bevy/crates/bevy_math/src/direction.rs

1085 lines
35 KiB
Rust
Raw Normal View History

use crate::{
primitives::{Primitive2d, Primitive3d},
Quat, Rot2, Vec2, Vec3, Vec3A,
};
use core::f32::consts::FRAC_1_SQRT_2;
#[cfg(feature = "bevy_reflect")]
use bevy_reflect::Reflect;
#[cfg(all(feature = "serialize", feature = "bevy_reflect"))]
use bevy_reflect::{ReflectDeserialize, ReflectSerialize};
/// An error indicating that a direction is invalid.
#[derive(Debug, PartialEq)]
pub enum InvalidDirectionError {
/// The length of the direction vector is zero or very close to zero.
Zero,
/// The length of the direction vector is `std::f32::INFINITY`.
Infinite,
/// The length of the direction vector is `NaN`.
NaN,
}
impl InvalidDirectionError {
/// Creates an [`InvalidDirectionError`] from the length of an invalid direction vector.
pub fn from_length(length: f32) -> Self {
if length.is_nan() {
InvalidDirectionError::NaN
} else if !length.is_finite() {
// If the direction is non-finite but also not NaN, it must be infinite
InvalidDirectionError::Infinite
} else {
// If the direction is invalid but neither NaN nor infinite, it must be zero
InvalidDirectionError::Zero
}
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::fmt::Display for InvalidDirectionError {
fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
write!(
f,
"Direction can not be zero (or very close to zero), or non-finite."
)
}
}
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
/// Checks that a vector with the given squared length is normalized.
///
/// Warns for small error with a length threshold of approximately `1e-4`,
/// and panics for large error with a length threshold of approximately `1e-2`.
///
/// The format used for the logged warning is `"Warning: {warning} The length is {length}`,
/// and similarly for the error.
#[cfg(debug_assertions)]
fn assert_is_normalized(message: &str, length_squared: f32) {
let length_error_squared = (length_squared - 1.0).abs();
// Panic for large error and warn for slight error.
if length_error_squared > 2e-2 || length_error_squared.is_nan() {
// Length error is approximately 1e-2 or more.
panic!("Error: {message} The length is {}.", length_squared.sqrt());
} else if length_error_squared > 2e-4 {
// Length error is approximately 1e-4 or more.
eprintln!(
"Warning: {message} The length is {}.",
length_squared.sqrt()
);
}
}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
/// A normalized vector pointing in a direction in 2D space
#[deprecated(
since = "0.14.0",
note = "`Direction2d` has been renamed. Please use `Dir2` instead."
)]
pub type Direction2d = Dir2;
/// A normalized vector pointing in a direction in 3D space
#[deprecated(
since = "0.14.0",
note = "`Direction3d` has been renamed. Please use `Dir3` instead."
)]
pub type Direction3d = Dir3;
/// A normalized vector pointing in a direction in 2D space
#[derive(Clone, Copy, Debug, PartialEq)]
#[cfg_attr(feature = "serialize", derive(serde::Serialize, serde::Deserialize))]
#[cfg_attr(feature = "bevy_reflect", derive(Reflect), reflect(Debug, PartialEq))]
#[cfg_attr(
all(feature = "serialize", feature = "bevy_reflect"),
reflect(Serialize, Deserialize)
)]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
#[doc(alias = "Direction2d")]
pub struct Dir2(Vec2);
impl Primitive2d for Dir2 {}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl Dir2 {
/// A unit vector pointing along the positive X axis.
pub const X: Self = Self(Vec2::X);
/// A unit vector pointing along the positive Y axis.
pub const Y: Self = Self(Vec2::Y);
/// A unit vector pointing along the negative X axis.
pub const NEG_X: Self = Self(Vec2::NEG_X);
/// A unit vector pointing along the negative Y axis.
pub const NEG_Y: Self = Self(Vec2::NEG_Y);
/// The directional axes.
pub const AXES: [Self; 2] = [Self::X, Self::Y];
/// The "north" direction, equivalent to [`Dir2::Y`].
pub const NORTH: Self = Self(Vec2::Y);
/// The "south" direction, equivalent to [`Dir2::NEG_Y`].
pub const SOUTH: Self = Self(Vec2::NEG_Y);
/// The "east" direction, equivalent to [`Dir2::X`].
pub const EAST: Self = Self(Vec2::X);
/// The "west" direction, equivalent to [`Dir2::NEG_X`].
pub const WEST: Self = Self(Vec2::NEG_X);
/// The "north-east" direction, between [`Dir2::NORTH`] and [`Dir2::EAST`].
pub const NORTH_EAST: Self = Self(Vec2::new(FRAC_1_SQRT_2, FRAC_1_SQRT_2));
/// The "north-west" direction, between [`Dir2::NORTH`] and [`Dir2::WEST`].
pub const NORTH_WEST: Self = Self(Vec2::new(-FRAC_1_SQRT_2, FRAC_1_SQRT_2));
/// The "south-east" direction, between [`Dir2::SOUTH`] and [`Dir2::EAST`].
pub const SOUTH_EAST: Self = Self(Vec2::new(FRAC_1_SQRT_2, -FRAC_1_SQRT_2));
/// The "south-west" direction, between [`Dir2::SOUTH`] and [`Dir2::WEST`].
pub const SOUTH_WEST: Self = Self(Vec2::new(-FRAC_1_SQRT_2, -FRAC_1_SQRT_2));
/// Create a direction from a finite, nonzero [`Vec2`], normalizing it.
///
/// Returns [`Err(InvalidDirectionError)`](InvalidDirectionError) if the length
/// of the given vector is zero (or very close to zero), infinite, or `NaN`.
pub fn new(value: Vec2) -> Result<Self, InvalidDirectionError> {
Self::new_and_length(value).map(|(dir, _)| dir)
}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
/// Create a [`Dir2`] from a [`Vec2`] that is already normalized.
///
/// # Warning
///
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
/// `value` must be normalized, i.e its length must be `1.0`.
pub fn new_unchecked(value: Vec2) -> Self {
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
#[cfg(debug_assertions)]
assert_is_normalized(
"The vector given to `Dir2::new_unchecked` is not normalized.",
value.length_squared(),
);
Self(value)
}
/// Create a direction from a finite, nonzero [`Vec2`], normalizing it and
/// also returning its original length.
///
/// Returns [`Err(InvalidDirectionError)`](InvalidDirectionError) if the length
/// of the given vector is zero (or very close to zero), infinite, or `NaN`.
pub fn new_and_length(value: Vec2) -> Result<(Self, f32), InvalidDirectionError> {
let length = value.length();
let direction = (length.is_finite() && length > 0.0).then_some(value / length);
direction
.map(|dir| (Self(dir), length))
.ok_or(InvalidDirectionError::from_length(length))
}
/// Create a direction from its `x` and `y` components.
///
/// Returns [`Err(InvalidDirectionError)`](InvalidDirectionError) if the length
/// of the vector formed by the components is zero (or very close to zero), infinite, or `NaN`.
pub fn from_xy(x: f32, y: f32) -> Result<Self, InvalidDirectionError> {
Self::new(Vec2::new(x, y))
}
Add methods to return the inner value for direction types (#12516) # Objective Currently in order to retrieve the inner values from direction types is that you need to use the `Deref` trait or `From`/`Into`. `Deref` that is currently implemented is an anti-pattern that I believe should be less relied upon. This pull-request add getters for retrieving the inner values for direction types. Advantages of getters: - Let rust-analyzer to list out available methods for users to understand better to on how to get the inner value. (This happens to me. I really don't know how to get the value until I look through the source code.) - They are simple. - Generally won't be ambiguous in most context. Traits such as `From`/`Into` will require fully qualified syntax from time to time. - Unsurprising result. Disadvantages of getters: - More verbose Advantages of deref polymorphism: - You reduce boilerplate for getting the value and call inner methods by: ```rust let dir = Dir3::new(Vec3::UP).unwrap(); // getting value let value = *dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` Disadvantages of deref polymorphism: - When under more level of indirection, it will requires more dereferencing which will get ugly in some part: ```rust // getting value let value = **dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` [More detail here](https://rust-unofficial.github.io/patterns/anti_patterns/deref.html). Edit: Update information for From/Into trait. Edit: Advantages and disadvantages. ## Solution Add `vec2` method for Dir2. Add `vec3` method for Dir3. Add `vec3a` method for Dir3A.
