
# Objective - Fixes #17960 ## Solution - Followed the [edition upgrade guide](https://doc.rust-lang.org/edition-guide/editions/transitioning-an-existing-project-to-a-new-edition.html) ## Testing - CI --- ## Summary of Changes ### Documentation Indentation When using lists in documentation, proper indentation is now linted for. This means subsequent lines within the same list item must start at the same indentation level as the item. ```rust /* Valid */ /// - Item 1 /// Run-on sentence. /// - Item 2 struct Foo; /* Invalid */ /// - Item 1 /// Run-on sentence. /// - Item 2 struct Foo; ``` ### Implicit `!` to `()` Conversion `!` (the never return type, returned by `panic!`, etc.) no longer implicitly converts to `()`. This is particularly painful for systems with `todo!` or `panic!` statements, as they will no longer be functions returning `()` (or `Result<()>`), making them invalid systems for functions like `add_systems`. The ideal fix would be to accept functions returning `!` (or rather, _not_ returning), but this is blocked on the [stabilisation of the `!` type itself](https://doc.rust-lang.org/std/primitive.never.html), which is not done. The "simple" fix would be to add an explicit `-> ()` to system signatures (e.g., `|| { todo!() }` becomes `|| -> () { todo!() }`). However, this is _also_ banned, as there is an existing lint which (IMO, incorrectly) marks this as an unnecessary annotation. So, the "fix" (read: workaround) is to put these kinds of `|| -> ! { ... }` closuers into variables and give the variable an explicit type (e.g., `fn()`). ```rust // Valid let system: fn() = || todo!("Not implemented yet!"); app.add_systems(..., system); // Invalid app.add_systems(..., || todo!("Not implemented yet!")); ``` ### Temporary Variable Lifetimes The order in which temporary variables are dropped has changed. The simple fix here is _usually_ to just assign temporaries to a named variable before use. ### `gen` is a keyword We can no longer use the name `gen` as it is reserved for a future generator syntax. This involved replacing uses of the name `gen` with `r#gen` (the raw-identifier syntax). ### Formatting has changed Use statements have had the order of imports changed, causing a substantial +/-3,000 diff when applied. For now, I have opted-out of this change by amending `rustfmt.toml` ```toml style_edition = "2021" ``` This preserves the original formatting for now, reducing the size of this PR. It would be a simple followup to update this to 2024 and run `cargo fmt`. ### New `use<>` Opt-Out Syntax Lifetimes are now implicitly included in RPIT types. There was a handful of instances where it needed to be added to satisfy the borrow checker, but there may be more cases where it _should_ be added to avoid breakages in user code. ### `MyUnitStruct { .. }` is an invalid pattern Previously, you could match against unit structs (and unit enum variants) with a `{ .. }` destructuring. This is no longer valid. ### Pretty much every use of `ref` and `mut` are gone Pattern binding has changed to the point where these terms are largely unused now. They still serve a purpose, but it is far more niche now. ### `iter::repeat(...).take(...)` is bad New lint recommends using the more explicit `iter::repeat_n(..., ...)` instead. ## Migration Guide The lifetimes of functions using return-position impl-trait (RPIT) are likely _more_ conservative than they had been previously. If you encounter lifetime issues with such a function, please create an issue to investigate the addition of `+ use<...>`. ## Notes - Check the individual commits for a clearer breakdown for what _actually_ changed. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2138 lines
86 KiB
Rust
2138 lines
86 KiB
Rust
use crate::{
|
|
archetype::{Archetype, ArchetypeComponentId, ArchetypeGeneration, ArchetypeId},
|
|
component::{ComponentId, Tick},
|
|
entity::{Entity, EntityBorrow, EntitySet},
|
|
entity_disabling::DefaultQueryFilters,
|
|
prelude::FromWorld,
|
|
query::{Access, FilteredAccess, QueryCombinationIter, QueryIter, QueryParIter, WorldQuery},
|
|
storage::{SparseSetIndex, TableId},
|
|
system::Query,
|
|
world::{unsafe_world_cell::UnsafeWorldCell, World, WorldId},
|
|
};
|
|
|
|
#[cfg(all(not(target_arch = "wasm32"), feature = "multi_threaded"))]
|
|
use crate::entity::{TrustedEntityBorrow, UniqueEntitySlice};
|
|
|
|
use alloc::vec::Vec;
|
|
use core::{fmt, ptr};
|
|
use fixedbitset::FixedBitSet;
|
|
use log::warn;
|
|
#[cfg(feature = "trace")]
|
|
use tracing::Span;
|
|
|
|
use super::{
|
|
NopWorldQuery, QueryBuilder, QueryData, QueryEntityError, QueryFilter, QueryManyIter,
|
|
QueryManyUniqueIter, QuerySingleError, ROQueryItem, ReadOnlyQueryData,
|
|
};
|
|
|
|
/// An ID for either a table or an archetype. Used for Query iteration.
|
|
///
|
|
/// Query iteration is exclusively dense (over tables) or archetypal (over archetypes) based on whether
|
|
/// the query filters are dense or not. This is represented by the [`QueryState::is_dense`] field.
|
|
///
|
|
/// Note that `D::IS_DENSE` and `F::IS_DENSE` have no relationship with `QueryState::is_dense` and
|
|
/// any combination of their values can happen.
|
|
///
|
|
/// This is a union instead of an enum as the usage is determined at compile time, as all [`StorageId`]s for
|
|
/// a [`QueryState`] will be all [`TableId`]s or all [`ArchetypeId`]s, and not a mixture of both. This
|
|
/// removes the need for discriminator to minimize memory usage and branching during iteration, but requires
|
|
/// a safety invariant be verified when disambiguating them.
|
|
///
|
|
/// # Safety
|
|
/// Must be initialized and accessed as a [`TableId`], if both generic parameters to the query are dense.
|
|
/// Must be initialized and accessed as an [`ArchetypeId`] otherwise.
|
|
#[derive(Clone, Copy)]
|
|
pub(super) union StorageId {
|
|
pub(super) table_id: TableId,
|
|
pub(super) archetype_id: ArchetypeId,
|
|
}
|
|
|
|
/// Provides scoped access to a [`World`] state according to a given [`QueryData`] and [`QueryFilter`].
|
|
///
|
|
/// This data is cached between system runs, and is used to:
|
|
/// - store metadata about which [`Table`] or [`Archetype`] are matched by the query. "Matched" means
|
|
/// that the query will iterate over the data in the matched table/archetype.
|
|
/// - cache the [`State`] needed to compute the [`Fetch`] struct used to retrieve data
|
|
/// from a specific [`Table`] or [`Archetype`]
|
|
/// - build iterators that can iterate over the query results
|
|
///
|
|
/// [`State`]: crate::query::world_query::WorldQuery::State
|
|
/// [`Fetch`]: crate::query::world_query::WorldQuery::Fetch
|
|
/// [`Table`]: crate::storage::Table
|
|
#[repr(C)]
|
|
// SAFETY NOTE:
|
|
// Do not add any new fields that use the `D` or `F` generic parameters as this may
|
|
// make `QueryState::as_transmuted_state` unsound if not done with care.
|
|
pub struct QueryState<D: QueryData, F: QueryFilter = ()> {
|
|
world_id: WorldId,
|
|
pub(crate) archetype_generation: ArchetypeGeneration,
|
|
/// Metadata about the [`Table`](crate::storage::Table)s matched by this query.
|
|
pub(crate) matched_tables: FixedBitSet,
|
|
/// Metadata about the [`Archetype`]s matched by this query.
|
|
pub(crate) matched_archetypes: FixedBitSet,
|
|
/// [`FilteredAccess`] computed by combining the `D` and `F` access. Used to check which other queries
|
|
/// this query can run in parallel with.
|
|
pub(crate) component_access: FilteredAccess<ComponentId>,
|
|
// NOTE: we maintain both a bitset and a vec because iterating the vec is faster
|
|
pub(super) matched_storage_ids: Vec<StorageId>,
|
|
// Represents whether this query iteration is dense or not. When this is true
|
|
// `matched_storage_ids` stores `TableId`s, otherwise it stores `ArchetypeId`s.
|
|
pub(super) is_dense: bool,
|
|
pub(crate) fetch_state: D::State,
|
|
pub(crate) filter_state: F::State,
|
|
#[cfg(feature = "trace")]
|
|
par_iter_span: Span,
|
|
}
|
|
|
|
impl<D: QueryData, F: QueryFilter> fmt::Debug for QueryState<D, F> {
|
|
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
|
|
f.debug_struct("QueryState")
|
|
.field("world_id", &self.world_id)
|
|
.field("matched_table_count", &self.matched_tables.count_ones(..))
|
|
.field(
|
|
"matched_archetype_count",
|
|
&self.matched_archetypes.count_ones(..),
|
|
)
|
|
.finish_non_exhaustive()
|
|
}
|
|
}
|
|
|
|
impl<D: QueryData, F: QueryFilter> FromWorld for QueryState<D, F> {
|
|
fn from_world(world: &mut World) -> Self {
|
|
world.query_filtered()
|
|
}
|
|
}
|
|
|
|
impl<D: QueryData, F: QueryFilter> QueryState<D, F> {
|
|
/// Converts this `QueryState` reference to a `QueryState` that does not access anything mutably.
|
|
pub fn as_readonly(&self) -> &QueryState<D::ReadOnly, F> {
|
|
// SAFETY: invariant on `WorldQuery` trait upholds that `D::ReadOnly` and `F::ReadOnly`
|
|
// have a subset of the access, and match the exact same archetypes/tables as `D`/`F` respectively.
