Converting sf::Event::MouseMoveEvent to sf::Vector2i

[kim366]

Let's say we have

sf::Vector2i v;

We get

error: conversion from 'sf::Event::MouseMoveEvent' to non-scalar type 'sf::Vector2i {aka sf::Vector2<int>}' requested

when doing

v = event.mouseMove;

The following works fine, but is ugly

v = sf::Vector2i(event.mouseMove.x, event.mouseMove.y)

There also is no explicit conversion.

Is there a nicer way of doing this?

Thanks,
Kim

PS: A bit off-topic, but why does Vector2 have no parenthesis overload?

[Hapax]

Assuming you can use C++11, you can use:

v = { event.mouseMove.x, event.mouseMove.y };

which is less ugly.

Since both types there are both int, it'll work fine. However, to convert, you can either construct a new vector:

sf::Vector2f floatVector;
// ...
floatVector = sf::Vector2f(intVector);

or cast the values directly (again, from C++11):

sf::Vector2f floatVector;
// ...
floatVector = { static_cast<float>(intVector.x), static_cast<float>(intVector.y) };

[kim366]

Okay, thanks, but it is not possible to cast it all at once, you do have to make a new vector from the x and y values?

[Hapax]

You need to create a new vector to convert as one (this is still pretty quick).

Casting, however, is not creating a new vector. It's simply casting each value and placing that casted value into the existing vector.
I casted each value separately because I tend to need to do that as I generally cast separately when it's not just a direct conversion. I would use the vector's constructor for that.
Statically casting the vector does work, though; it simply uses the vector's constructor to convert.

[eXpl0it3r]

Casting from a float to an int vector often is also a sign of wrong usage. Why do you need integer positions when you have floating point values? "Converting" from float to int will simply cut off the decimal values, which isn't what you'd always want (3.9 is closer to 4 than to 3).
And finally if you want to convert from game space to screen space, you should use the mapCoordsToPixel function that takes the view into account.

[Laurent]

MouseMoveEvent gives X and Y separately, so yes, if you want to create a Vector2i from these coordinates, you have to... create a Vector2i from these coordinates (the thing you're calling "ugly").

A bit off-topic, but why does Vector2 have no parenthesis overload?

I have no idea what you mean. An overload of the "parenthesis" operator () is defined when you need to "call" the object, so I guess this is not what you're talking about since it makes no sense to "call" vectors.

[DarkRoku12]

A bit off-topic, but why does Vector2 have no parenthesis overload?

I have no idea what you mean. An overload of the "parenthesis" operator () is defined when you need to "call" the object, so I guess this is not what you're talking about since it makes no sense to "call" vectors.

I think he means: user-defined conversion

[Hapax]

I just wanted to add that the idea of casting from one vector to another just makes it more "ugly":

static_cast<sf::Vector2f>(intVector)

vs

sf::Vector2f(intVector)

I presume by "ugly", you mean more characters...

Casting from a float to an int vector often is also a sign of wrong usage. Why do you need integer positions when you have floating point values?[...]

Just to be clear, there was no conversion from float to int. The original post was about creating a vector from components and the conversion in my reply (from int to float) was in response to their question about conversion.
I agree with you that casting from float to int can be more troublesome than originally expected. 3.9f casted to int would be 3 (as you already mentioned) but even rounding the float first doesn't guarantee it rounding upwards. round(3.9f) might result in a value very slightly below 4 so would still cast to in as 3.
That said, it's possible to fix this: static_cast<int>(std::round(3.9f) + 0.5f);. Well, that is how I would do it