2024-03-18 17:49:58 +00:00
/// Create a direction from its `x` and `y` components, assuming the resulting vector is normalized.
///
/// # Warning
///
/// The vector produced from `x` and `y` must be normalized, i.e its length must be `1.0`.
pub fn from_xy_unchecked(x: f32, y: f32) -> Self {
Self::new_unchecked(Vec2::new(x, y))
}
Add methods to return the inner value for direction types (#12516) # Objective Currently in order to retrieve the inner values from direction types is that you need to use the `Deref` trait or `From`/`Into`. `Deref` that is currently implemented is an anti-pattern that I believe should be less relied upon. This pull-request add getters for retrieving the inner values for direction types. Advantages of getters: - Let rust-analyzer to list out available methods for users to understand better to on how to get the inner value. (This happens to me. I really don't know how to get the value until I look through the source code.) - They are simple. - Generally won't be ambiguous in most context. Traits such as `From`/`Into` will require fully qualified syntax from time to time. - Unsurprising result. Disadvantages of getters: - More verbose Advantages of deref polymorphism: - You reduce boilerplate for getting the value and call inner methods by: ```rust let dir = Dir3::new(Vec3::UP).unwrap(); // getting value let value = *dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` Disadvantages of deref polymorphism: - When under more level of indirection, it will requires more dereferencing which will get ugly in some part: ```rust // getting value let value = **dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` [More detail here](https://rust-unofficial.github.io/patterns/anti_patterns/deref.html). Edit: Update information for From/Into trait. Edit: Advantages and disadvantages. ## Solution Add `vec2` method for Dir2. Add `vec3` method for Dir3. Add `vec3a` method for Dir3A.
2024-03-18 17:49:58 +00:00
/// Returns the inner [`Vec2`]
pub const fn as_vec2(&self) -> Vec2 {
self.0
}
/// Performs a spherical linear interpolation between `self` and `rhs`
/// based on the value `s`.
///
/// This corresponds to interpolating between the two directions at a constant angular velocity.
///
/// When `s == 0.0`, the result will be equal to `self`.
/// When `s == 1.0`, the result will be equal to `rhs`.
///
/// # Example
///
/// ```
/// # use bevy_math::Dir2;
/// # use approx::{assert_relative_eq, RelativeEq};
/// #
/// let dir1 = Dir2::X;
/// let dir2 = Dir2::Y;
///
/// let result1 = dir1.slerp(dir2, 1.0 / 3.0);
/// assert_relative_eq!(result1, Dir2::from_xy(0.75_f32.sqrt(), 0.5).unwrap());
///
/// let result2 = dir1.slerp(dir2, 0.5);
/// assert_relative_eq!(result2, Dir2::from_xy(0.5_f32.sqrt(), 0.5_f32.sqrt()).unwrap());
/// ```
#[inline]
pub fn slerp(self, rhs: Self, s: f32) -> Self {
let angle = self.angle_to(rhs.0);
Rot2::radians(angle * s) * self
}
/// Get the rotation that rotates this direction to `other`.
#[inline]
pub fn rotation_to(self, other: Self) -> Rot2 {
// Rotate `self` to X-axis, then X-axis to `other`:
other.rotation_from_x() * self.rotation_to_x()
}
/// Get the rotation that rotates `other` to this direction.
#[inline]
pub fn rotation_from(self, other: Self) -> Rot2 {
other.rotation_to(self)
}
/// Get the rotation that rotates the X-axis to this direction.
#[inline]
pub fn rotation_from_x(self) -> Rot2 {
Rot2::from_sin_cos(self.0.y, self.0.x)
}
/// Get the rotation that rotates this direction to the X-axis.
#[inline]
pub fn rotation_to_x(self) -> Rot2 {
// (This is cheap, it just negates one component.)
self.rotation_from_x().inverse()
}
/// Get the rotation that rotates the Y-axis to this direction.
#[inline]
pub fn rotation_from_y(self) -> Rot2 {
// `x <- y`, `y <- -x` correspond to rotating clockwise by pi/2;
// this transforms the Y-axis into the X-axis, maintaining the relative position
// of our direction. Then we just use the same technique as `rotation_from_x`.
Rot2::from_sin_cos(-self.0.x, self.0.y)
}
/// Get the rotation that rotates this direction to the Y-axis.
#[inline]
pub fn rotation_to_y(self) -> Rot2 {
self.rotation_from_y().inverse()
}
/// Returns `self` after an approximate normalization, assuming the value is already nearly normalized.
/// Useful for preventing numerical error accumulation.
/// See [`Dir3::fast_renormalize`] for an example of when such error accumulation might occur.
#[inline]
pub fn fast_renormalize(self) -> Self {
let length_squared = self.0.length_squared();
// Based on a Taylor approximation of the inverse square root, see [`Dir3::fast_renormalize`] for more details.
Self(self * (0.5 * (3.0 - length_squared)))
}
}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl TryFrom<Vec2> for Dir2 {
type Error = InvalidDirectionError;
fn try_from(value: Vec2) -> Result<Self, Self::Error> {
Self::new(value)
}
}
impl From<Dir2> for Vec2 {
fn from(value: Dir2) -> Self {
value.as_vec2()
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Deref for Dir2 {
type Target = Vec2;
fn deref(&self) -> &Self::Target {
&self.0
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Neg for Dir2 {
type Output = Self;
fn neg(self) -> Self::Output {
Self(-self.0)
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Mul<f32> for Dir2 {
type Output = Vec2;
fn mul(self, rhs: f32) -> Self::Output {
self.0 * rhs
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Mul<Dir2> for f32 {
type Output = Vec2;
fn mul(self, rhs: Dir2) -> Self::Output {
self * rhs.0
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Mul<Dir2> for Rot2 {
Add `Rotation2d` (#11658) # Objective Rotating vectors is a very common task. It is required for a variety of things both within Bevy itself and in many third party plugins, for example all over physics and collision detection, and for things like Bevy's bounding volumes and several gizmo implementations. For 3D, we can do this using a `Quat`, but for 2D, we do not have a clear and efficient option. `Mat2` can be used for rotating vectors if created using `Mat2::from_angle`, but this is not obvious to many users, it doesn't have many rotation helpers, and the type does not give any guarantees that it represents a valid rotation. We should have a proper type for 2D rotations. In addition to allowing for potential optimization, it would allow us to have a consistent and explicitly documented representation used throughout the engine, i.e. counterclockwise and in radians. ## Representation The mathematical formula for rotating a 2D vector is the following: ``` new_x = x * cos - y * sin new_y = x * sin + y * cos ``` Here, `sin` and `cos` are the sine and cosine of the rotation angle. Computing these every time when a vector needs to be rotated can be expensive, so the rotation shouldn't be just an `f32` angle. Instead, it is often more efficient to represent the rotation using the sine and cosine of the angle instead of storing the angle itself. This can be freely passed around and reused without unnecessary computations. The two options are either a 2x2 rotation matrix or a unit complex number where the cosine is the real part and the sine is the imaginary part. These are equivalent for the most part, but the unit complex representation is a bit more memory efficient (two `f32`s instead of four), so I chose that. This is like Nalgebra's [`UnitComplex`](https://docs.rs/nalgebra/latest/nalgebra/geometry/type.UnitComplex.html) type, which can be used for the [`Rotation2`](https://docs.rs/nalgebra/latest/nalgebra/geometry/type.Rotation2.html) type. ## Implementation Add a `Rotation2d` type represented as a unit complex number: ```rust /// A counterclockwise 2D rotation in radians. /// /// The rotation angle is wrapped to be within the `]-pi, pi]` range. pub struct Rotation2d { /// The cosine of the rotation angle in radians. /// /// This is the real part of the unit complex number representing the rotation. pub cos: f32, /// The sine of the rotation angle in radians. /// /// This is the imaginary part of the unit complex number representing the rotation. pub sin: f32, } ``` Using it is similar to using `Quat`, but in 2D: ```rust let rotation = Rotation2d::radians(PI / 2.0); // Rotate vector (also works on Direction2d!) assert_eq!(rotation * Vec2::X, Vec2::Y); // Get angle as degrees assert_eq!(rotation.as_degrees(), 90.0); // Getting sin and cos is free let (sin, cos) = rotation.sin_cos(); // "Subtract" rotations let rotation2 = Rotation2d::FRAC_PI_4; // there are constants! let diff = rotation * rotation2.inverse(); assert_eq!(diff.as_radians(), PI / 4.0); // This is equivalent to the above assert_eq!(rotation2.angle_between(rotation), PI / 4.0); // Lerp let rotation1 = Rotation2d::IDENTITY; let rotation2 = Rotation2d::FRAC_PI_2; let result = rotation1.lerp(rotation2, 0.5); assert_eq!(result.as_radians(), std::f32::consts::FRAC_PI_4); // Slerp let rotation1 = Rotation2d::FRAC_PI_4); let rotation2 = Rotation2d::degrees(-180.0); // we can use degrees too! let result = rotation1.slerp(rotation2, 1.0 / 3.0); assert_eq!(result.as_radians(), std::f32::consts::FRAC_PI_2); ``` There's also a `From<f32>` implementation for `Rotation2d`, which means that methods can still accept radians as floats if the argument uses `impl Into<Rotation2d>`. This means that adding `Rotation2d` shouldn't even be a breaking change. --- ## Changelog - Added `Rotation2d` - Bounding volume methods now take an `impl Into<Rotation2d>` - Gizmo methods with rotation now take an `impl Into<Rotation2d>` ## Future use cases - Collision detection (a type like this is quite essential considering how common vector rotations are) - `Transform` helpers (e.g. return a 2D rotation about the Z axis from a `Transform`) - The rotation used for `Transform2d` (#8268) - More gizmos, maybe meshes... everything in 2D that uses rotation --------- Co-authored-by: Tristan Guichaoua <33934311+tguichaoua@users.noreply.github.com> Co-authored-by: Robert Walter <robwalter96@gmail.com> Co-authored-by: IQuick 143 <IQuick143cz@gmail.com>
2024-03-11 19:11:57 +00:00
type Output = Dir2;
/// Rotates the [`Dir2`] using a [`Rot2`].
