Correct mouse position after moving view?

[Hmmz]

How do I retrieve the correct mouse position after moving view?

At the moment, I am doing it this way:

pixelPos = Mouse::getPosition(window);
coordPos = window.mapPixelToCoords(pixelPos);

I tried passing the currently used view as the second parameter to the 'mapPixelToCoords' method, but it didn't help. I am trying to snap objects to a 48x48 grid, but the farther I move the view, the more mouse input deviates from the expected.

What am I doing wrong?

http://pastebin.com/6tGSwLQz

[G.]

I tried your code and it works perfectly. How far are we supposed to move the view to notice the problem? I moved it up to 15000, 13000.

[Hmmz]

I tried your code and it works perfectly. How far are we supposed to move the view to notice the problem? I moved it up to 15000, 13000.

Oh. Sorry. I am moving it to the left.
Here's how it looks:
http://i.imgur.com/icSpNmx.png

Notice the cursor and where the object spawned. Hm... strangely enough, it does seem to work okay if moving to the right. I wonder if I am perhaps not handling negative values correctly... but if that were the case, the objects would snap to the left of the cursor and not to the right.

Or are we not allowed to use negative coords in general? If not, maybe I should just limit myself to positive values and be done with (actually, that would be one heck of a way to avoid cryptic algorithms)

[Estivo]

I have my view class with getPosition method, I think is the best option. If you have your custom view you move object by

obj.pos = mouse + view.getPosition();

I think it is the easiest way.

[G.]

When x is negative, the remainder of x % 48 is also negative. You then remove the negative remainder from x, -- gives + so it actually adds the remainder and your rectangle is drawn to a multiple of 48 to the right of where it should be.

For an obvious solution, remove 48 to x when x is < 0. (do the same for y) There are probably better solutions.

int x = static_cast<int>(std::trunc(coordPos.x));
int y = static_cast<int>(std::trunc(coordPos.y));

if (x < 0) x -=48;
if (y < 0) y -=48;

x -= x % 48;
y -= y % 48;

coordPos.x = static_cast<float>(x);
coordPos.y = static_cast<float>(y);

[Hmmz]

Well... now I feel dumb. Can't believe I was thinking that I was doing something incorrectly from the coding point of view rather than doing simple mathematics!

But I decided to limit myself to just positive coordinates. Since my game is going to be data-driven, negative positions inside a JSON file would make a really ugly sight, I presume. :p