|
|
unsafe { self.as_transmuted_state::<D::ReadOnly, F>() }
|
|
}
|
|
|
|
/// Converts this `QueryState` reference to a `QueryState` that does not return any data
|
|
/// which can be faster.
|
|
///
|
|
/// This doesn't use `NopWorldQuery` as it loses filter functionality, for example
|
|
/// `NopWorldQuery<Changed<T>>` is functionally equivalent to `With<T>`.
|
|
pub(crate) fn as_nop(&self) -> &QueryState<NopWorldQuery<D>, F> {
|
|
// SAFETY: `NopWorldQuery` doesn't have any accesses and defers to
|
|
// `D` for table/archetype matching
|
|
unsafe { self.as_transmuted_state::<NopWorldQuery<D>, F>() }
|
|
}
|
|
|
|
/// Converts this `QueryState` reference to any other `QueryState` with
|
|
/// the same `WorldQuery::State` associated types.
|
|
///
|
|
/// Consider using `as_readonly` or `as_nop` instead which are safe functions.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// `NewD` must have a subset of the access that `D` does and match the exact same archetypes/tables
|
|
/// `NewF` must have a subset of the access that `F` does and match the exact same archetypes/tables
|
|
pub(crate) unsafe fn as_transmuted_state<
|
|
NewD: QueryData<State = D::State>,
|
|
NewF: QueryFilter<State = F::State>,
|
|
>(
|
|
&self,
|
|
) -> &QueryState<NewD, NewF> {
|
|
&*ptr::from_ref(self).cast::<QueryState<NewD, NewF>>()
|
|
}
|
|
|
|
/// Returns the components accessed by this query.
|
|
pub fn component_access(&self) -> &FilteredAccess<ComponentId> {
|
|
&self.component_access
|
|
}
|
|
|
|
/// Returns the tables matched by this query.
|
|
pub fn matched_tables(&self) -> impl Iterator<Item = TableId> + '_ {
|
|
self.matched_tables.ones().map(TableId::from_usize)
|
|
}
|
|
|
|
/// Returns the archetypes matched by this query.
|
|
pub fn matched_archetypes(&self) -> impl Iterator<Item = ArchetypeId> + '_ {
|
|
self.matched_archetypes.ones().map(ArchetypeId::new)
|
|
}
|
|
|
|
/// Creates a new [`QueryState`] from a given [`World`] and inherits the result of `world.id()`.
|
|
pub fn new(world: &mut World) -> Self {
|
|
let mut state = Self::new_uninitialized(world);
|
|
state.update_archetypes(world);
|
|
state
|
|
}
|
|
|
|
/// Creates a new [`QueryState`] from an immutable [`World`] reference and inherits the result of `world.id()`.
|
|
///
|
|
/// This function may fail if, for example,
|
|
/// the components that make up this query have not been registered into the world.
|
|
pub fn try_new(world: &World) -> Option<Self> {
|
|
let mut state = Self::try_new_uninitialized(world)?;
|
|
state.update_archetypes(world);
|
|
Some(state)
|
|
}
|
|
|
|
/// Identical to `new`, but it populates the provided `access` with the matched results.
|
|
pub(crate) fn new_with_access(
|
|
world: &mut World,
|
|
access: &mut Access<ArchetypeComponentId>,
|
|
) -> Self {
|
|
let mut state = Self::new_uninitialized(world);
|
|
for archetype in world.archetypes.iter() {
|
|
// SAFETY: The state was just initialized from the `world` above, and the archetypes being added
|
|
// come directly from the same world.
|
|
unsafe {
|
|
if state.new_archetype_internal(archetype) {
|
|
state.update_archetype_component_access(archetype, access);
|
|
}
|
|
}
|
|
}
|
|
state.archetype_generation = world.archetypes.generation();
|
|
|
|
// Resource access is not part of any archetype and must be handled separately
|
|
if state.component_access.access().has_read_all_resources() {
|
|
access.read_all_resources();
|
|
} else {
|
|
for component_id in state.component_access.access().resource_reads() {
|
|
access.add_resource_read(world.initialize_resource_internal(component_id).id());
|
|
}
|
|
}
|
|
|
|
debug_assert!(
|
|
!state.component_access.access().has_any_resource_write(),
|
|
"Mutable resource access in queries is not allowed"
|
|
);
|
|
|
|
state
|
|
}
|
|
|
|
/// Creates a new [`QueryState`] but does not populate it with the matched results from the World yet
|
|
///
|
|
/// `new_archetype` and its variants must be called on all of the World's archetypes before the
|
|
/// state can return valid query results.
|
|
fn new_uninitialized(world: &mut World) -> Self {
|
|
let fetch_state = D::init_state(world);
|
|
let filter_state = F::init_state(world);
|
|
Self::from_states_uninitialized(world, fetch_state, filter_state)
|
|
}
|
|
|
|
/// Creates a new [`QueryState`] but does not populate it with the matched results from the World yet
|
|
///
|
|
/// `new_archetype` and its variants must be called on all of the World's archetypes before the
|
|
/// state can return valid query results.
|
|
fn try_new_uninitialized(world: &World) -> Option<Self> {
|
|
let fetch_state = D::get_state(world.components())?;
|
|
let filter_state = F::get_state(world.components())?;
|
|
Some(Self::from_states_uninitialized(
|
|
world,
|
|
fetch_state,
|
|
filter_state,
|
|
))
|
|
}
|
|
|
|
/// Creates a new [`QueryState`] but does not populate it with the matched results from the World yet
|
|
///
|
|
/// `new_archetype` and its variants must be called on all of the World's archetypes before the
|
|
/// state can return valid query results.
|
|
fn from_states_uninitialized(
|
|
world: &World,
|
|
fetch_state: <D as WorldQuery>::State,
|
|
filter_state: <F as WorldQuery>::State,
|
|
) -> Self {
|
|
let mut component_access = FilteredAccess::default();
|
|
D::update_component_access(&fetch_state, &mut component_access);
|
|
|
|
// Use a temporary empty FilteredAccess for filters. This prevents them from conflicting with the
|
|
// main Query's `fetch_state` access. Filters are allowed to conflict with the main query fetch
|
|
// because they are evaluated *before* a specific reference is constructed.
|
|
let mut filter_component_access = FilteredAccess::default();
|
|
F::update_component_access(&filter_state, &mut filter_component_access);
|
|
|
|
// Merge the temporary filter access with the main access. This ensures that filter access is
|
|
// properly considered in a global "cross-query" context (both within systems and across systems).
|
|
component_access.extend(&filter_component_access);
|
|
|
|
// For queries without dynamic filters the dense-ness of the query is equal to the dense-ness
|
|
// of its static type parameters.
|
|
let mut is_dense = D::IS_DENSE && F::IS_DENSE;
|
|
|
|
if let Some(default_filters) = world.get_resource::<DefaultQueryFilters>() {
|
|
default_filters.modify_access(&mut component_access);
|
|
is_dense &= default_filters.is_dense(world.components());
|
|
}
|
|
|
|
Self {
|
|
world_id: world.id(),
|
|
archetype_generation: ArchetypeGeneration::initial(),
|
|
matched_storage_ids: Vec::new(),
|
|
is_dense,
|
|
fetch_state,
|
|
filter_state,
|
|
component_access,
|
|
matched_tables: Default::default(),
|
|
matched_archetypes: Default::default(),
|
|
#[cfg(feature = "trace")]
|
|
par_iter_span: tracing::info_span!(
|
|
"par_for_each",
|
|
query = core::any::type_name::<D>(),
|
|
filter = core::any::type_name::<F>(),
|
|
),
|
|
}
|
|
}
|
|
|
|
/// Creates a new [`QueryState`] from a given [`QueryBuilder`] and inherits its [`FilteredAccess`].
|
|
pub fn from_builder(builder: &mut QueryBuilder<D, F>) -> Self {
|
|
let mut fetch_state = D::init_state(builder.world_mut());
|
|
let filter_state = F::init_state(builder.world_mut());
|
|
D::set_access(&mut fetch_state, builder.access());
|
|
|
|
let mut component_access = builder.access().clone();
|
|
|
|
// For dynamic queries the dense-ness is given by the query builder.
|
|
let mut is_dense = builder.is_dense();
|
|
|
|
if let Some(default_filters) = builder.world().get_resource::<DefaultQueryFilters>() {
|
|
default_filters.modify_access(&mut component_access);
|
|
is_dense &= default_filters.is_dense(builder.world().components());
|
|
}
|
|
|
|
let mut state = Self {
|
|
world_id: builder.world().id(),
|
|
archetype_generation: ArchetypeGeneration::initial(),
|
|
matched_storage_ids: Vec::new(),
|
|
is_dense,
|
|
fetch_state,
|
|
filter_state,
|
|
component_access,
|
|
matched_tables: Default::default(),
|
|
matched_archetypes: Default::default(),
|
|
#[cfg(feature = "trace")]
|
|
par_iter_span: tracing::info_span!(
|
|
"par_for_each",
|
|
data = core::any::type_name::<D>(),
|
|
filter = core::any::type_name::<F>(),
|
|
),
|
|
};
|
|
state.update_archetypes(builder.world());
|
|
state
|
|
}
|
|
|
|
/// Creates a [`Query`] from the given [`QueryState`] and [`World`].
|
|
///
|
|
/// This will create read-only queries, see [`Self::query_mut`] for mutable queries.