Add `Rotation2d` (#11658) # Objective Rotating vectors is a very common task. It is required for a variety of things both within Bevy itself and in many third party plugins, for example all over physics and collision detection, and for things like Bevy's bounding volumes and several gizmo implementations. For 3D, we can do this using a `Quat`, but for 2D, we do not have a clear and efficient option. `Mat2` can be used for rotating vectors if created using `Mat2::from_angle`, but this is not obvious to many users, it doesn't have many rotation helpers, and the type does not give any guarantees that it represents a valid rotation. We should have a proper type for 2D rotations. In addition to allowing for potential optimization, it would allow us to have a consistent and explicitly documented representation used throughout the engine, i.e. counterclockwise and in radians. ## Representation The mathematical formula for rotating a 2D vector is the following: ``` new_x = x * cos - y * sin new_y = x * sin + y * cos ``` Here, `sin` and `cos` are the sine and cosine of the rotation angle. Computing these every time when a vector needs to be rotated can be expensive, so the rotation shouldn't be just an `f32` angle. Instead, it is often more efficient to represent the rotation using the sine and cosine of the angle instead of storing the angle itself. This can be freely passed around and reused without unnecessary computations. The two options are either a 2x2 rotation matrix or a unit complex number where the cosine is the real part and the sine is the imaginary part. These are equivalent for the most part, but the unit complex representation is a bit more memory efficient (two `f32`s instead of four), so I chose that. This is like Nalgebra's [`UnitComplex`](https://docs.rs/nalgebra/latest/nalgebra/geometry/type.UnitComplex.html) type, which can be used for the [`Rotation2`](https://docs.rs/nalgebra/latest/nalgebra/geometry/type.Rotation2.html) type. ## Implementation Add a `Rotation2d` type represented as a unit complex number: ```rust /// A counterclockwise 2D rotation in radians. /// /// The rotation angle is wrapped to be within the `]-pi, pi]` range. pub struct Rotation2d { /// The cosine of the rotation angle in radians. /// /// This is the real part of the unit complex number representing the rotation. pub cos: f32, /// The sine of the rotation angle in radians. /// /// This is the imaginary part of the unit complex number representing the rotation. pub sin: f32, } ``` Using it is similar to using `Quat`, but in 2D: ```rust let rotation = Rotation2d::radians(PI / 2.0); // Rotate vector (also works on Direction2d!) assert_eq!(rotation * Vec2::X, Vec2::Y); // Get angle as degrees assert_eq!(rotation.as_degrees(), 90.0); // Getting sin and cos is free let (sin, cos) = rotation.sin_cos(); // "Subtract" rotations let rotation2 = Rotation2d::FRAC_PI_4; // there are constants! let diff = rotation * rotation2.inverse(); assert_eq!(diff.as_radians(), PI / 4.0); // This is equivalent to the above assert_eq!(rotation2.angle_between(rotation), PI / 4.0); // Lerp let rotation1 = Rotation2d::IDENTITY; let rotation2 = Rotation2d::FRAC_PI_2; let result = rotation1.lerp(rotation2, 0.5); assert_eq!(result.as_radians(), std::f32::consts::FRAC_PI_4); // Slerp let rotation1 = Rotation2d::FRAC_PI_4); let rotation2 = Rotation2d::degrees(-180.0); // we can use degrees too! let result = rotation1.slerp(rotation2, 1.0 / 3.0); assert_eq!(result.as_radians(), std::f32::consts::FRAC_PI_2); ``` There's also a `From<f32>` implementation for `Rotation2d`, which means that methods can still accept radians as floats if the argument uses `impl Into<Rotation2d>`. This means that adding `Rotation2d` shouldn't even be a breaking change. --- ## Changelog - Added `Rotation2d` - Bounding volume methods now take an `impl Into<Rotation2d>` - Gizmo methods with rotation now take an `impl Into<Rotation2d>` ## Future use cases - Collision detection (a type like this is quite essential considering how common vector rotations are) - `Transform` helpers (e.g. return a 2D rotation about the Z axis from a `Transform`) - The rotation used for `Transform2d` (#8268) - More gizmos, maybe meshes... everything in 2D that uses rotation --------- Co-authored-by: Tristan Guichaoua <33934311+tguichaoua@users.noreply.github.com> Co-authored-by: Robert Walter <robwalter96@gmail.com> Co-authored-by: IQuick 143 <IQuick143cz@gmail.com>
2024-03-11 19:11:57 +00:00
fn mul(self, direction: Dir2) -> Self::Output {
let rotated = self * *direction;
#[cfg(debug_assertions)]
assert_is_normalized(
"`Dir2` is denormalized after rotation.",
rotated.length_squared(),
);
Dir2(rotated)
}
}
#[cfg(any(feature = "approx", test))]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl approx::AbsDiffEq for Dir2 {
type Epsilon = f32;
fn default_epsilon() -> f32 {
f32::EPSILON
}
fn abs_diff_eq(&self, other: &Self, epsilon: f32) -> bool {
self.as_ref().abs_diff_eq(other.as_ref(), epsilon)
}
}
#[cfg(any(feature = "approx", test))]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl approx::RelativeEq for Dir2 {
fn default_max_relative() -> f32 {
f32::EPSILON
}
fn relative_eq(&self, other: &Self, epsilon: f32, max_relative: f32) -> bool {
self.as_ref()
.relative_eq(other.as_ref(), epsilon, max_relative)
}
}
#[cfg(any(feature = "approx", test))]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl approx::UlpsEq for Dir2 {
fn default_max_ulps() -> u32 {
4
}
fn ulps_eq(&self, other: &Self, epsilon: f32, max_ulps: u32) -> bool {
self.as_ref().ulps_eq(other.as_ref(), epsilon, max_ulps)
}
}
/// A normalized vector pointing in a direction in 3D space
#[derive(Clone, Copy, Debug, PartialEq)]
#[cfg_attr(feature = "serialize", derive(serde::Serialize, serde::Deserialize))]
#[cfg_attr(feature = "bevy_reflect", derive(Reflect), reflect(Debug, PartialEq))]
#[cfg_attr(
all(feature = "serialize", feature = "bevy_reflect"),
reflect(Serialize, Deserialize)
)]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
#[doc(alias = "Direction3d")]
pub struct Dir3(Vec3);
impl Primitive3d for Dir3 {}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl Dir3 {
/// A unit vector pointing along the positive X axis.
pub const X: Self = Self(Vec3::X);
/// A unit vector pointing along the positive Y axis.
pub const Y: Self = Self(Vec3::Y);
/// A unit vector pointing along the positive Z axis.
pub const Z: Self = Self(Vec3::Z);
/// A unit vector pointing along the negative X axis.
pub const NEG_X: Self = Self(Vec3::NEG_X);
/// A unit vector pointing along the negative Y axis.
pub const NEG_Y: Self = Self(Vec3::NEG_Y);
/// A unit vector pointing along the negative Z axis.
pub const NEG_Z: Self = Self(Vec3::NEG_Z);
/// The directional axes.
pub const AXES: [Self; 3] = [Self::X, Self::Y, Self::Z];
/// Create a direction from a finite, nonzero [`Vec3`], normalizing it.
