[Solved specifically]Detecting click position?

[Naiten]

Well, myitemclass.onLeftClickEntity() allows me to perform different stuff with the entity i clicked. But how do i know the position at wich player clicked? I.e. i want to know if player has clicked front/rear side of my entity. Any ways to do that?

[hydroflame]

look at the angle between the entity facing direction and the player facing direction

[Naiten]

Way too complicated. Even considering i'm good at math and trigonometry...

[hydroflame]

front/rear side of my entity.

suspicious look

 

float angleDifference = entity1.rotationYaw-entity2.rotationYaw;

if they are facing itll be roughly 180, if its on the side itll be roughly 90 or 270 (depending on sides)

if its from the back itll be roughly 0

 

 

... now you're just mean. be honest, tell me if you dont want my help

[Naiten]

You code is obviously not going to work. If you think it will, take a look at the picture:

 

 

E — wagon with front and rear couplers; P1, P2 — players; rY — entity rotationYaw; players have equal rotationYaw, that in this case is zero.

If i use your code it will work next way:

angleDifference_1 = rY - 0;

angleDifference_2 = rY - 0;

angleDifference_1 = angleDifference_2 =>

it will return that both players have clicked same couplers, what is false.

 

Next time think a little, before advising something that is not working and saying i'm „just mean“ and „dont want your help“.

[hydroflame]

valid,

raytracing is the only option left then

[Mew]

You code is obviously not going to work. If you think it will, take a look at the picture:

 

 

E — wagon with front and rear couplers; P1, P2 — players; rY — entity rotationYaw; players have equal rotationYaw, that in this case is zero.

If i use your code it will work next way:

angleDifference_1 = rY - 0;

angleDifference_2 = rY - 0;

angleDifference_1 = angleDifference_2 =>

it will return that both players have clicked same couplers, that is false.

 

Next time think a little, before advising something that is not working and saying i'm „just mean“ and „dont want our help“.

 

I don't want another rage fight to happen (had one on FB last night, no ones business ).

 

By looking at your drawing, it seems to be that you are right. BUT!!!!! You did not specify wether you wanted a front or back of the same thing, thus not fully explaining your question. For what WAS asked, you where given the answer. For what you didn't ask, you obviously weren't given an answer.

 

Therefore the lessons we can ALL take away from this is: Ask the ENTIRE question you want answered, don't jump to conclusions, and for me, don't jump into random business (except for that fact that hydroflame is technically my boss in a way is my coworker...).

[Naiten]

You did not specify wether you wanted a front or back of the same thing

Haven't i wrote i want to know if player has clicked front/rear side of my entity ? Are my eyes and memory tricking me?

thus not fully explaining your question. For what WAS asked, you where given the answer. For what you didn't ask, you obviously weren't given an answer.

 

Therefore the lessons we can ALL take away from this is: Ask the ENTIRE question you want answered, don't jump to conclusions, and for me, don't jump into random business (except for that fact that hydroflame is technically my boss in a way is my coworker...).

The lesson you should take from this — read more attentively, and don't be a smart-ass.

[hydroflame]

fuck it you're on my blacklist, maybe someone else has time to put up with your shit.

[Naiten]

hydroflame, huh, what happened now? Have i unfairly offended you or what?

[Mew]

hydroflame, huh, what happened now? Have i unfairly offended you or what?

 

I am sorry I did not explain well enough. You never said that you had a mirrored model. Therefore, it still gives you the side you clicked on. And double checking your drawing, it is well and truly wrong...

[pelep]

wth. why does your post have a thank you from me?o_o

 

anyway, you can use hydroflame's suggestion. you just have to tweak it a little. since you say you're good in math, you should be able to understand this:

 

take the player's position to be a point of origin then check what quadrant the entity's position is in relative to that point. if it's in quadrant 1, the difference in angle (hydroflame's suggestion) should be x degrees. if it's in quadrant 2, it should be something different, and so on

[GotoLink]

I would use getLookVec() from EntityLivingBase.