|
|
pub fn query<'w, 's>(&'s mut self, world: &'w World) -> Query<'w, 's, D::ReadOnly, F> {
|
|
self.update_archetypes(world);
|
|
self.query_manual(world)
|
|
}
|
|
|
|
/// Creates a [`Query`] from the given [`QueryState`] and [`World`].
|
|
///
|
|
/// This method is slightly more efficient than [`QueryState::query`] in some situations, since
|
|
/// it does not update this instance's internal cache. The resulting query may skip an entity that
|
|
/// belongs to an archetype that has not been cached.
|
|
///
|
|
/// To ensure that the cache is up to date, call [`QueryState::update_archetypes`] before this method.
|
|
/// The cache is also updated in [`QueryState::new`], [`QueryState::get`], or any method with mutable
|
|
/// access to `self`.
|
|
///
|
|
/// This will create read-only queries, see [`Self::query_mut`] for mutable queries.
|
|
pub fn query_manual<'w, 's>(&'s self, world: &'w World) -> Query<'w, 's, D::ReadOnly, F> {
|
|
self.validate_world(world.id());
|
|
// SAFETY:
|
|
// - We have read access to the entire world, and we call `as_readonly()` so the query only performs read access.
|
|
// - We called `validate_world`.
|
|
unsafe {
|
|
self.as_readonly()
|
|
.query_unchecked_manual(world.as_unsafe_world_cell_readonly())
|
|
}
|
|
}
|
|
|
|
/// Creates a [`Query`] from the given [`QueryState`] and [`World`].
|
|
pub fn query_mut<'w, 's>(&'s mut self, world: &'w mut World) -> Query<'w, 's, D, F> {
|
|
let last_run = world.last_change_tick();
|
|
let this_run = world.change_tick();
|
|
// SAFETY: We have exclusive access to the entire world.
|
|
unsafe { self.query_unchecked_with_ticks(world.as_unsafe_world_cell(), last_run, this_run) }
|
|
}
|
|
|
|
/// Creates a [`Query`] from the given [`QueryState`] and [`World`].
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
pub unsafe fn query_unchecked<'w, 's>(
|
|
&'s mut self,
|
|
world: UnsafeWorldCell<'w>,
|
|
) -> Query<'w, 's, D, F> {
|
|
self.update_archetypes_unsafe_world_cell(world);
|
|
// SAFETY: Caller ensures we have the required access
|
|
unsafe { self.query_unchecked_manual(world) }
|
|
}
|
|
|
|
/// Creates a [`Query`] from the given [`QueryState`] and [`World`].
|
|
///
|
|
/// This method is slightly more efficient than [`QueryState::query_unchecked`] in some situations, since
|
|
/// it does not update this instance's internal cache. The resulting query may skip an entity that
|
|
/// belongs to an archetype that has not been cached.
|
|
///
|
|
/// To ensure that the cache is up to date, call [`QueryState::update_archetypes`] before this method.
|
|
/// The cache is also updated in [`QueryState::new`], [`QueryState::get`], or any method with mutable
|
|
/// access to `self`.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
/// This does not validate that `world.id()` matches `self.world_id`. Calling this on a `world`
|
|
/// with a mismatched [`WorldId`] is unsound.
|
|
pub unsafe fn query_unchecked_manual<'w, 's>(
|
|
&'s self,
|
|
world: UnsafeWorldCell<'w>,
|
|
) -> Query<'w, 's, D, F> {
|
|
let last_run = world.last_change_tick();
|
|
let this_run = world.change_tick();
|
|
// SAFETY:
|
|
// - The caller ensured we have the correct access to the world.
|
|
// - The caller ensured that the world matches.
|
|
unsafe { self.query_unchecked_manual_with_ticks(world, last_run, this_run) }
|
|
}
|
|
|
|
/// Creates a [`Query`] from the given [`QueryState`] and [`World`].
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
pub unsafe fn query_unchecked_with_ticks<'w, 's>(
|
|
&'s mut self,
|
|
world: UnsafeWorldCell<'w>,
|
|
last_run: Tick,
|
|
this_run: Tick,
|
|
) -> Query<'w, 's, D, F> {
|
|
self.update_archetypes_unsafe_world_cell(world);
|
|
// SAFETY:
|
|
// - The caller ensured we have the correct access to the world.
|
|
// - We called `update_archetypes_unsafe_world_cell`, which calls `validate_world`.
|
|
unsafe { self.query_unchecked_manual_with_ticks(world, last_run, this_run) }
|
|
}
|
|
|
|
/// Creates a [`Query`] from the given [`QueryState`] and [`World`].
|
|
///
|
|
/// This method is slightly more efficient than [`QueryState::query_unchecked_with_ticks`] in some situations, since
|
|
/// it does not update this instance's internal cache. The resulting query may skip an entity that
|
|
/// belongs to an archetype that has not been cached.
|
|
///
|
|
/// To ensure that the cache is up to date, call [`QueryState::update_archetypes`] before this method.
|
|
/// The cache is also updated in [`QueryState::new`], [`QueryState::get`], or any method with mutable
|
|
/// access to `self`.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
/// This does not validate that `world.id()` matches `self.world_id`. Calling this on a `world`
|
|
/// with a mismatched [`WorldId`] is unsound.
|
|
pub unsafe fn query_unchecked_manual_with_ticks<'w, 's>(
|
|
&'s self,
|
|
world: UnsafeWorldCell<'w>,
|
|
last_run: Tick,
|
|
this_run: Tick,
|
|
) -> Query<'w, 's, D, F> {
|
|
// SAFETY:
|
|
// - The caller ensured we have the correct access to the world.
|
|
// - The caller ensured that the world matches.
|
|
unsafe { Query::new(world, self, last_run, this_run) }
|
|
}
|
|
|
|
/// Checks if the query is empty for the given [`World`], where the last change and current tick are given.
|
|
///
|
|
/// This is equivalent to `self.iter().next().is_none()`, and thus the worst case runtime will be `O(n)`
|
|
/// where `n` is the number of *potential* matches. This can be notably expensive for queries that rely
|
|
/// on non-archetypal filters such as [`Added`] or [`Changed`] which must individually check each query
|
|
/// result for a match.
|
|
///
|
|
/// # Panics
|
|
///
|
|
/// If `world` does not match the one used to call `QueryState::new` for this instance.
|
|
///
|
|
/// [`Added`]: crate::query::Added
|
|
/// [`Changed`]: crate::query::Changed
|
|
#[inline]
|
|
pub fn is_empty(&self, world: &World, last_run: Tick, this_run: Tick) -> bool {
|
|
self.validate_world(world.id());
|
|
// SAFETY:
|
|
// - We have read access to the entire world, and `is_empty()` only performs read access.
|
|
// - We called `validate_world`.
|
|
unsafe {
|
|
self.query_unchecked_manual_with_ticks(
|
|
world.as_unsafe_world_cell_readonly(),
|
|
last_run,
|
|
this_run,
|
|
)
|
|
}
|
|
.is_empty()
|
|
}
|
|
|
|
/// Returns `true` if the given [`Entity`] matches the query.
|
|
///
|
|
/// This is always guaranteed to run in `O(1)` time.
|
|
#[inline]
|
|
pub fn contains(&self, entity: Entity, world: &World, last_run: Tick, this_run: Tick) -> bool {
|
|
self.validate_world(world.id());
|
|
// SAFETY:
|
|
// - We have read access to the entire world, and `is_empty()` only performs read access.
|
|
// - We called `validate_world`.
|
|
unsafe {
|
|
self.query_unchecked_manual_with_ticks(
|
|
world.as_unsafe_world_cell_readonly(),
|
|
last_run,
|
|
this_run,
|
|
)
|
|
}
|
|
.contains(entity)
|
|
}
|
|
|
|
/// Updates the state's internal view of the [`World`]'s archetypes. If this is not called before querying data,
|
|
/// the results may not accurately reflect what is in the `world`.
|
|
///
|
|
/// This is only required if a `manual` method (such as [`Self::get_manual`]) is being called, and it only needs to
|
|
/// be called if the `world` has been structurally mutated (i.e. added/removed a component or resource). Users using
|
|
/// non-`manual` methods such as [`QueryState::get`] do not need to call this as it will be automatically called for them.
|
|
///
|
|
/// If you have an [`UnsafeWorldCell`] instead of `&World`, consider using [`QueryState::update_archetypes_unsafe_world_cell`].
|
|
///
|
|
/// # Panics
|
|
///
|
|
/// If `world` does not match the one used to call `QueryState::new` for this instance.
|
|
#[inline]
|
|
pub fn update_archetypes(&mut self, world: &World) {
|
|
self.update_archetypes_unsafe_world_cell(world.as_unsafe_world_cell_readonly());
|
|
}
|
|
|
|
/// Updates the state's internal view of the `world`'s archetypes. If this is not called before querying data,
|
|
/// the results may not accurately reflect what is in the `world`.
|
|
///
|
|
/// This is only required if a `manual` method (such as [`Self::get_manual`]) is being called, and it only needs to
|
|
/// be called if the `world` has been structurally mutated (i.e. added/removed a component or resource). Users using
|
|
/// non-`manual` methods such as [`QueryState::get`] do not need to call this as it will be automatically called for them.
|
|
///
|
|
/// # Note
|
|
///
|
|
/// This method only accesses world metadata.
|
|
///
|
|
/// # Panics
|
|
///
|
|
/// If `world` does not match the one used to call `QueryState::new` for this instance.