///
/// Returns [`Err(InvalidDirectionError)`](InvalidDirectionError) if the length
/// of the given vector is zero (or very close to zero), infinite, or `NaN`.
pub fn new(value: Vec3) -> Result<Self, InvalidDirectionError> {
Self::new_and_length(value).map(|(dir, _)| dir)
}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
/// Create a [`Dir3`] from a [`Vec3`] that is already normalized.
///
/// # Warning
///
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
/// `value` must be normalized, i.e its length must be `1.0`.
pub fn new_unchecked(value: Vec3) -> Self {
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
#[cfg(debug_assertions)]
assert_is_normalized(
"The vector given to `Dir3::new_unchecked` is not normalized.",
value.length_squared(),
);
Self(value)
}
/// Create a direction from a finite, nonzero [`Vec3`], normalizing it and
/// also returning its original length.
///
/// Returns [`Err(InvalidDirectionError)`](InvalidDirectionError) if the length
/// of the given vector is zero (or very close to zero), infinite, or `NaN`.
pub fn new_and_length(value: Vec3) -> Result<(Self, f32), InvalidDirectionError> {
let length = value.length();
let direction = (length.is_finite() && length > 0.0).then_some(value / length);
direction
.map(|dir| (Self(dir), length))
.ok_or(InvalidDirectionError::from_length(length))
}
/// Create a direction from its `x`, `y`, and `z` components.
///
/// Returns [`Err(InvalidDirectionError)`](InvalidDirectionError) if the length
/// of the vector formed by the components is zero (or very close to zero), infinite, or `NaN`.
pub fn from_xyz(x: f32, y: f32, z: f32) -> Result<Self, InvalidDirectionError> {
Self::new(Vec3::new(x, y, z))
}
Add methods to return the inner value for direction types (#12516) # Objective Currently in order to retrieve the inner values from direction types is that you need to use the `Deref` trait or `From`/`Into`. `Deref` that is currently implemented is an anti-pattern that I believe should be less relied upon. This pull-request add getters for retrieving the inner values for direction types. Advantages of getters: - Let rust-analyzer to list out available methods for users to understand better to on how to get the inner value. (This happens to me. I really don't know how to get the value until I look through the source code.) - They are simple. - Generally won't be ambiguous in most context. Traits such as `From`/`Into` will require fully qualified syntax from time to time. - Unsurprising result. Disadvantages of getters: - More verbose Advantages of deref polymorphism: - You reduce boilerplate for getting the value and call inner methods by: ```rust let dir = Dir3::new(Vec3::UP).unwrap(); // getting value let value = *dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` Disadvantages of deref polymorphism: - When under more level of indirection, it will requires more dereferencing which will get ugly in some part: ```rust // getting value let value = **dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` [More detail here](https://rust-unofficial.github.io/patterns/anti_patterns/deref.html). Edit: Update information for From/Into trait. Edit: Advantages and disadvantages. ## Solution Add `vec2` method for Dir2. Add `vec3` method for Dir3. Add `vec3a` method for Dir3A.
2024-03-18 17:49:58 +00:00
/// Create a direction from its `x`, `y`, and `z` components, assuming the resulting vector is normalized.
///
/// # Warning
///
/// The vector produced from `x`, `y`, and `z` must be normalized, i.e its length must be `1.0`.
pub fn from_xyz_unchecked(x: f32, y: f32, z: f32) -> Self {
Self::new_unchecked(Vec3::new(x, y, z))
}
Add methods to return the inner value for direction types (#12516) # Objective Currently in order to retrieve the inner values from direction types is that you need to use the `Deref` trait or `From`/`Into`. `Deref` that is currently implemented is an anti-pattern that I believe should be less relied upon. This pull-request add getters for retrieving the inner values for direction types. Advantages of getters: - Let rust-analyzer to list out available methods for users to understand better to on how to get the inner value. (This happens to me. I really don't know how to get the value until I look through the source code.) - They are simple. - Generally won't be ambiguous in most context. Traits such as `From`/`Into` will require fully qualified syntax from time to time. - Unsurprising result. Disadvantages of getters: - More verbose Advantages of deref polymorphism: - You reduce boilerplate for getting the value and call inner methods by: ```rust let dir = Dir3::new(Vec3::UP).unwrap(); // getting value let value = *dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` Disadvantages of deref polymorphism: - When under more level of indirection, it will requires more dereferencing which will get ugly in some part: ```rust // getting value let value = **dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` [More detail here](https://rust-unofficial.github.io/patterns/anti_patterns/deref.html). Edit: Update information for From/Into trait. Edit: Advantages and disadvantages. ## Solution Add `vec2` method for Dir2. Add `vec3` method for Dir3. Add `vec3a` method for Dir3A.
2024-03-18 17:49:58 +00:00
/// Returns the inner [`Vec3`]
pub const fn as_vec3(&self) -> Vec3 {
self.0
}
/// Performs a spherical linear interpolation between `self` and `rhs`
/// based on the value `s`.
///
/// This corresponds to interpolating between the two directions at a constant angular velocity.
///
/// When `s == 0.0`, the result will be equal to `self`.
/// When `s == 1.0`, the result will be equal to `rhs`.
///
/// # Example
///
/// ```
/// # use bevy_math::Dir3;
/// # use approx::{assert_relative_eq, RelativeEq};
/// #
/// let dir1 = Dir3::X;
/// let dir2 = Dir3::Y;
///
/// let result1 = dir1.slerp(dir2, 1.0 / 3.0);
/// assert_relative_eq!(
/// result1,
/// Dir3::from_xyz(0.75_f32.sqrt(), 0.5, 0.0).unwrap(),
/// epsilon = 0.000001
/// );
///
/// let result2 = dir1.slerp(dir2, 0.5);
/// assert_relative_eq!(result2, Dir3::from_xyz(0.5_f32.sqrt(), 0.5_f32.sqrt(), 0.0).unwrap());
/// ```
#[inline]
pub fn slerp(self, rhs: Self, s: f32) -> Self {
let quat = Quat::IDENTITY.slerp(Quat::from_rotation_arc(self.0, rhs.0), s);
Dir3(quat.mul_vec3(self.0))
}
/// Returns `self` after an approximate normalization, assuming the value is already nearly normalized.
/// Useful for preventing numerical error accumulation.
///
/// # Example
/// The following seemingly benign code would start accumulating errors over time,
/// leading to `dir` eventually not being normalized anymore.
/// ```
/// # use bevy_math::prelude::*;
/// # let N: usize = 200;
/// let mut dir = Dir3::X;
/// let quaternion = Quat::from_euler(EulerRot::XYZ, 1.0, 2.0, 3.0);
/// for i in 0..N {
/// dir = quaternion * dir;
/// }
/// ```
/// Instead, do the following.
/// ```
/// # use bevy_math::prelude::*;
/// # let N: usize = 200;
/// let mut dir = Dir3::X;
/// let quaternion = Quat::from_euler(EulerRot::XYZ, 1.0, 2.0, 3.0);
/// for i in 0..N {
/// dir = quaternion * dir;
/// dir = dir.fast_renormalize();
/// }
/// ```
#[inline]
pub fn fast_renormalize(self) -> Self {
// We numerically approximate the inverse square root by a Taylor series around 1
// As we expect the error (x := length_squared - 1) to be small
// inverse_sqrt(length_squared) = (1 + x)^(-1/2) = 1 - 1/2 x + O(x²)
// inverse_sqrt(length_squared) ≈ 1 - 1/2 (length_squared - 1) = 1/2 (3 - length_squared)
// Iterative calls to this method quickly converge to a normalized value,
// so long as the denormalization is not large ~ O(1/10).
