Roland

And that means? Hey I'm just the C++ guy here Groping 4 walls into one drawcall ... Sorry to chock you all with my total lack of knowledge about drawcall, or it may be so that I just don't know what the term means. 1. 4 separate meshes in same model file 2. all as one mesh 3. or is it meshes with same materials Now you really caught me with the trousers down

So ... conclusion would be that method 1 is appropriate for occlusion reasons and/or if the part have thousands of polys. Other than that, one can safely go for the 'all in one mesh' method 2. Did I get that right?

Example: Building a room with 4 plain walls (no doors, no windows), say each wall needs 100 polys. What is better? 1) Make one wall model and copy that to 3 new entity's. This will give 100 polys and 4 entities 2) Make one model with all 4 walls. This will give 400 polys and 1 entity

Gravity works Y is decreasing

And that's a good choice in my humble opinion

Hard to say besides LE3 run on Android. Anyway. Great that you are happy

Besides the to be or not to be discussion. Did my example solve your problem Josh?

Using STL is quite standard and would be a positive thing. There are many advantages with STL iterators, as the ability to have several iterators operation on the same list at once, it also have a bunch of nice features that I would miss. So my recommendation is STL, but Josh has the internal overview of the code and may have some reason to use non-standard lists.

But why... can't see the genius in doing that! But its your engine, not mine However, think it will be something like this... Have not tested, just written. Link.h #pragma once namespace Leadwerks3D { template<typename T> class LinkedList; template <typename T> class Link { friend class LinkedList<T>; T* value; Link* prev; Link* next; LinkedList<T>* list; public: Link() : value(nullptr), prev(nullptr), next(nullptr), list(nullptr) {} ~Link() { if (list->firstlink==this) list->firstlink=next; if (list->lastlink==this) list->lastlink=prev; if (prev) prev->next=next; if (next) next->prev=prev; value=nullptr; } Link* NextLink() const { return next; } Link* PrevLink() const { return prev; } T* GetValue() const { return value; } }; } LinkedList.h #include "Link.h" namespace Leadwerks3D { template <typename> class Link; template <typename T> class LinkedList { friend class Link<T>; Link<T>* firstlink; Link<T>* lastlink; public: LinkedList() : firstlink(nullptr), lastlink(nullptr) {} ~LinkedList() { Link<T>* link = firstlink; Link<T>* nextlink; while (link!=nullptr) { nextlink=link->next; link->next=nullptr; link->prev=nullptr; delete link; } } Link<T>* GetFirstLink() const { return firstlink; } Link<T>* GetLastLink() const { return lastlink; } Link<T>* AddLast(T* o) { Link<T>* link = new Link<T>; link->value = o ; link->prev = lastlink ; link->list = this; if (lastlink) { lastlink->next=link; } else { firstlink=link; } lastlink = link; return link; } Link<T>* AddFirst(T* o) { Link<T>* link = new Link<T>; link->value= o; link->next = firstlink; link->list = this; if (firstlink) { firstlink->prev=link; } else { lastlink=link; } firstlink = link; return link; } static LinkedList<T>* Create() { return new LinkedList<T>; } }; } Use like this #include "LinkedList.h" using namespace Leadwerks3D; class ABC { }; int main(int argc, char* argv[]) { LinkedList<ABC> list; list.AddFirst(new ABC() ); list.AddLast(new ABC() ); return 0; }

As usual with Shadmar's shader wonders I'm knocked to the floor

Looking forward to use this. Previously I did not actually use 3DWS, mostly because it seemed to simple or ... well I don't really know why. Anyway I have lately (better late than never) discovered how good that tools is. So I really look forward to this super-3dws-thing. Question Is there any kind of snap to object function in this new beast. That would be sooooo great and end of trying to get things (like walls) built together will endless small movements with either the mouse or editing the positions values.

Wow. Thanks thanks thanks. Shadmar Now I can get good details using tiles and at the same get dirt there with good resolution. That was really kind of you to take the time for this. Much appreciated. Hope this of some usage for others also

@Shadmar I'm probably doing something wrong. With those changes my MaterialEditor and Editor crashes. I have attached my files.

Not quite. Grrrr .. my lack of English. Let's illustrate then Lets say I have this tile (256x256) and an oyerlay (512x512) Then I want to add the tile to plane (tile 4 times = 512x512) and the place the overlay on the same plane, like this or without the red tile markers and of course I want to place the overlay in code (or bye mat-file) so I can use different overlays.

We are tuned

You guys are just so awesome Can't wait until I get home a test

Programmers making art ... yeah.. I know ... Anyway. I making some stone walls and had an idea of using two texture layers. One layer that is tiled and another that is not. The tiled layer is the base diffuse color (and normal map) while the non-tiled layer would be some dirt and grime added. My idea was to get a high detail on the base color and then add some flavor to that using the second layer. But how is this done? Or is it even possible to do. I have seen that its possible to have two UV-maps on the models and use different textures for each of them. Is that how to do it? And if so. Can Leadwerks handle two uv-maps with different textures? Have you guys any suggestions?

Soul brothers maybe

I fully understand what Joh is saying. I also do this for fun, no more, no less. Developers can have different goals. Some want to get things done as quickly as possible and care less about how. The main goal is to get things up working as fast a possible with as little effort as possible. I can fully understand the need for such an approach, specially if the output is going to be a commercial game giving some income. In that case, lib's that save time and effort is gold. Then we have another type of developer like my self. I'm not into this to get an income or hit the world with something spectacular. All this is just for fun, a hobby. Of course I also want to get things done, but in my case I care more about how its done and less about the time it takes to get it done. I guess that Joh may belong to this group of developers as I do. Both types of developers has their own justification and different methods, and that's a good thing. In my eyes there is no 'better' or 'worse' method of developing using lib's and third-party code. It all depends of what the intent is. So use lib's anytime you like to, or make your own if that's fun, as long as you are enjoying what you do.

I know that feeling

Though Josh said it was fun, not breakthrough

I also want to have fun.... Hurry up Josh

Naugthy. I trust in your experience. Think I will try to reduce the cylinder diameter a bit and go with that. Later on I will have to dive into code manipulated movement of limbs anyway (some simple IK for moving the hand to pick up things), so I will test Scarlets suggestion at that point. Thank's guys You are great.

Wow. Thanks all for the input. I will test with all your advices Josh! Wouldn't it be just like piece of nothing to replace the cylinder with a capsule... wishful thinking ... Ken. Great video. Thanks

Well.. actually my feet's are on the ground, but my game character has that problem. Walking on flat terrain is no big problem, but when she enters hills and rough terrain her feet's floats above the ground. I guess that is cause by some collision mesh thing and is hard to avoid. Above she is standing on a slope. (left image with Debug Physics on). I guess make the diameter of the collision cylinder would make the problem less prominent, but is that the only way to go without writing your own controller. How do you guys solve this or at least minimize the problem?