|
|
pub fn update_archetypes_unsafe_world_cell(&mut self, world: UnsafeWorldCell) {
|
|
self.validate_world(world.id());
|
|
if self.component_access.required.is_empty() {
|
|
let archetypes = world.archetypes();
|
|
let old_generation =
|
|
core::mem::replace(&mut self.archetype_generation, archetypes.generation());
|
|
|
|
for archetype in &archetypes[old_generation..] {
|
|
// SAFETY: The validate_world call ensures that the world is the same the QueryState
|
|
// was initialized from.
|
|
unsafe {
|
|
self.new_archetype_internal(archetype);
|
|
}
|
|
}
|
|
} else {
|
|
// skip if we are already up to date
|
|
if self.archetype_generation == world.archetypes().generation() {
|
|
return;
|
|
}
|
|
// if there are required components, we can optimize by only iterating through archetypes
|
|
// that contain at least one of the required components
|
|
let potential_archetypes = self
|
|
.component_access
|
|
.required
|
|
.ones()
|
|
.filter_map(|idx| {
|
|
let component_id = ComponentId::get_sparse_set_index(idx);
|
|
world
|
|
.archetypes()
|
|
.component_index()
|
|
.get(&component_id)
|
|
.map(|index| index.keys())
|
|
})
|
|
// select the component with the fewest archetypes
|
|
.min_by_key(ExactSizeIterator::len);
|
|
if let Some(archetypes) = potential_archetypes {
|
|
for archetype_id in archetypes {
|
|
// exclude archetypes that have already been processed
|
|
if archetype_id < &self.archetype_generation.0 {
|
|
continue;
|
|
}
|
|
// SAFETY: get_potential_archetypes only returns archetype ids that are valid for the world
|
|
let archetype = &world.archetypes()[*archetype_id];
|
|
// SAFETY: The validate_world call ensures that the world is the same the QueryState
|
|
// was initialized from.
|
|
unsafe {
|
|
self.new_archetype_internal(archetype);
|
|
}
|
|
}
|
|
}
|
|
self.archetype_generation = world.archetypes().generation();
|
|
}
|
|
}
|
|
|
|
/// # Panics
|
|
///
|
|
/// If `world_id` does not match the [`World`] used to call `QueryState::new` for this instance.
|
|
///
|
|
/// Many unsafe query methods require the world to match for soundness. This function is the easiest
|
|
/// way of ensuring that it matches.
|
|
#[inline]
|
|
#[track_caller]
|
|
pub fn validate_world(&self, world_id: WorldId) {
|
|
#[inline(never)]
|
|
#[track_caller]
|
|
#[cold]
|
|
fn panic_mismatched(this: WorldId, other: WorldId) -> ! {
|
|
panic!("Encountered a mismatched World. This QueryState was created from {this:?}, but a method was called using {other:?}.");
|
|
}
|
|
|
|
if self.world_id != world_id {
|
|
panic_mismatched(self.world_id, world_id);
|
|
}
|
|
}
|
|
|
|
/// Update the current [`QueryState`] with information from the provided [`Archetype`]
|
|
/// (if applicable, i.e. if the archetype has any intersecting [`ComponentId`] with the current [`QueryState`]).
|
|
///
|
|
/// The passed in `access` will be updated with any new accesses introduced by the new archetype.
|
|
///
|
|
/// # Safety
|
|
/// `archetype` must be from the `World` this state was initialized from.
|
|
pub unsafe fn new_archetype(
|
|
&mut self,
|
|
archetype: &Archetype,
|
|
access: &mut Access<ArchetypeComponentId>,
|
|
) {
|
|
// SAFETY: The caller ensures that `archetype` is from the World the state was initialized from.
|
|
let matches = unsafe { self.new_archetype_internal(archetype) };
|
|
if matches {
|
|
// SAFETY: The caller ensures that `archetype` is from the World the state was initialized from.
|
|
unsafe { self.update_archetype_component_access(archetype, access) };
|
|
}
|
|
}
|
|
|
|
/// Process the given [`Archetype`] to update internal metadata about the [`Table`](crate::storage::Table)s
|
|
/// and [`Archetype`]s that are matched by this query.
|
|
///
|
|
/// Returns `true` if the given `archetype` matches the query. Otherwise, returns `false`.
|
|
/// If there is no match, then there is no need to update the query's [`FilteredAccess`].
|
|
///
|
|
/// # Safety
|
|
/// `archetype` must be from the `World` this state was initialized from.
|
|
unsafe fn new_archetype_internal(&mut self, archetype: &Archetype) -> bool {
|
|
if D::matches_component_set(&self.fetch_state, &|id| archetype.contains(id))
|
|
&& F::matches_component_set(&self.filter_state, &|id| archetype.contains(id))
|
|
&& self.matches_component_set(&|id| archetype.contains(id))
|
|
{
|
|
let archetype_index = archetype.id().index();
|
|
if !self.matched_archetypes.contains(archetype_index) {
|
|
self.matched_archetypes.grow_and_insert(archetype_index);
|
|
if !self.is_dense {
|
|
self.matched_storage_ids.push(StorageId {
|
|
archetype_id: archetype.id(),
|
|
});
|
|
}
|
|
}
|
|
let table_index = archetype.table_id().as_usize();
|
|
if !self.matched_tables.contains(table_index) {
|
|
self.matched_tables.grow_and_insert(table_index);
|
|
if self.is_dense {
|
|
self.matched_storage_ids.push(StorageId {
|
|
table_id: archetype.table_id(),
|
|
});
|
|
}
|
|
}
|
|
true
|
|
} else {
|
|
false
|
|
}
|
|
}
|
|
|
|
/// Returns `true` if this query matches a set of components. Otherwise, returns `false`.
|
|
pub fn matches_component_set(&self, set_contains_id: &impl Fn(ComponentId) -> bool) -> bool {
|
|
self.component_access.filter_sets.iter().any(|set| {
|
|
set.with
|
|
.ones()
|
|
.all(|index| set_contains_id(ComponentId::get_sparse_set_index(index)))
|
|
&& set
|
|
.without
|
|
.ones()
|
|
.all(|index| !set_contains_id(ComponentId::get_sparse_set_index(index)))
|
|
})
|
|
}
|
|
|
|
/// For the given `archetype`, adds any component accessed used by this query's underlying [`FilteredAccess`] to `access`.
|
|
///
|
|
/// The passed in `access` will be updated with any new accesses introduced by the new archetype.
|
|
///
|
|
/// # Safety
|
|
/// `archetype` must be from the `World` this state was initialized from.
|
|
pub unsafe fn update_archetype_component_access(
|
|
&mut self,
|
|
archetype: &Archetype,
|
|
access: &mut Access<ArchetypeComponentId>,
|
|
) {
|
|
// As a fast path, we can iterate directly over the components involved
|
|
// if the `access` isn't inverted.
|
|
let (component_reads_and_writes, component_reads_and_writes_inverted) =
|
|
self.component_access.access.component_reads_and_writes();
|
|
let (component_writes, component_writes_inverted) =
|
|
self.component_access.access.component_writes();
|
|
|
|
if !component_reads_and_writes_inverted && !component_writes_inverted {
|
|
component_reads_and_writes.for_each(|id| {
|
|
if let Some(id) = archetype.get_archetype_component_id(id) {
|
|
access.add_component_read(id);
|
|
}
|
|
});
|
|
component_writes.for_each(|id| {
|
|
if let Some(id) = archetype.get_archetype_component_id(id) {
|
|
access.add_component_write(id);
|
|
}
|
|
});
|
|
return;
|
|
}
|
|
|
|
for (component_id, archetype_component_id) in
|
|
archetype.components_with_archetype_component_id()
|
|
{
|
|
if self
|
|
.component_access
|
|
.access
|
|
.has_component_read(component_id)
|
|
{
|
|
access.add_component_read(archetype_component_id);
|
|
}
|
|
if self
|
|
.component_access
|
|
.access
|
|
.has_component_write(component_id)
|
|
{
|
|
access.add_component_write(archetype_component_id);
|
|
}
|
|
}
|
|
}
|
|
|
|
/// Use this to transform a [`QueryState`] into a more generic [`QueryState`].
|
|
/// This can be useful for passing to another function that might take the more general form.
|
|
/// See [`Query::transmute_lens`](crate::system::Query::transmute_lens) for more details.
|
|
///
|
|
/// You should not call [`update_archetypes`](Self::update_archetypes) on the returned [`QueryState`] as the result will be unpredictable.
|
|
/// You might end up with a mix of archetypes that only matched the original query + archetypes that only match
|
|
/// the new [`QueryState`]. Most of the safe methods on [`QueryState`] call [`QueryState::update_archetypes`] internally, so this
|
|
/// best used through a [`Query`]
|
|
pub fn transmute<'a, NewD: QueryData>(
|
|
&self,
|
|
world: impl Into<UnsafeWorldCell<'a>>,
|
|
) -> QueryState<NewD> {
|
|
self.transmute_filtered::<NewD, ()>(world.into())
|
|
}
|
|
|
|
/// Creates a new [`QueryState`] with the same underlying [`FilteredAccess`], matched tables and archetypes
|
|
/// as self but with a new type signature.
|
|
///
|
|
/// Panics if `NewD` or `NewF` require accesses that this query does not have.