// One iteration can be described as:
// l_sq <- l_sq * (1 - 1/2 (l_sq - 1))²;
// Rewriting in terms of the error x:
// 1 + x <- (1 + x) * (1 - 1/2 x)²
// 1 + x <- (1 + x) * (1 - x + 1/4 x²)
// 1 + x <- 1 - x + 1/4 x² + x - x² + 1/4 x³
// x <- -1/4 x² (3 - x)
// If the error is small, say in a range of (-1/2, 1/2), then:
// |-1/4 x² (3 - x)| <= (3/4 + 1/4 * |x|) * x² <= (3/4 + 1/4 * 1/2) * x² < x² < 1/2 x
// Therefore the sequence of iterates converges to 0 error as a second order method.
let length_squared = self.0.length_squared();
Self(self * (0.5 * (3.0 - length_squared)))
}
}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl TryFrom<Vec3> for Dir3 {
type Error = InvalidDirectionError;
fn try_from(value: Vec3) -> Result<Self, Self::Error> {
Self::new(value)
}
}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl From<Dir3> for Vec3 {
fn from(value: Dir3) -> Self {
value.0
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Deref for Dir3 {
type Target = Vec3;
fn deref(&self) -> &Self::Target {
&self.0
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Neg for Dir3 {
type Output = Self;
fn neg(self) -> Self::Output {
Self(-self.0)
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Mul<f32> for Dir3 {
type Output = Vec3;
fn mul(self, rhs: f32) -> Self::Output {
self.0 * rhs
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Mul<Dir3> for f32 {
type Output = Vec3;
fn mul(self, rhs: Dir3) -> Self::Output {
self * rhs.0
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Mul<Dir3> for Quat {
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
type Output = Dir3;
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
/// Rotates the [`Dir3`] using a [`Quat`].
fn mul(self, direction: Dir3) -> Self::Output {
let rotated = self * *direction;
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
#[cfg(debug_assertions)]
assert_is_normalized(
"`Dir3` is denormalized after rotation.",
rotated.length_squared(),
);
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
Dir3(rotated)
}
}
#[cfg(feature = "approx")]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl approx::AbsDiffEq for Dir3 {
type Epsilon = f32;
fn default_epsilon() -> f32 {
f32::EPSILON
}
fn abs_diff_eq(&self, other: &Self, epsilon: f32) -> bool {
self.as_ref().abs_diff_eq(other.as_ref(), epsilon)
}
}
#[cfg(feature = "approx")]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl approx::RelativeEq for Dir3 {
fn default_max_relative() -> f32 {
f32::EPSILON
}
fn relative_eq(&self, other: &Self, epsilon: f32, max_relative: f32) -> bool {
self.as_ref()
.relative_eq(other.as_ref(), epsilon, max_relative)
}
}
#[cfg(feature = "approx")]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl approx::UlpsEq for Dir3 {
fn default_max_ulps() -> u32 {
4
}
fn ulps_eq(&self, other: &Self, epsilon: f32, max_ulps: u32) -> bool {
self.as_ref().ulps_eq(other.as_ref(), epsilon, max_ulps)
}
}
/// A normalized SIMD vector pointing in a direction in 3D space.
///
/// This type stores a 16 byte aligned [`Vec3A`].
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
/// This may or may not be faster than [`Dir3`]: make sure to benchmark!
#[derive(Clone, Copy, Debug, PartialEq)]
#[cfg_attr(feature = "serialize", derive(serde::Serialize, serde::Deserialize))]
#[cfg_attr(feature = "bevy_reflect", derive(Reflect), reflect(Debug, PartialEq))]
#[cfg_attr(
all(feature = "serialize", feature = "bevy_reflect"),
reflect(Serialize, Deserialize)
)]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
#[doc(alias = "Direction3dA")]
pub struct Dir3A(Vec3A);
impl Primitive3d for Dir3A {}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl Dir3A {
/// A unit vector pointing along the positive X axis.
pub const X: Self = Self(Vec3A::X);
/// A unit vector pointing along the positive Y axis.
pub const Y: Self = Self(Vec3A::Y);
/// A unit vector pointing along the positive Z axis.
pub const Z: Self = Self(Vec3A::Z);
/// A unit vector pointing along the negative X axis.
pub const NEG_X: Self = Self(Vec3A::NEG_X);
/// A unit vector pointing along the negative Y axis.
pub const NEG_Y: Self = Self(Vec3A::NEG_Y);
/// A unit vector pointing along the negative Z axis.
pub const NEG_Z: Self = Self(Vec3A::NEG_Z);
/// The directional axes.
pub const AXES: [Self; 3] = [Self::X, Self::Y, Self::Z];
/// Create a direction from a finite, nonzero [`Vec3A`], normalizing it.
///
/// Returns [`Err(InvalidDirectionError)`](InvalidDirectionError) if the length
/// of the given vector is zero (or very close to zero), infinite, or `NaN`.
pub fn new(value: Vec3A) -> Result<Self, InvalidDirectionError> {
Self::new_and_length(value).map(|(dir, _)| dir)
}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
/// Create a [`Dir3A`] from a [`Vec3A`] that is already normalized.
///
/// # Warning
///
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
/// `value` must be normalized, i.e its length must be `1.0`.
pub fn new_unchecked(value: Vec3A) -> Self {
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
#[cfg(debug_assertions)]
assert_is_normalized(
"The vector given to `Dir3A::new_unchecked` is not normalized.",
value.length_squared(),
);
Self(value)
}
/// Create a direction from a finite, nonzero [`Vec3A`], normalizing it and
/// also returning its original length.
///
/// Returns [`Err(InvalidDirectionError)`](InvalidDirectionError) if the length
/// of the given vector is zero (or very close to zero), infinite, or `NaN`.
pub fn new_and_length(value: Vec3A) -> Result<(Self, f32), InvalidDirectionError> {
let length = value.length();
let direction = (length.is_finite() && length > 0.0).then_some(value / length);
direction
.map(|dir| (Self(dir), length))
.ok_or(InvalidDirectionError::from_length(length))
}
/// Create a direction from its `x`, `y`, and `z` components.
///
/// Returns [`Err(InvalidDirectionError)`](InvalidDirectionError) if the length
/// of the vector formed by the components is zero (or very close to zero), infinite, or `NaN`.
pub fn from_xyz(x: f32, y: f32, z: f32) -> Result<Self, InvalidDirectionError> {
Self::new(Vec3A::new(x, y, z))
}
Add methods to return the inner value for direction types (#12516) # Objective Currently in order to retrieve the inner values from direction types is that you need to use the `Deref` trait or `From`/`Into`. `Deref` that is currently implemented is an anti-pattern that I believe should be less relied upon. This pull-request add getters for retrieving the inner values for direction types. Advantages of getters: - Let rust-analyzer to list out available methods for users to understand better to on how to get the inner value. (This happens to me. I really don't know how to get the value until I look through the source code.) - They are simple. - Generally won't be ambiguous in most context. Traits such as `From`/`Into` will require fully qualified syntax from time to time. - Unsurprising result. Disadvantages of getters: - More verbose Advantages of deref polymorphism: - You reduce boilerplate for getting the value and call inner methods by: ```rust let dir = Dir3::new(Vec3::UP).unwrap(); // getting value let value = *dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` Disadvantages of deref polymorphism: - When under more level of indirection, it will requires more dereferencing which will get ugly in some part: ```rust // getting value let value = **dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` [More detail here](https://rust-unofficial.github.io/patterns/anti_patterns/deref.html). Edit: Update information for From/Into trait. Edit: Advantages and disadvantages. ## Solution Add `vec2` method for Dir2. Add `vec3` method for Dir3. Add `vec3a` method for Dir3A.
2024-03-18 17:49:58 +00:00
/// Create a direction from its `x`, `y`, and `z` components, assuming the resulting vector is normalized.
