Gaulish Gaming Banquet: hskal

pegperegogaucho

Hmm, weird, I never edited the value and it still seemed to work just fine :FrogThinking:

hesopesomeso

``` frame_a2 = a2->data.data.data.data.rw_frame; frame_a1 = a1->data.data.data.data.rw_frame; rsum = a1->data.data.radius + a2->data.data.radius; xdiff = frame_a1->ltm.pos.x - frame_a2->ltm.pos.x; if (xdiff) > (double)rsum ) return 0; zdiff = frame_a1->ltm.pos.z - frame_a2->ltm.pos.z; if ( zdiff > (double)rsum ) return 0; ydiff = frame_a1->ltm.pos.y - frame_a2->ltm.pos.y; if (ydiff > (double)rsum ) return 0; ... ``` This is a kobb-kobb collision detection check, and as you can see they first check the radii for a quick check.

pegperegogaucho

Makes sense, although might be superfluous for shapes like AABB or spheres themselves

pegperegogaucho

Maybe for my stuff the radius was always big enough...

hesopesomeso

Spheres actually use that value, don't they?

pegperegogaucho

Okay, you got me

hesopesomeso

And this?

pegperegogaucho

That is probably for the cannon button, if it would be present

pegperegogaucho

I'm sure that their editor automatically made it regardless of machine type

pegperegogaucho

I believe bombs can also trigger that shape and make the machine extend the rollers despite it not being the other type of machine. This happened a couple of times in my UXXL playthroughs :PogelixOkay:

hesopesomeso

Lol

hesopesomeso

Another one of those...

hesopesomeso

Either way, I got something going finally....

pegperegogaucho

Yes that is a fiendish bug, you have to make sure to increment the counter each iteration

pegperegogaucho

I have also fallen victim to it many times with range-based for loops. I believe it might be smarter to just use normal for loops since this issue is very dangerous

pegperegogaucho

Maybe another solution would be to make a "enumerate" function like in python, so you can use structured bindings for both the index and the element and always have it match, but I haven't tried that out yet, might cause the compiler to emit less efficient code for no reason

hesopesomeso

Yeah... It's just so convenient though- I don't have to type 10 more characters :Obelul:

I'll have to make sure I always put it at the end of the for scope...

hesopesomeso

Why not just make C++ into Python altogether? :KEKW:

adrientd

There is a `std::ranges::views::enumerate` function doing that, but only since C++23...

hesopesomeso

```c++ for (auto [i, action] : std::views::enumerate(std::ranges::ref_view(actions))) { if (action.is_active()) { action.process(action_values[i], time); } } ``` I'd have to do something like this.

#hskal