|
|
pub fn transmute_filtered<'a, NewD: QueryData, NewF: QueryFilter>(
|
|
&self,
|
|
world: impl Into<UnsafeWorldCell<'a>>,
|
|
) -> QueryState<NewD, NewF> {
|
|
let world = world.into();
|
|
self.validate_world(world.id());
|
|
|
|
let mut component_access = FilteredAccess::default();
|
|
let mut fetch_state = NewD::get_state(world.components()).expect("Could not create fetch_state, Please initialize all referenced components before transmuting.");
|
|
let filter_state = NewF::get_state(world.components()).expect("Could not create filter_state, Please initialize all referenced components before transmuting.");
|
|
|
|
NewD::set_access(&mut fetch_state, &self.component_access);
|
|
NewD::update_component_access(&fetch_state, &mut component_access);
|
|
|
|
let mut filter_component_access = FilteredAccess::default();
|
|
NewF::update_component_access(&filter_state, &mut filter_component_access);
|
|
|
|
component_access.extend(&filter_component_access);
|
|
assert!(
|
|
component_access.is_subset(&self.component_access),
|
|
"Transmuted state for {} attempts to access terms that are not allowed by original state {}.",
|
|
core::any::type_name::<(NewD, NewF)>(), core::any::type_name::<(D, F)>()
|
|
);
|
|
|
|
QueryState {
|
|
world_id: self.world_id,
|
|
archetype_generation: self.archetype_generation,
|
|
matched_storage_ids: self.matched_storage_ids.clone(),
|
|
is_dense: self.is_dense,
|
|
fetch_state,
|
|
filter_state,
|
|
component_access: self.component_access.clone(),
|
|
matched_tables: self.matched_tables.clone(),
|
|
matched_archetypes: self.matched_archetypes.clone(),
|
|
#[cfg(feature = "trace")]
|
|
par_iter_span: tracing::info_span!(
|
|
"par_for_each",
|
|
query = core::any::type_name::<NewD>(),
|
|
filter = core::any::type_name::<NewF>(),
|
|
),
|
|
}
|
|
}
|
|
|
|
/// Use this to combine two queries. The data accessed will be the intersection
|
|
/// of archetypes included in both queries. This can be useful for accessing a
|
|
/// subset of the entities between two queries.
|
|
///
|
|
/// You should not call `update_archetypes` on the returned `QueryState` as the result
|
|
/// could be unpredictable. You might end up with a mix of archetypes that only matched
|
|
/// the original query + archetypes that only match the new `QueryState`. Most of the
|
|
/// safe methods on `QueryState` call [`QueryState::update_archetypes`] internally, so
|
|
/// this is best used through a `Query`.
|
|
///
|
|
/// ## Performance
|
|
///
|
|
/// This will have similar performance as constructing a new `QueryState` since much of internal state
|
|
/// needs to be reconstructed. But it will be a little faster as it only needs to compare the intersection
|
|
/// of matching archetypes rather than iterating over all archetypes.
|
|
///
|
|
/// ## Panics
|
|
///
|
|
/// Will panic if `NewD` contains accesses not in `Q` or `OtherQ`.
|
|
pub fn join<'a, OtherD: QueryData, NewD: QueryData>(
|
|
&self,
|
|
world: impl Into<UnsafeWorldCell<'a>>,
|
|
other: &QueryState<OtherD>,
|
|
) -> QueryState<NewD, ()> {
|
|
self.join_filtered::<_, (), NewD, ()>(world, other)
|
|
}
|
|
|
|
/// Use this to combine two queries. The data accessed will be the intersection
|
|
/// of archetypes included in both queries.
|
|
///
|
|
/// ## Panics
|
|
///
|
|
/// Will panic if `NewD` or `NewF` requires accesses not in `Q` or `OtherQ`.
|
|
pub fn join_filtered<
|
|
'a,
|
|
OtherD: QueryData,
|
|
OtherF: QueryFilter,
|
|
NewD: QueryData,
|
|
NewF: QueryFilter,
|
|
>(
|
|
&self,
|
|
world: impl Into<UnsafeWorldCell<'a>>,
|
|
other: &QueryState<OtherD, OtherF>,
|
|
) -> QueryState<NewD, NewF> {
|
|
if self.world_id != other.world_id {
|
|
panic!("Joining queries initialized on different worlds is not allowed.");
|
|
}
|
|
|
|
let world = world.into();
|
|
|
|
self.validate_world(world.id());
|
|
|
|
let mut component_access = FilteredAccess::default();
|
|
let mut new_fetch_state = NewD::get_state(world.components())
|
|
.expect("Could not create fetch_state, Please initialize all referenced components before transmuting.");
|
|
let new_filter_state = NewF::get_state(world.components())
|
|
.expect("Could not create filter_state, Please initialize all referenced components before transmuting.");
|
|
|
|
NewD::set_access(&mut new_fetch_state, &self.component_access);
|
|
NewD::update_component_access(&new_fetch_state, &mut component_access);
|
|
|
|
let mut new_filter_component_access = FilteredAccess::default();
|
|
NewF::update_component_access(&new_filter_state, &mut new_filter_component_access);
|
|
|
|
component_access.extend(&new_filter_component_access);
|
|
|
|
let mut joined_component_access = self.component_access.clone();
|
|
joined_component_access.extend(&other.component_access);
|
|
|
|
assert!(
|
|
component_access.is_subset(&joined_component_access),
|
|
"Joined state for {} attempts to access terms that are not allowed by state {} joined with {}.",
|
|
core::any::type_name::<(NewD, NewF)>(), core::any::type_name::<(D, F)>(), core::any::type_name::<(OtherD, OtherF)>()
|
|
);
|
|
|
|
if self.archetype_generation != other.archetype_generation {
|
|
warn!("You have tried to join queries with different archetype_generations. This could lead to unpredictable results.");
|
|
}
|
|
|
|
// the join is dense of both the queries were dense.
|
|
let is_dense = self.is_dense && other.is_dense;
|
|
|
|
// take the intersection of the matched ids
|
|
let mut matched_tables = self.matched_tables.clone();
|
|
let mut matched_archetypes = self.matched_archetypes.clone();
|
|
matched_tables.intersect_with(&other.matched_tables);
|
|
matched_archetypes.intersect_with(&other.matched_archetypes);
|
|
let matched_storage_ids = if is_dense {
|
|
matched_tables
|
|
.ones()
|
|
.map(|id| StorageId {
|
|
table_id: TableId::from_usize(id),
|
|
})
|
|
.collect()
|
|
} else {
|
|
matched_archetypes
|
|
.ones()
|
|
.map(|id| StorageId {
|
|
archetype_id: ArchetypeId::new(id),
|
|
})
|
|
.collect()
|
|
};
|
|
|
|
QueryState {
|
|
world_id: self.world_id,
|
|
archetype_generation: self.archetype_generation,
|
|
matched_storage_ids,
|
|
is_dense,
|
|
fetch_state: new_fetch_state,
|
|
filter_state: new_filter_state,
|
|
component_access: joined_component_access,
|
|
matched_tables,
|
|
matched_archetypes,
|
|
#[cfg(feature = "trace")]
|
|
par_iter_span: tracing::info_span!(
|
|
"par_for_each",
|
|
query = core::any::type_name::<NewD>(),
|
|
filter = core::any::type_name::<NewF>(),
|
|
),
|
|
}
|
|
}
|
|
|
|
/// Gets the query result for the given [`World`] and [`Entity`].
|
|
///
|
|
/// This can only be called for read-only queries, see [`Self::get_mut`] for write-queries.
|
|
///
|
|
/// This is always guaranteed to run in `O(1)` time.
|
|
#[inline]
|
|
pub fn get<'w>(
|
|
&mut self,
|
|
world: &'w World,
|
|
entity: Entity,
|
|
) -> Result<ROQueryItem<'w, D>, QueryEntityError<'w>> {
|
|
self.query(world).get_inner(entity)
|
|
}
|
|
|
|
/// Returns the read-only query results for the given array of [`Entity`].
|
|
///
|
|
/// In case of a nonexisting entity or mismatched component, a [`QueryEntityError`] is
|
|
/// returned instead.
|
|
///
|
|
/// Note that the unlike [`QueryState::get_many_mut`], the entities passed in do not need to be unique.
|
|
///
|
|
/// # Examples
|
|
///
|
|
/// ```
|
|
/// use bevy_ecs::prelude::*;
|
|
/// use bevy_ecs::query::QueryEntityError;
|
|
///
|
|
/// #[derive(Component, PartialEq, Debug)]
|
|
/// struct A(usize);
|
|
///
|
|
/// let mut world = World::new();
|
|
/// let entity_vec: Vec<Entity> = (0..3).map(|i|world.spawn(A(i)).id()).collect();
|
|
/// let entities: [Entity; 3] = entity_vec.try_into().unwrap();
|
|
///
|
|
/// world.spawn(A(73));
|
|
///
|
|
/// let mut query_state = world.query::<&A>();
|
|
///
|
|
/// let component_values = query_state.get_many(&world, entities).unwrap();
|
|
///
|
|
/// assert_eq!(component_values, [&A(0), &A(1), &A(2)]);
|
|
///
|
|
/// let wrong_entity = Entity::from_raw(365);
|
|
///
|
|
/// assert_eq!(match query_state.get_many(&mut world, [wrong_entity]).unwrap_err() {QueryEntityError::EntityDoesNotExist(error) => error.entity, _ => panic!()}, wrong_entity);
|
|
/// ```
|
|
#[inline]
|
|
pub fn get_many<'w, const N: usize>(
|
|
&mut self,
|
|
world: &'w World,
|
|
entities: [Entity; N],
|
|
) -> Result<[ROQueryItem<'w, D>; N], QueryEntityError<'w>> {
|
|
self.query(world).get_many_inner(entities)
|
|
}
|
|
|
|
/// Gets the query result for the given [`World`] and [`Entity`].