///
/// # Warning
///
/// The vector produced from `x`, `y`, and `z` must be normalized, i.e its length must be `1.0`.
pub fn from_xyz_unchecked(x: f32, y: f32, z: f32) -> Self {
Self::new_unchecked(Vec3A::new(x, y, z))
}
Add methods to return the inner value for direction types (#12516) # Objective Currently in order to retrieve the inner values from direction types is that you need to use the `Deref` trait or `From`/`Into`. `Deref` that is currently implemented is an anti-pattern that I believe should be less relied upon. This pull-request add getters for retrieving the inner values for direction types. Advantages of getters: - Let rust-analyzer to list out available methods for users to understand better to on how to get the inner value. (This happens to me. I really don't know how to get the value until I look through the source code.) - They are simple. - Generally won't be ambiguous in most context. Traits such as `From`/`Into` will require fully qualified syntax from time to time. - Unsurprising result. Disadvantages of getters: - More verbose Advantages of deref polymorphism: - You reduce boilerplate for getting the value and call inner methods by: ```rust let dir = Dir3::new(Vec3::UP).unwrap(); // getting value let value = *dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` Disadvantages of deref polymorphism: - When under more level of indirection, it will requires more dereferencing which will get ugly in some part: ```rust // getting value let value = **dir; // instead of using getters let value = dir.vec3(); // calling methods for the inner vector dir.xy(); // instead of using getters dir.vec3().xy(); ``` [More detail here](https://rust-unofficial.github.io/patterns/anti_patterns/deref.html). Edit: Update information for From/Into trait. Edit: Advantages and disadvantages. ## Solution Add `vec2` method for Dir2. Add `vec3` method for Dir3. Add `vec3a` method for Dir3A.
2024-03-18 17:49:58 +00:00
/// Returns the inner [`Vec3A`]
pub const fn as_vec3a(&self) -> Vec3A {
self.0
}
/// Performs a spherical linear interpolation between `self` and `rhs`
/// based on the value `s`.
///
/// This corresponds to interpolating between the two directions at a constant angular velocity.
///
/// When `s == 0.0`, the result will be equal to `self`.
/// When `s == 1.0`, the result will be equal to `rhs`.
///
/// # Example
///
/// ```
/// # use bevy_math::Dir3A;
/// # use approx::{assert_relative_eq, RelativeEq};
/// #
/// let dir1 = Dir3A::X;
/// let dir2 = Dir3A::Y;
///
/// let result1 = dir1.slerp(dir2, 1.0 / 3.0);
/// assert_relative_eq!(
/// result1,
/// Dir3A::from_xyz(0.75_f32.sqrt(), 0.5, 0.0).unwrap(),
/// epsilon = 0.000001
/// );
///
/// let result2 = dir1.slerp(dir2, 0.5);
/// assert_relative_eq!(result2, Dir3A::from_xyz(0.5_f32.sqrt(), 0.5_f32.sqrt(), 0.0).unwrap());
/// ```
#[inline]
pub fn slerp(self, rhs: Self, s: f32) -> Self {
let quat = Quat::IDENTITY.slerp(
Quat::from_rotation_arc(Vec3::from(self.0), Vec3::from(rhs.0)),
s,
);
Dir3A(quat.mul_vec3a(self.0))
}
/// Returns `self` after an approximate normalization, assuming the value is already nearly normalized.
/// Useful for preventing numerical error accumulation.
///
/// See [`Dir3::fast_renormalize`] for an example of when such error accumulation might occur.
#[inline]
pub fn fast_renormalize(self) -> Self {
let length_squared = self.0.length_squared();
// Based on a Taylor approximation of the inverse square root, see [`Dir3::fast_renormalize`] for more details.
Self(self * (0.5 * (3.0 - length_squared)))
}
}
impl From<Dir3> for Dir3A {
fn from(value: Dir3) -> Self {
Self(value.0.into())
}
}
impl From<Dir3A> for Dir3 {
fn from(value: Dir3A) -> Self {
Self(value.0.into())
}
}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl TryFrom<Vec3A> for Dir3A {
type Error = InvalidDirectionError;
fn try_from(value: Vec3A) -> Result<Self, Self::Error> {
Self::new(value)
}
}
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl From<Dir3A> for Vec3A {
fn from(value: Dir3A) -> Self {
value.0
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Deref for Dir3A {
type Target = Vec3A;
fn deref(&self) -> &Self::Target {
&self.0
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Neg for Dir3A {
type Output = Self;
fn neg(self) -> Self::Output {
Self(-self.0)
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Mul<f32> for Dir3A {
type Output = Vec3A;
fn mul(self, rhs: f32) -> Self::Output {
self.0 * rhs
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Mul<Dir3A> for f32 {
type Output = Vec3A;
fn mul(self, rhs: Dir3A) -> Self::Output {
self * rhs.0
}
}
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
impl core::ops::Mul<Dir3A> for Quat {
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
type Output = Dir3A;
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
/// Rotates the [`Dir3A`] using a [`Quat`].
fn mul(self, direction: Dir3A) -> Self::Output {
let rotated = self * *direction;
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
#[cfg(debug_assertions)]
assert_is_normalized(
"`Dir3A` is denormalized after rotation.",
rotated.length_squared(),
);
Improve error messages for denormalized directions (#12278) # Objective `Dir3` and `Dir3A` can be rotated using `Quat`s. However, if enough floating point error accumulates or (more commonly) the rotation itself is degenerate (like not normalized), the resulting direction can also become denormalized. Currently, with debug assertions enabled, it panics in these cases with the message `rotated.is_normalized()`. This error message is unclear, doesn't give information about *how* it is denormalized (like is the length too large, NaN, or something else), and is overall not very helpful. Panicking for small-ish error might also be a bit too strict, and has lead to unwanted crashes in crates like `bevy_xpbd` (although it has also helped in finding actual bugs). The error message should be clearer and give more context, and it shouldn't cause unwanted crashes. ## Solution Change the `debug_assert!` to a warning for small error with a (squared length) threshold of 2e-4 and a panic for clear error with a threshold of 2e-2. The warnings mention the direction type and the length of the denormalized vector. Here's what the error and warning look like: ``` Error: `Dir3` is denormalized after rotation. The length is 1.014242. ``` ``` Warning: `Dir3A` is denormalized after rotation. The length is 1.0001414. ``` I gave the same treatment to `new_unchecked`: ``` Error: The vector given to `Dir3::new_unchecked` is not normalized. The length is 1.014242. ``` ``` Warning: The vector given to `Dir3A::new_unchecked` is not normalized. The length is 1.0001414. ``` --- ## Discussion ### Threshold values The thresholds are somewhat arbitrary. 2e-4 is what Glam uses for the squared length in `is_normalized` (after I corrected it in bitshifter/glam-rs#480), and 2e-2 is just what I thought could be a clear sign of something being critically wrong. I can definitely tune them if there are better thresholds though. ### Logging `bevy_math` doesn't have `bevy_log`, so we can't use `warn!` or `error!`. This is why I made it use just `eprintln!` and `panic!` for now. Let me know if there's a better way of logging errors in `bevy_math`.
