
Also, this project keeps me from being bored all the time since I have something to do.[/QUOTE]
You forgot to add in the fact that this should be far more optimized, since bullet physics can use multiple cores
[QUOTE=da space core;43752179]You forgot to add in the fact that this should be far more optimized, since bullet physics can use multiple cores[/QUOTE]
Yeah, that's mostly true, but it's going to be me who will parallelize bullet (bullet's threading support is total crap).
How's it like on Linux at the moment? Might give it a try on Linux if it's stable enough.
[QUOTE=Bo98;43752401]How's it like on Linux at the moment? Might give it a try on Linux if it's stable enough.[/QUOTE]
Pretty much the same as windows. Also, it looks like that class tables thing was fixed.
Anyways, if you can, see if you can build it yourself. Should be simple enough to set up (read about the submodules on the github readme, and configure src/makefile) on linux.
Also, if you're compiling on a 64 bit machine, you'll have to setup g++ for 32 bit compiling (apt-get all of the 32 bit libraries needed for building)
I don't want to provide linux builds because I have to log into my virtualbox linux machine every time, and it takes a while.
Alright, I'll give it a try.
[QUOTE=Bo98;43752734]Alright, I'll give it a try.[/QUOTE]
Did you get it working?
Sorry, not had the chance today (UK time). Will definitely get on it tomorrow.
[QUOTE=Bo98;43763644]Sorry, not had the chance today (UK time). Will definitely get on it tomorrow.[/QUOTE]
Alrighty, well if you have any problems be sure to detail them here, because I'm sure others may want to know how to fix any problems with linux builds.
[QUOTE=Grayman;43751420]This project aims to replace valve's closed source physics engine with an open source physics engine that aims to be better than valve's.
Since valve's physics is closed source, even to licensees, developers can't correct any of the problems it has (like a bunch of crashes, at one point jeeps on linux caused a crash and it wasn't fixed for months[/QUOTE]
It was never fixed instead Garry did a workaround by making jeeps unfreezable. Forcing you to weld the jeep to the world oh wait.
[QUOTE=Sally;43766138]It was never fixed instead Garry did a workaround by making jeeps unfreezable. Forcing you to weld the jeep to the world oh wait.[/QUOTE]
Really? Huh, I thought that was a problem with my vphysics.
[QUOTE=Sally;43766138]It was never fixed instead Garry did a workaround by making jeeps unfreezable. Forcing you to weld the jeep to the world oh wait.[/QUOTE]
Sounds like a Garry-solution alright.
[QUOTE=Bo98;43769320]Sounds like a Garry-solution alright.[/QUOTE]
Not much he could do anyways. This project didn't exist then and he couldn't fix it himself.
Anyways, for an update on vphysics (which I'm sure a lot of people are hungry for), I'm working on profiling some slow code, and I'm testing a bunch of maps to fix crashes or put the bugs on the issue tracker for me to look at later.
[QUOTE=Grayman;43770781]Not much he could do anyways. This project didn't exist then and he couldn't fix it himself.
Anyways, for an update on vphysics (which I'm sure a lot of people are hungry for), I'm working on profiling some slow code, and I'm testing a bunch of maps to fix crashes or put the bugs on the issue tracker for me to look at later.[/QUOTE]
I just wanted to thank you for doing this. Do you think tank tracks could be possible with the success of your project? If so, I will donate $20.00 :)
[QUOTE=Skyfire12;43779538]I just wanted to thank you for doing this. Do you think tank tracks could be possible with the success of your project? If so, I will donate $20.00 :)[/QUOTE]
I suppose, depending on how you want them. Could be a bunch of raycast "wheels" along the track or just a bunch of constrained props (which probably would really lag).
Once I get the custom vehicle controller setup and bind it to gmod lua it should be possible.
I think he means more for contraption building
[QUOTE=A_Pigeon;43779860]I think he means more for contraption building[/QUOTE]
Well, they're already possible with valve's vphysics. Weld a bunch of props or you could use a ragdoll model for tracks.
A tank vehicle will certainly be possible under my vphysics by using the custom vehicle controller.
Anyways, any feature requests? Be sure to check that they aren't already possible in valve's vphysics.
[QUOTE=Grayman;43784059] any feature requests? Be sure to check that they aren't already possible in valve's vphysics.[/QUOTE]
An actual working welding tool and a non-spazztastic slider tool
One of the things I hate the most is that hydraulics spazz out in zero gravity. I think that was the single change that started the death of the spacebuild community.
Also, that physics mesh merging thing that bugged due to clientside prediction errors would be nice, but that's not really something the physics can change.
If you somehow were able to rework the constraint system in a way such that welds didn't sag nearly as severely and sliders actually worked under normal circumstances, you'd gain the admiration of a fair deal of people.
[QUOTE=da space core;43787418]An actual working welding tool and a non-spazztastic slider tool[/QUOTE]
Except for that, which was explained earlier in the thread that the constraint solver can't do stuff like that. Constraints are solved by setting up a formula for the solver, and the solver solves it using forces and whatever else.
Although, it looks like bullet constraints are pretty rock solid. From what it looks like, large differences in mass don't affect the constraint's stability. Same with other constraints, for that matter. I think the only weakness would be in chaining constraints.
Try it yourself: weld 4 barrels together in a square, and then weld something really heavy on top. Compare the results between my vphysics and valve's vphysics. If someone wants to make a video of it, go ahead.
Here's a video of what I was talking about:
[video=youtube;1thX2ucj_-w]http://www.youtube.com/watch?v=1thX2ucj_-w[/video]
Although it's a bit inconsistent with bullet. The barrels have to be the reference objects for these rock solid constraints.
I tested it last night, except some strange bugs, props eaten by the map an performance problems, that's a nice job !
Holy crap, that's already 100x better than valve physics
[QUOTE=ExtReMLapin;43797271][IMG]http://extrem-team.com/test4.gif[/IMG]
[IMG]http://extrem-team.com/test5.gif[/IMG]
:v:[/QUOTE]
Just a few of the amazing features vphysics has over vphysics!
Sorry, you need to Log In to post a reply to this thread.