|
|
///
|
|
/// This is always guaranteed to run in `O(1)` time.
|
|
#[inline]
|
|
pub fn get_mut<'w>(
|
|
&mut self,
|
|
world: &'w mut World,
|
|
entity: Entity,
|
|
) -> Result<D::Item<'w>, QueryEntityError<'w>> {
|
|
self.query_mut(world).get_inner(entity)
|
|
}
|
|
|
|
/// Returns the query results for the given array of [`Entity`].
|
|
///
|
|
/// In case of a nonexisting entity or mismatched component, a [`QueryEntityError`] is
|
|
/// returned instead.
|
|
///
|
|
/// ```
|
|
/// use bevy_ecs::prelude::*;
|
|
/// use bevy_ecs::query::QueryEntityError;
|
|
///
|
|
/// #[derive(Component, PartialEq, Debug)]
|
|
/// struct A(usize);
|
|
///
|
|
/// let mut world = World::new();
|
|
///
|
|
/// let entities: Vec<Entity> = (0..3).map(|i|world.spawn(A(i)).id()).collect();
|
|
/// let entities: [Entity; 3] = entities.try_into().unwrap();
|
|
///
|
|
/// world.spawn(A(73));
|
|
///
|
|
/// let mut query_state = world.query::<&mut A>();
|
|
///
|
|
/// let mut mutable_component_values = query_state.get_many_mut(&mut world, entities).unwrap();
|
|
///
|
|
/// for mut a in &mut mutable_component_values {
|
|
/// a.0 += 5;
|
|
/// }
|
|
///
|
|
/// let component_values = query_state.get_many(&world, entities).unwrap();
|
|
///
|
|
/// assert_eq!(component_values, [&A(5), &A(6), &A(7)]);
|
|
///
|
|
/// let wrong_entity = Entity::from_raw(57);
|
|
/// let invalid_entity = world.spawn_empty().id();
|
|
///
|
|
/// assert_eq!(match query_state.get_many(&mut world, [wrong_entity]).unwrap_err() {QueryEntityError::EntityDoesNotExist(error) => error.entity, _ => panic!()}, wrong_entity);
|
|
/// assert_eq!(match query_state.get_many_mut(&mut world, [invalid_entity]).unwrap_err() {QueryEntityError::QueryDoesNotMatch(entity, _) => entity, _ => panic!()}, invalid_entity);
|
|
/// assert_eq!(query_state.get_many_mut(&mut world, [entities[0], entities[0]]).unwrap_err(), QueryEntityError::AliasedMutability(entities[0]));
|
|
/// ```
|
|
#[inline]
|
|
pub fn get_many_mut<'w, const N: usize>(
|
|
&mut self,
|
|
world: &'w mut World,
|
|
entities: [Entity; N],
|
|
) -> Result<[D::Item<'w>; N], QueryEntityError<'w>> {
|
|
self.query_mut(world).get_many_inner(entities)
|
|
}
|
|
|
|
/// Gets the query result for the given [`World`] and [`Entity`].
|
|
///
|
|
/// This method is slightly more efficient than [`QueryState::get`] in some situations, since
|
|
/// it does not update this instance's internal cache. This method will return an error if `entity`
|
|
/// belongs to an archetype that has not been cached.
|
|
///
|
|
/// To ensure that the cache is up to date, call [`QueryState::update_archetypes`] before this method.
|
|
/// The cache is also updated in [`QueryState::new`], `QueryState::get`, or any method with mutable
|
|
/// access to `self`.
|
|
///
|
|
/// This can only be called for read-only queries, see [`Self::get_mut`] for mutable queries.
|
|
///
|
|
/// This is always guaranteed to run in `O(1)` time.
|
|
#[inline]
|
|
pub fn get_manual<'w>(
|
|
&self,
|
|
world: &'w World,
|
|
entity: Entity,
|
|
) -> Result<ROQueryItem<'w, D>, QueryEntityError<'w>> {
|
|
self.query_manual(world).get_inner(entity)
|
|
}
|
|
|
|
/// Gets the query result for the given [`World`] and [`Entity`].
|
|
///
|
|
/// This is always guaranteed to run in `O(1)` time.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
#[inline]
|
|
pub unsafe fn get_unchecked<'w>(
|
|
&mut self,
|
|
world: UnsafeWorldCell<'w>,
|
|
entity: Entity,
|
|
) -> Result<D::Item<'w>, QueryEntityError<'w>> {
|
|
self.query_unchecked(world).get_inner(entity)
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over the query results for the given [`World`].
|
|
///
|
|
/// This can only be called for read-only queries, see [`Self::iter_mut`] for write-queries.
|
|
#[inline]
|
|
pub fn iter<'w, 's>(&'s mut self, world: &'w World) -> QueryIter<'w, 's, D::ReadOnly, F> {
|
|
self.query(world).into_iter()
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over the query results for the given [`World`].
|
|
///
|
|
/// This iterator is always guaranteed to return results from each matching entity once and only once.
|
|
/// Iteration order is not guaranteed.
|
|
#[inline]
|
|
pub fn iter_mut<'w, 's>(&'s mut self, world: &'w mut World) -> QueryIter<'w, 's, D, F> {
|
|
self.query_mut(world).into_iter()
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over the query results for the given [`World`] without updating the query's archetypes.
|
|
/// Archetypes must be manually updated before by using [`Self::update_archetypes`].
|
|
///
|
|
/// This iterator is always guaranteed to return results from each matching entity once and only once.
|
|
/// Iteration order is not guaranteed.
|
|
///
|
|
/// This can only be called for read-only queries.
|
|
#[inline]
|
|
pub fn iter_manual<'w, 's>(&'s self, world: &'w World) -> QueryIter<'w, 's, D::ReadOnly, F> {
|
|
self.query_manual(world).into_iter()
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over all possible combinations of `K` query results without repetition.
|
|
/// This can only be called for read-only queries.
|
|
///
|
|
/// A combination is an arrangement of a collection of items where order does not matter.
|
|
///
|
|
/// `K` is the number of items that make up each subset, and the number of items returned by the iterator.
|
|
/// `N` is the number of total entities output by query.
|
|
///
|
|
/// For example, given the list [1, 2, 3, 4], where `K` is 2, the combinations without repeats are
|
|
/// [1, 2], [1, 3], [1, 4], [2, 3], [2, 4], [3, 4].
|
|
/// And in this case, `N` would be defined as 4 since the size of the input list is 4.
|
|
///
|
|
/// For combinations of size `K` of query taking `N` inputs, you will get:
|
|
/// - if `K == N`: one combination of all query results
|
|
/// - if `K < N`: all possible `K`-sized combinations of query results, without repetition
|
|
/// - if `K > N`: empty set (no `K`-sized combinations exist)
|
|
///
|
|
/// The `iter_combinations` method does not guarantee order of iteration.
|
|
///
|
|
/// This iterator is always guaranteed to return results from each unique pair of matching entities.
|
|
/// Iteration order is not guaranteed.
|
|
///
|
|
/// This can only be called for read-only queries, see [`Self::iter_combinations_mut`] for
|
|
/// write-queries.
|
|
#[inline]
|
|
pub fn iter_combinations<'w, 's, const K: usize>(
|
|
&'s mut self,
|
|
world: &'w World,
|
|
) -> QueryCombinationIter<'w, 's, D::ReadOnly, F, K> {
|
|
self.query(world).iter_combinations_inner()
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over all possible combinations of `K` query results without repetition.
|
|
///
|
|
/// A combination is an arrangement of a collection of items where order does not matter.
|
|
///
|
|
/// `K` is the number of items that make up each subset, and the number of items returned by the iterator.
|
|
/// `N` is the number of total entities output by query.
|
|
///
|
|
/// For example, given the list [1, 2, 3, 4], where `K` is 2, the combinations without repeats are
|
|
/// [1, 2], [1, 3], [1, 4], [2, 3], [2, 4], [3, 4].
|
|
/// And in this case, `N` would be defined as 4 since the size of the input list is 4.
|
|
///
|
|
/// For combinations of size `K` of query taking `N` inputs, you will get:
|
|
/// - if `K == N`: one combination of all query results
|
|
/// - if `K < N`: all possible `K`-sized combinations of query results, without repetition
|
|
/// - if `K > N`: empty set (no `K`-sized combinations exist)
|
|
///
|
|
/// The `iter_combinations_mut` method does not guarantee order of iteration.
|
|
#[inline]
|
|
pub fn iter_combinations_mut<'w, 's, const K: usize>(
|
|
&'s mut self,
|
|
world: &'w mut World,
|
|
) -> QueryCombinationIter<'w, 's, D, F, K> {
|
|
self.query_mut(world).iter_combinations_inner()
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over the read-only query items generated from an [`Entity`] list.
|
|
///
|
|
/// Items are returned in the order of the list of entities.
|
|
/// Entities that don't match the query are skipped.
|
|
///
|
|
/// # See also
|
|
///
|
|
/// - [`iter_many_mut`](Self::iter_many_mut) to get mutable query items.