2024-03-04 00:01:32 +00:00
Dir3A(rotated)
}
}
#[cfg(feature = "approx")]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl approx::AbsDiffEq for Dir3A {
type Epsilon = f32;
fn default_epsilon() -> f32 {
f32::EPSILON
}
fn abs_diff_eq(&self, other: &Self, epsilon: f32) -> bool {
self.as_ref().abs_diff_eq(other.as_ref(), epsilon)
}
}
#[cfg(feature = "approx")]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl approx::RelativeEq for Dir3A {
fn default_max_relative() -> f32 {
f32::EPSILON
}
fn relative_eq(&self, other: &Self, epsilon: f32, max_relative: f32) -> bool {
self.as_ref()
.relative_eq(other.as_ref(), epsilon, max_relative)
}
}
#[cfg(feature = "approx")]
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
impl approx::UlpsEq for Dir3A {
fn default_max_ulps() -> u32 {
4
}
fn ulps_eq(&self, other: &Self, epsilon: f32, max_ulps: u32) -> bool {
self.as_ref().ulps_eq(other.as_ref(), epsilon, max_ulps)
}
}
#[cfg(test)]
mod tests {
Make bevy_math's `libm` feature use `libm` for all `f32`methods with unspecified precision (#14693) # Objective Closes #14474 Previously, the `libm` feature of bevy_math would just pass the same feature flag down to glam. However, bevy_math itself had many uses of floating-point arithmetic with unspecified precision. For example, `f32::sin_cos` and `f32::powi` have unspecified precision, which means that the exact details of their output are not guaranteed to be stable across different systems and/or versions of Rust. This means that users of bevy_math could observe slightly different behavior on different systems if these methods were used. The goal of this PR is to make it so that the `libm` feature flag actually guarantees some degree of determinacy within bevy_math itself by switching to the libm versions of these functions when the `libm` feature is enabled. ## Solution bevy_math now has an internal module `bevy_math::ops`, which re-exports either the standard versions of the operations or the libm versions depending on whether the `libm` feature is enabled. For example, `ops::sin` compiles to `f32::sin` without the `libm` feature and to `libm::sinf` with it. This approach has a small shortfall, which is that `f32::powi` (integer powers of floating point numbers) does not have an equivalent in `libm`. On the other hand, this method is only used for squaring and cubing numbers in bevy_math. Accordingly, this deficit is covered by the introduction of a trait `ops::FloatPow`: ```rust pub(crate) trait FloatPow { fn squared(self) -> Self; fn cubed(self) -> Self; } ``` Next, each current usage of the unspecified-precision methods has been replaced by its equivalent in `ops`, so that when `libm` is enabled, the libm version is used instead. The exception, of course, is that `.powi(2)`/`.powi(3)` have been replaced with `.squared()`/`.cubed()`. Finally, the usage of the plain `f32` methods with unspecified precision is now linted out of bevy_math (and hence disallowed in CI). For example, using `f32::sin` within bevy_math produces a warning that tells the user to use the `ops::sin` version instead. ## Testing Ran existing tests. It would be nice to check some benchmarks on NURBS things once #14677 merges. I'm happy to wait until then if the rest of this PR is fine. --- ## Discussion In the future, it might make sense to actually expose `bevy_math::ops` as public if any downstream Bevy crates want to provide similar determinacy guarantees. For now, it's all just `pub(crate)`. This PR also only covers `f32`. If we find ourselves using `f64` internally in parts of bevy_math for better robustness, we could extend the module and lints to cover the `f64` versions easily enough. I don't know how feasible it is, but it would also be nice if we could standardize the bevy_math tests with the `libm` feature in CI, since their success is currently platform-dependent (e.g. 8 of them fail on my machine when run locally). --------- Co-authored-by: IQuick 143 <IQuick143cz@gmail.com>
2024-08-12 16:13:36 +00:00
use crate::ops;
use super::*;
use approx::assert_relative_eq;
#[test]
fn dir2_creation() {
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
assert_eq!(Dir2::new(Vec2::X * 12.5), Ok(Dir2::X));
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir2::new(Vec2::new(0.0, 0.0)),
Err(InvalidDirectionError::Zero)
);
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir2::new(Vec2::new(f32::INFINITY, 0.0)),
Err(InvalidDirectionError::Infinite)
);
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir2::new(Vec2::new(f32::NEG_INFINITY, 0.0)),
Err(InvalidDirectionError::Infinite)
);
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir2::new(Vec2::new(f32::NAN, 0.0)),
Err(InvalidDirectionError::NaN)
);
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
assert_eq!(Dir2::new_and_length(Vec2::X * 6.5), Ok((Dir2::X, 6.5)));
}
#[test]
fn dir2_slerp() {
assert_relative_eq!(
Dir2::X.slerp(Dir2::Y, 0.5),
Dir2::from_xy(0.5_f32.sqrt(), 0.5_f32.sqrt()).unwrap()
);
assert_eq!(Dir2::Y.slerp(Dir2::X, 0.0), Dir2::Y);
assert_relative_eq!(Dir2::X.slerp(Dir2::Y, 1.0), Dir2::Y);
assert_relative_eq!(
Dir2::Y.slerp(Dir2::X, 1.0 / 3.0),
Dir2::from_xy(0.5, 0.75_f32.sqrt()).unwrap()
);
assert_relative_eq!(
Dir2::X.slerp(Dir2::Y, 2.0 / 3.0),
Dir2::from_xy(0.5, 0.75_f32.sqrt()).unwrap()
);
}
#[test]
fn dir2_to_rotation2d() {
assert_relative_eq!(Dir2::EAST.rotation_to(Dir2::NORTH_EAST), Rot2::FRAC_PI_4);
assert_relative_eq!(Dir2::NORTH.rotation_from(Dir2::NORTH_EAST), Rot2::FRAC_PI_4);
assert_relative_eq!(Dir2::SOUTH.rotation_to_x(), Rot2::FRAC_PI_2);
assert_relative_eq!(Dir2::SOUTH.rotation_to_y(), Rot2::PI);
assert_relative_eq!(Dir2::NORTH_WEST.rotation_from_x(), Rot2::degrees(135.0));
assert_relative_eq!(Dir2::NORTH_WEST.rotation_from_y(), Rot2::FRAC_PI_4);
}
#[test]
fn dir2_renorm() {
// Evil denormalized Rot2
Make bevy_math's `libm` feature use `libm` for all `f32`methods with unspecified precision (#14693) # Objective Closes #14474 Previously, the `libm` feature of bevy_math would just pass the same feature flag down to glam. However, bevy_math itself had many uses of floating-point arithmetic with unspecified precision. For example, `f32::sin_cos` and `f32::powi` have unspecified precision, which means that the exact details of their output are not guaranteed to be stable across different systems and/or versions of Rust. This means that users of bevy_math could observe slightly different behavior on different systems if these methods were used. The goal of this PR is to make it so that the `libm` feature flag actually guarantees some degree of determinacy within bevy_math itself by switching to the libm versions of these functions when the `libm` feature is enabled. ## Solution bevy_math now has an internal module `bevy_math::ops`, which re-exports either the standard versions of the operations or the libm versions depending on whether the `libm` feature is enabled. For example, `ops::sin` compiles to `f32::sin` without the `libm` feature and to `libm::sinf` with it. This approach has a small shortfall, which is that `f32::powi` (integer powers of floating point numbers) does not have an equivalent in `libm`. On the other hand, this method is only used for squaring and cubing numbers in bevy_math. Accordingly, this deficit is covered by the introduction of a trait `ops::FloatPow`: ```rust pub(crate) trait FloatPow { fn squared(self) -> Self; fn cubed(self) -> Self; } ``` Next, each current usage of the unspecified-precision methods has been replaced by its equivalent in `ops`, so that when `libm` is enabled, the libm version is used instead. The exception, of course, is that `.powi(2)`/`.powi(3)` have been replaced with `.squared()`/`.cubed()`. Finally, the usage of the plain `f32` methods with unspecified precision is now linted out of bevy_math (and hence disallowed in CI). For example, using `f32::sin` within bevy_math produces a warning that tells the user to use the `ops::sin` version instead. ## Testing Ran existing tests. It would be nice to check some benchmarks on NURBS things once #14677 merges. I'm happy to wait until then if the rest of this PR is fine. --- ## Discussion In the future, it might make sense to actually expose `bevy_math::ops` as public if any downstream Bevy crates want to provide similar determinacy guarantees. For now, it's all just `pub(crate)`. This PR also only covers `f32`. If we find ourselves using `f64` internally in parts of bevy_math for better robustness, we could extend the module and lints to cover the `f64` versions easily enough. I don't know how feasible it is, but it would also be nice if we could standardize the bevy_math tests with the `libm` feature in CI, since their success is currently platform-dependent (e.g. 8 of them fail on my machine when run locally). --------- Co-authored-by: IQuick 143 <IQuick143cz@gmail.com>
2024-08-12 16:13:36 +00:00
let (sin, cos) = ops::sin_cos(1.0_f32);
let rot2 = Rot2::from_sin_cos(sin * (1.0 + 1e-5), cos * (1.0 + 1e-5));
let mut dir_a = Dir2::X;
let mut dir_b = Dir2::X;
// We test that renormalizing an already normalized dir doesn't do anything
assert_relative_eq!(dir_b, dir_b.fast_renormalize(), epsilon = 0.000001);
for _ in 0..50 {
dir_a = rot2 * dir_a;
dir_b = rot2 * dir_b;
dir_b = dir_b.fast_renormalize();
}
// `dir_a` should've gotten denormalized, meanwhile `dir_b` should stay normalized.