This is a code I made for Battlegear's dagger backstab ability:

protected boolean performBackStab(Item item, EntityLivingBase entityHit, EntityLivingBase entityHitting) {
        //Get victim and murderer vector views at hit time
        double[] victimView = new double[]{entityHit.getLookVec().xCoord,entityHit.getLookVec().zCoord};
        double[] murdererView = new double[]{entityHitting.getLookVec().xCoord,entityHitting.getLookVec().zCoord};
        //back-stab conditions: vectors are closely enough aligned, (fuzzy parameter might need testing)
        //but not in opposite directions (face to face or sideways)
        if(Math.abs(victimView[0]*murdererView[1]-victimView[1]*murdererView[0])<backstabFuzzy &&
                Math.signum(victimView[0])==Math.signum(murdererView[0]) &&
                Math.signum(victimView[1])==Math.signum(murdererView[1])){
            return ((IBackStabbable)item).onBackStab(entityHit, entityHitting);//Perform back stab effect
        }
        return false;
    }

I use a fuzzy parameter (close to 0) to allow player some chance of doing it.

Hope that helped.

[Naiten]

I am sorry I did not explain well enough. You never said that you had a mirrored model. Therefore, it still gives you the side you clicked on. And double checking your drawing, it is well and truly wrong...

I don't understand wtf you are talking about, especially 'mirrored model' and following part. And why you are quoting my question about such anger, presented to hydroplane.

 

pelep, GotoLink, thanks for suggestions, but i suddenly got an idea, that will definitely work and which i like more, because i have half of it already done before. I'm currently developing it and i'll share the result if anybody wants...

[Naiten]

So, i'm almost done with it. Currently the code works fine only on flat plane, because it's one o'clock am, i'm tired and don't remember how to write line equations in three dimensions. Also, it will work fine only in my or similar case, because i don't really need side-detection, rather than point-detection ('cause i represent wagon couplers as points where they are located).

 

My draft:

[/img]

 

So, since couplers are points that are placed on wagon axis and rotate together with it, i can define their global positions by wagon position, it's rotation and preset distances between couplers and wagon centre. See the top of the draft.

 

//This is what i've made before
//somewhere in my wagon class
rotationCos = Math.cos(Math.toRadians(this.rotationYaw + 90)); //note that i add 90 degrees, because 
rotationSin = Math.sin(Math.toRadians(this.rotationYaw + 90)); //rotationYaw equals zero when parallel to Z-axis
frontCouplerX = posX + rotationCos * frontCouplerShift);
frontCouplerZ = posZ + rotationSin * frontCouplerShift);
rearCouplerX = posX + rotationCos * rearCouplerShift);
rearCouplerZ = posZ + rotationSin * rearCouplerShift);

 

Then i define distance between the front/rear point and the line of player's sight direction (which is length of a perpendicular passing from point to line). Lesser value indicates needed coupler. See the top middle of the draft.

To define that distance i've written a function, because i've not found anything about it in MC sources. Maybe i was searching bad, idk. I'm bad at explaining, but however: both given line and it's perpendicular could be described by angular coefficient equations, which look like y = kx + b, where k is angular coefficient (or a tangent of the angle between line and X-axis) and b is y-intercept of the line.

After equations are made, they are combined and then solved, so we know coordinates of the interception point H. Finally, i use Pythagorean to define the distance between given and intersection points. See the bottom of the draft.

 

public float dist_point2line(float mx, float mz, float qx, float qz, float angle){
//[b]mx-mz[/b] - coordinates of the separate point (coupler), [b]qx-qz[/b] - coordinates of any point that belongs to the line
//in my case it's player position. [b]angle[/b] = angle between the line and X-axis.
float k  = (float) Math.tan(Math.toRadians(angle));
float k1 = (float) Math.tan(Math.toRadians(angle + 90F));
float b  = (float) (qz - k * qx);
float b1  = (float) (mz - k1 * mx);
float hx = (b - b1)/(k1 - k);
float hz = k * hx + b;
return (float) Math.sqrt((hx - mx) * (hx - mx) + (hz - mz) * (hz - mz));
//this was not optimized, because written to show the algorithm.
}

 

And the last piece of code:

 

float dist_front = dist_point2line(cart.frontCouplerX, cart.frontCouplerZ, 
(float) player.posX, (float) player.posZ, player.rotationYaw + 90);
float dist_rear = dist_point2line(cart.rearCouplerX, cart.rearCouplerZ, 
(float) player.posX, (float) player.posZ, player.rotationYaw + 90);
if(dist_front < dist_rear) player.addChatMessage("Clicked on front coupler");
else player.addChatMessage("Clicked on rear coupler");

 

 

Final note: this code was written especially for my case, if you want something more or less specific, i don't promise good results...