|
|
#[inline]
|
|
pub fn iter_many<'w, 's, EntityList: IntoIterator<Item: EntityBorrow>>(
|
|
&'s mut self,
|
|
world: &'w World,
|
|
entities: EntityList,
|
|
) -> QueryManyIter<'w, 's, D::ReadOnly, F, EntityList::IntoIter> {
|
|
self.query(world).iter_many_inner(entities)
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over the read-only query items generated from an [`Entity`] list.
|
|
///
|
|
/// Items are returned in the order of the list of entities.
|
|
/// Entities that don't match the query are skipped.
|
|
///
|
|
/// If `world` archetypes changed since [`Self::update_archetypes`] was last called,
|
|
/// this will skip entities contained in new archetypes.
|
|
///
|
|
/// This can only be called for read-only queries.
|
|
///
|
|
/// # See also
|
|
///
|
|
/// - [`iter_many`](Self::iter_many) to update archetypes.
|
|
/// - [`iter_manual`](Self::iter_manual) to iterate over all query items.
|
|
#[inline]
|
|
pub fn iter_many_manual<'w, 's, EntityList: IntoIterator<Item: EntityBorrow>>(
|
|
&'s self,
|
|
world: &'w World,
|
|
entities: EntityList,
|
|
) -> QueryManyIter<'w, 's, D::ReadOnly, F, EntityList::IntoIter> {
|
|
self.query_manual(world).iter_many_inner(entities)
|
|
}
|
|
|
|
/// Returns an iterator over the query items generated from an [`Entity`] list.
|
|
///
|
|
/// Items are returned in the order of the list of entities.
|
|
/// Entities that don't match the query are skipped.
|
|
#[inline]
|
|
pub fn iter_many_mut<'w, 's, EntityList: IntoIterator<Item: EntityBorrow>>(
|
|
&'s mut self,
|
|
world: &'w mut World,
|
|
entities: EntityList,
|
|
) -> QueryManyIter<'w, 's, D, F, EntityList::IntoIter> {
|
|
self.query_mut(world).iter_many_inner(entities)
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over the unique read-only query items generated from an [`EntitySet`].
|
|
///
|
|
/// Items are returned in the order of the list of entities.
|
|
/// Entities that don't match the query are skipped.
|
|
///
|
|
/// # See also
|
|
///
|
|
/// - [`iter_many_unique_mut`](Self::iter_many_unique_mut) to get mutable query items.
|
|
#[inline]
|
|
pub fn iter_many_unique<'w, 's, EntityList: EntitySet>(
|
|
&'s mut self,
|
|
world: &'w World,
|
|
entities: EntityList,
|
|
) -> QueryManyUniqueIter<'w, 's, D::ReadOnly, F, EntityList::IntoIter> {
|
|
self.query(world).iter_many_unique_inner(entities)
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over the unique read-only query items generated from an [`EntitySet`].
|
|
///
|
|
/// Items are returned in the order of the list of entities.
|
|
/// Entities that don't match the query are skipped.
|
|
///
|
|
/// If `world` archetypes changed since [`Self::update_archetypes`] was last called,
|
|
/// this will skip entities contained in new archetypes.
|
|
///
|
|
/// This can only be called for read-only queries.
|
|
///
|
|
/// # See also
|
|
///
|
|
/// - [`iter_many_unique`](Self::iter_many) to update archetypes.
|
|
/// - [`iter_many`](Self::iter_many) to iterate over a non-unique entity list.
|
|
/// - [`iter_manual`](Self::iter_manual) to iterate over all query items.
|
|
#[inline]
|
|
pub fn iter_many_unique_manual<'w, 's, EntityList: EntitySet>(
|
|
&'s self,
|
|
world: &'w World,
|
|
entities: EntityList,
|
|
) -> QueryManyUniqueIter<'w, 's, D::ReadOnly, F, EntityList::IntoIter> {
|
|
self.query_manual(world).iter_many_unique_inner(entities)
|
|
}
|
|
|
|
/// Returns an iterator over the unique query items generated from an [`EntitySet`].
|
|
///
|
|
/// Items are returned in the order of the list of entities.
|
|
/// Entities that don't match the query are skipped.
|
|
#[inline]
|
|
pub fn iter_many_unique_mut<'w, 's, EntityList: EntitySet>(
|
|
&'s mut self,
|
|
world: &'w mut World,
|
|
entities: EntityList,
|
|
) -> QueryManyUniqueIter<'w, 's, D, F, EntityList::IntoIter> {
|
|
self.query_mut(world).iter_many_unique_inner(entities)
|
|
}
|
|
/// Returns an [`Iterator`] over the query results for the given [`World`].
|
|
///
|
|
/// This iterator is always guaranteed to return results from each matching entity once and only once.
|
|
/// Iteration order is not guaranteed.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
#[inline]
|
|
pub unsafe fn iter_unchecked<'w, 's>(
|
|
&'s mut self,
|
|
world: UnsafeWorldCell<'w>,
|
|
) -> QueryIter<'w, 's, D, F> {
|
|
self.query_unchecked(world).into_iter()
|
|
}
|
|
|
|
/// Returns an [`Iterator`] over all possible combinations of `K` query results for the
|
|
/// given [`World`] without repetition.
|
|
/// This can only be called for read-only queries.
|
|
///
|
|
/// This iterator is always guaranteed to return results from each unique pair of matching entities.
|
|
/// Iteration order is not guaranteed.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
#[inline]
|
|
pub unsafe fn iter_combinations_unchecked<'w, 's, const K: usize>(
|
|
&'s mut self,
|
|
world: UnsafeWorldCell<'w>,
|
|
) -> QueryCombinationIter<'w, 's, D, F, K> {
|
|
self.query_unchecked(world).iter_combinations_inner()
|
|
}
|
|
|
|
/// Returns a parallel iterator over the query results for the given [`World`].
|
|
///
|
|
/// This can only be called for read-only queries, see [`par_iter_mut`] for write-queries.
|
|
///
|
|
/// Note that you must use the `for_each` method to iterate over the
|
|
/// results, see [`par_iter_mut`] for an example.
|
|
///
|
|
/// [`par_iter_mut`]: Self::par_iter_mut
|
|
#[inline]
|
|
pub fn par_iter<'w, 's>(
|
|
&'s mut self,
|
|
world: &'w World,
|
|
) -> QueryParIter<'w, 's, D::ReadOnly, F> {
|
|
self.query(world).par_iter_inner()
|
|
}
|
|
|
|
/// Returns a parallel iterator over the query results for the given [`World`].
|
|
///
|
|
/// This can only be called for mutable queries, see [`par_iter`] for read-only-queries.
|
|
///
|
|
/// # Examples
|
|
///
|
|
/// ```
|
|
/// use bevy_ecs::prelude::*;
|
|
/// use bevy_ecs::query::QueryEntityError;
|
|
///
|
|
/// #[derive(Component, PartialEq, Debug)]
|
|
/// struct A(usize);
|
|
///
|
|
/// # bevy_tasks::ComputeTaskPool::get_or_init(|| bevy_tasks::TaskPool::new());
|
|
///
|
|
/// let mut world = World::new();
|
|
///
|
|
/// # let entities: Vec<Entity> = (0..3).map(|i| world.spawn(A(i)).id()).collect();
|
|
/// # let entities: [Entity; 3] = entities.try_into().unwrap();
|
|
///
|
|
/// let mut query_state = world.query::<&mut A>();
|
|
///
|
|
/// query_state.par_iter_mut(&mut world).for_each(|mut a| {
|
|
/// a.0 += 5;
|
|
/// });
|
|
///
|
|
/// # let component_values = query_state.get_many(&world, entities).unwrap();
|
|
///
|
|
/// # assert_eq!(component_values, [&A(5), &A(6), &A(7)]);
|
|
///
|
|
/// # let wrong_entity = Entity::from_raw(57);
|
|
/// # let invalid_entity = world.spawn_empty().id();
|
|
///
|
|
/// # assert_eq!(match query_state.get_many(&mut world, [wrong_entity]).unwrap_err() {QueryEntityError::EntityDoesNotExist(error) => error.entity, _ => panic!()}, wrong_entity);
|
|
/// assert_eq!(match query_state.get_many_mut(&mut world, [invalid_entity]).unwrap_err() {QueryEntityError::QueryDoesNotMatch(entity, _) => entity, _ => panic!()}, invalid_entity);
|
|
/// # assert_eq!(query_state.get_many_mut(&mut world, [entities[0], entities[0]]).unwrap_err(), QueryEntityError::AliasedMutability(entities[0]));
|
|
/// ```
|
|
///
|
|
/// # Panics
|
|
/// The [`ComputeTaskPool`] is not initialized. If using this from a query that is being
|
|
/// initialized and run from the ECS scheduler, this should never panic.
|
|
///
|
|
/// [`par_iter`]: Self::par_iter
|
|
/// [`ComputeTaskPool`]: bevy_tasks::ComputeTaskPool
|
|
#[inline]
|
|
pub fn par_iter_mut<'w, 's>(&'s mut self, world: &'w mut World) -> QueryParIter<'w, 's, D, F> {
|
|
self.query_mut(world).par_iter_inner()
|
|
}
|
|
|
|
/// Runs `func` on each query result in parallel for the given [`World`], where the last change and
|
|
/// the current change tick are given. This is faster than the equivalent
|
|
/// `iter()` method, but cannot be chained like a normal [`Iterator`].
|
|
///
|
|
/// # Panics
|
|
/// The [`ComputeTaskPool`] is not initialized. If using this from a query that is being
|
|
/// initialized and run from the ECS scheduler, this should never panic.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
/// This does not validate that `world.id()` matches `self.world_id`. Calling this on a `world`
|
|
/// with a mismatched [`WorldId`] is unsound.