assert!(
!dir_a.is_normalized(),
"Dernormalization doesn't work, test is faulty"
);
assert!(dir_b.is_normalized(), "Renormalisation did not work.");
}
#[test]
fn dir3_creation() {
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
assert_eq!(Dir3::new(Vec3::X * 12.5), Ok(Dir3::X));
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir3::new(Vec3::new(0.0, 0.0, 0.0)),
Err(InvalidDirectionError::Zero)
);
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir3::new(Vec3::new(f32::INFINITY, 0.0, 0.0)),
Err(InvalidDirectionError::Infinite)
);
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir3::new(Vec3::new(f32::NEG_INFINITY, 0.0, 0.0)),
Err(InvalidDirectionError::Infinite)
);
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir3::new(Vec3::new(f32::NAN, 0.0, 0.0)),
Err(InvalidDirectionError::NaN)
);
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
assert_eq!(Dir3::new_and_length(Vec3::X * 6.5), Ok((Dir3::X, 6.5)));
// Test rotation
assert!(
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
(Quat::from_rotation_z(core::f32::consts::FRAC_PI_2) * Dir3::X)
.abs_diff_eq(Vec3::Y, 10e-6)
);
}
#[test]
fn dir3_slerp() {
assert_relative_eq!(
Dir3::X.slerp(Dir3::Y, 0.5),
Dir3::from_xyz(0.5f32.sqrt(), 0.5f32.sqrt(), 0.0).unwrap()
);
assert_relative_eq!(Dir3::Y.slerp(Dir3::Z, 0.0), Dir3::Y);
assert_relative_eq!(Dir3::Z.slerp(Dir3::X, 1.0), Dir3::X, epsilon = 0.000001);
assert_relative_eq!(
Dir3::X.slerp(Dir3::Z, 1.0 / 3.0),
Dir3::from_xyz(0.75f32.sqrt(), 0.0, 0.5).unwrap(),
epsilon = 0.000001
);
assert_relative_eq!(
Dir3::Z.slerp(Dir3::Y, 2.0 / 3.0),
Dir3::from_xyz(0.0, 0.75f32.sqrt(), 0.5).unwrap()
);
}
#[test]
fn dir3_renorm() {
// Evil denormalized quaternion
let rot3 = Quat::from_euler(glam::EulerRot::XYZ, 1.0, 2.0, 3.0) * (1.0 + 1e-5);
let mut dir_a = Dir3::X;
let mut dir_b = Dir3::X;
// We test that renormalizing an already normalized dir doesn't do anything
assert_relative_eq!(dir_b, dir_b.fast_renormalize(), epsilon = 0.000001);
for _ in 0..50 {
dir_a = rot3 * dir_a;
dir_b = rot3 * dir_b;
dir_b = dir_b.fast_renormalize();
}
// `dir_a` should've gotten denormalized, meanwhile `dir_b` should stay normalized.
assert!(
!dir_a.is_normalized(),
"Dernormalization doesn't work, test is faulty"
);
assert!(dir_b.is_normalized(), "Renormalisation did not work.");
}
#[test]
fn dir3a_creation() {
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
assert_eq!(Dir3A::new(Vec3A::X * 12.5), Ok(Dir3A::X));
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir3A::new(Vec3A::new(0.0, 0.0, 0.0)),
Err(InvalidDirectionError::Zero)
);
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir3A::new(Vec3A::new(f32::INFINITY, 0.0, 0.0)),
Err(InvalidDirectionError::Infinite)
);
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir3A::new(Vec3A::new(f32::NEG_INFINITY, 0.0, 0.0)),
Err(InvalidDirectionError::Infinite)
);
assert_eq!(
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
Dir3A::new(Vec3A::new(f32::NAN, 0.0, 0.0)),
Err(InvalidDirectionError::NaN)
);
Rename `Direction2d/3d` to `Dir2/3` (#12189) # Objective Split up from #12017, rename Bevy's direction types. Currently, Bevy has the `Direction2d`, `Direction3d`, and `Direction3dA` types, which provide a type-level guarantee that their contained vectors remain normalized. They can be very useful for a lot of APIs for safety, explicitness, and in some cases performance, as they can sometimes avoid unnecessary normalizations. However, many consider them to be inconvenient to use, and opt for standard vector types like `Vec3` because of this. One reason is that the direction type names are a bit long and can be annoying to write (of course you can use autocomplete, but just typing `Vec3` is still nicer), and in some intances, the extra characters can make formatting worse. The naming is also inconsistent with Glam's shorter type names, and results in names like `Direction3dA`, which (in my opinion) are difficult to read and even a bit ugly. This PR proposes renaming the types to `Dir2`, `Dir3`, and `Dir3A`. These names are nice and easy to write, consistent with Glam, and work well for variants like the SIMD aligned `Dir3A`. As a bonus, it can also result in nicer formatting in a lot of cases, which can be seen from the diff of this PR. Some examples of what it looks like: (copied from #12017) ```rust // Before let ray_cast = RayCast2d::new(Vec2::ZERO, Direction2d::X, 5.0); // After let ray_cast = RayCast2d::new(Vec2::ZERO, Dir2::X, 5.0); ``` ```rust // Before (an example using Bevy XPBD) let hit = spatial_query.cast_ray( Vec3::ZERO, Direction3d::X, f32::MAX, true, SpatialQueryFilter::default(), ); // After let hit = spatial_query.cast_ray( Vec3::ZERO, Dir3::X, f32::MAX, true, SpatialQueryFilter::default(), ); ``` ```rust // Before self.circle( Vec3::new(0.0, -2.0, 0.0), Direction3d::Y, 5.0, Color::TURQUOISE, ); // After (formatting is collapsed in this case) self.circle(Vec3::new(0.0, -2.0, 0.0), Dir3::Y, 5.0, Color::TURQUOISE); ``` ## Solution Rename `Direction2d`, `Direction3d`, and `Direction3dA` to `Dir2`, `Dir3`, and `Dir3A`. --- ## Migration Guide The `Direction2d` and `Direction3d` types have been renamed to `Dir2` and `Dir3`. ## Additional Context This has been brought up on the Discord a few times, and we had a small [poll](https://discord.com/channels/691052431525675048/1203087353850364004/1212465038711984158) on this. `Dir2`/`Dir3`/`Dir3A` was quite unanimously chosen as the best option, but of course it was a very small poll and inconclusive, so other opinions are certainly welcome too. --------- Co-authored-by: IceSentry <c.giguere42@gmail.com>
2024-02-28 22:48:43 +00:00
assert_eq!(Dir3A::new_and_length(Vec3A::X * 6.5), Ok((Dir3A::X, 6.5)));
// Test rotation
assert!(
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
(Quat::from_rotation_z(core::f32::consts::FRAC_PI_2) * Dir3A::X)
.abs_diff_eq(Vec3A::Y, 10e-6)
);
}
#[test]
fn dir3a_slerp() {
assert_relative_eq!(
Dir3A::X.slerp(Dir3A::Y, 0.5),
Dir3A::from_xyz(0.5f32.sqrt(), 0.5f32.sqrt(), 0.0).unwrap()
);
assert_relative_eq!(Dir3A::Y.slerp(Dir3A::Z, 0.0), Dir3A::Y);
assert_relative_eq!(Dir3A::Z.slerp(Dir3A::X, 1.0), Dir3A::X, epsilon = 0.000001);
assert_relative_eq!(
Dir3A::X.slerp(Dir3A::Z, 1.0 / 3.0),
Dir3A::from_xyz(0.75f32.sqrt(), 0.0, 0.5).unwrap(),
epsilon = 0.000001
);
assert_relative_eq!(
Dir3A::Z.slerp(Dir3A::Y, 2.0 / 3.0),
Dir3A::from_xyz(0.0, 0.75f32.sqrt(), 0.5).unwrap()
);
}
#[test]
fn dir3a_renorm() {
// Evil denormalized quaternion
let rot3 = Quat::from_euler(glam::EulerRot::XYZ, 1.0, 2.0, 3.0) * (1.0 + 1e-5);
let mut dir_a = Dir3A::X;
let mut dir_b = Dir3A::X;
// We test that renormalizing an already normalized dir doesn't do anything
assert_relative_eq!(dir_b, dir_b.fast_renormalize(), epsilon = 0.000001);
for _ in 0..50 {
dir_a = rot3 * dir_a;
dir_b = rot3 * dir_b;
dir_b = dir_b.fast_renormalize();
}
// `dir_a` should've gotten denormalized, meanwhile `dir_b` should stay normalized.
assert!(
!dir_a.is_normalized(),
"Dernormalization doesn't work, test is faulty"
);
assert!(dir_b.is_normalized(), "Renormalisation did not work.");
}
}