|
|
///
|
|
/// [`ComputeTaskPool`]: bevy_tasks::ComputeTaskPool
|
|
#[cfg(all(not(target_arch = "wasm32"), feature = "multi_threaded"))]
|
|
pub(crate) unsafe fn par_fold_init_unchecked_manual<'w, T, FN, INIT>(
|
|
&self,
|
|
init_accum: INIT,
|
|
world: UnsafeWorldCell<'w>,
|
|
batch_size: usize,
|
|
func: FN,
|
|
last_run: Tick,
|
|
this_run: Tick,
|
|
) where
|
|
FN: Fn(T, D::Item<'w>) -> T + Send + Sync + Clone,
|
|
INIT: Fn() -> T + Sync + Send + Clone,
|
|
{
|
|
// NOTE: If you are changing query iteration code, remember to update the following places, where relevant:
|
|
// QueryIter, QueryIterationCursor, QueryManyIter, QueryCombinationIter,QueryState::par_fold_init_unchecked_manual,
|
|
// QueryState::par_many_fold_init_unchecked_manual, QueryState::par_many_unique_fold_init_unchecked_manual
|
|
use arrayvec::ArrayVec;
|
|
|
|
bevy_tasks::ComputeTaskPool::get().scope(|scope| {
|
|
// SAFETY: We only access table data that has been registered in `self.archetype_component_access`.
|
|
let tables = unsafe { &world.storages().tables };
|
|
let archetypes = world.archetypes();
|
|
let mut batch_queue = ArrayVec::new();
|
|
let mut queue_entity_count = 0;
|
|
|
|
// submit a list of storages which smaller than batch_size as single task
|
|
let submit_batch_queue = |queue: &mut ArrayVec<StorageId, 128>| {
|
|
if queue.is_empty() {
|
|
return;
|
|
}
|
|
let queue = core::mem::take(queue);
|
|
let mut func = func.clone();
|
|
let init_accum = init_accum.clone();
|
|
scope.spawn(async move {
|
|
#[cfg(feature = "trace")]
|
|
let _span = self.par_iter_span.enter();
|
|
let mut iter = self
|
|
.query_unchecked_manual_with_ticks(world, last_run, this_run)
|
|
.into_iter();
|
|
let mut accum = init_accum();
|
|
for storage_id in queue {
|
|
accum = iter.fold_over_storage_range(accum, &mut func, storage_id, None);
|
|
}
|
|
});
|
|
};
|
|
|
|
// submit single storage larger than batch_size
|
|
let submit_single = |count, storage_id: StorageId| {
|
|
for offset in (0..count).step_by(batch_size) {
|
|
let mut func = func.clone();
|
|
let init_accum = init_accum.clone();
|
|
let len = batch_size.min(count - offset);
|
|
let batch = offset..offset + len;
|
|
scope.spawn(async move {
|
|
#[cfg(feature = "trace")]
|
|
let _span = self.par_iter_span.enter();
|
|
let accum = init_accum();
|
|
self.query_unchecked_manual_with_ticks(world, last_run, this_run)
|
|
.into_iter()
|
|
.fold_over_storage_range(accum, &mut func, storage_id, Some(batch));
|
|
});
|
|
}
|
|
};
|
|
|
|
let storage_entity_count = |storage_id: StorageId| -> usize {
|
|
if self.is_dense {
|
|
tables[storage_id.table_id].entity_count()
|
|
} else {
|
|
archetypes[storage_id.archetype_id].len()
|
|
}
|
|
};
|
|
|
|
for storage_id in &self.matched_storage_ids {
|
|
let count = storage_entity_count(*storage_id);
|
|
|
|
// skip empty storage
|
|
if count == 0 {
|
|
continue;
|
|
}
|
|
// immediately submit large storage
|
|
if count >= batch_size {
|
|
submit_single(count, *storage_id);
|
|
continue;
|
|
}
|
|
// merge small storage
|
|
batch_queue.push(*storage_id);
|
|
queue_entity_count += count;
|
|
|
|
// submit batch_queue
|
|
if queue_entity_count >= batch_size || batch_queue.is_full() {
|
|
submit_batch_queue(&mut batch_queue);
|
|
queue_entity_count = 0;
|
|
}
|
|
}
|
|
submit_batch_queue(&mut batch_queue);
|
|
});
|
|
}
|
|
|
|
/// Runs `func` on each query result in parallel for the given [`EntitySet`],
|
|
/// where the last change and the current change tick are given. This is faster than the
|
|
/// equivalent `iter_many_unique()` method, but cannot be chained like a normal [`Iterator`].
|
|
///
|
|
/// # Panics
|
|
/// The [`ComputeTaskPool`] is not initialized. If using this from a query that is being
|
|
/// initialized and run from the ECS scheduler, this should never panic.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
/// This does not validate that `world.id()` matches `self.world_id`. Calling this on a `world`
|
|
/// with a mismatched [`WorldId`] is unsound.
|
|
///
|
|
/// [`ComputeTaskPool`]: bevy_tasks::ComputeTaskPool
|
|
#[cfg(all(not(target_arch = "wasm32"), feature = "multi_threaded"))]
|
|
pub(crate) unsafe fn par_many_unique_fold_init_unchecked_manual<'w, T, FN, INIT, E>(
|
|
&self,
|
|
init_accum: INIT,
|
|
world: UnsafeWorldCell<'w>,
|
|
entity_list: &UniqueEntitySlice<E>,
|
|
batch_size: usize,
|
|
mut func: FN,
|
|
last_run: Tick,
|
|
this_run: Tick,
|
|
) where
|
|
FN: Fn(T, D::Item<'w>) -> T + Send + Sync + Clone,
|
|
INIT: Fn() -> T + Sync + Send + Clone,
|
|
E: TrustedEntityBorrow + Sync,
|
|
{
|
|
// NOTE: If you are changing query iteration code, remember to update the following places, where relevant:
|
|
// QueryIter, QueryIterationCursor, QueryManyIter, QueryCombinationIter,QueryState::par_fold_init_unchecked_manual
|
|
// QueryState::par_many_fold_init_unchecked_manual, QueryState::par_many_unique_fold_init_unchecked_manual
|
|
|
|
bevy_tasks::ComputeTaskPool::get().scope(|scope| {
|
|
let chunks = entity_list.chunks_exact(batch_size);
|
|
let remainder = chunks.remainder();
|
|
|
|
for batch in chunks {
|
|
let mut func = func.clone();
|
|
let init_accum = init_accum.clone();
|
|
scope.spawn(async move {
|
|
#[cfg(feature = "trace")]
|
|
let _span = self.par_iter_span.enter();
|
|
let accum = init_accum();
|
|
self.query_unchecked_manual_with_ticks(world, last_run, this_run)
|
|
.iter_many_unique_inner(batch)
|
|
.fold(accum, &mut func);
|
|
});
|
|
}
|
|
|
|
#[cfg(feature = "trace")]
|
|
let _span = self.par_iter_span.enter();
|
|
let accum = init_accum();
|
|
self.query_unchecked_manual_with_ticks(world, last_run, this_run)
|
|
.iter_many_unique_inner(remainder)
|
|
.fold(accum, &mut func);
|
|
});
|
|
}
|
|
}
|
|
|
|
impl<D: ReadOnlyQueryData, F: QueryFilter> QueryState<D, F> {
|
|
/// Runs `func` on each read-only query result in parallel for the given [`Entity`] list,
|
|
/// where the last change and the current change tick are given. This is faster than the equivalent
|
|
/// `iter_many()` method, but cannot be chained like a normal [`Iterator`].
|
|
///
|
|
/// # Panics
|
|
/// The [`ComputeTaskPool`] is not initialized. If using this from a query that is being
|
|
/// initialized and run from the ECS scheduler, this should never panic.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
/// This does not validate that `world.id()` matches `self.world_id`. Calling this on a `world`
|
|
/// with a mismatched [`WorldId`] is unsound.
|
|
///
|
|
/// [`ComputeTaskPool`]: bevy_tasks::ComputeTaskPool
|
|
#[cfg(all(not(target_arch = "wasm32"), feature = "multi_threaded"))]
|
|
pub(crate) unsafe fn par_many_fold_init_unchecked_manual<'w, T, FN, INIT, E>(
|
|
&self,
|
|
init_accum: INIT,
|
|
world: UnsafeWorldCell<'w>,
|
|
entity_list: &[E],
|
|
batch_size: usize,
|
|
mut func: FN,
|
|
last_run: Tick,
|
|
this_run: Tick,
|
|
) where
|
|
FN: Fn(T, D::Item<'w>) -> T + Send + Sync + Clone,
|
|
INIT: Fn() -> T + Sync + Send + Clone,
|
|
E: EntityBorrow + Sync,
|
|
{
|
|
// NOTE: If you are changing query iteration code, remember to update the following places, where relevant:
|
|
// QueryIter, QueryIterationCursor, QueryManyIter, QueryCombinationIter, QueryState::par_fold_init_unchecked_manual
|
|
// QueryState::par_many_fold_init_unchecked_manual, QueryState::par_many_unique_fold_init_unchecked_manual
|
|
|
|
bevy_tasks::ComputeTaskPool::get().scope(|scope| {
|
|
let chunks = entity_list.chunks_exact(batch_size);
|
|
let remainder = chunks.remainder();
|
|
|
|
for batch in chunks {
|
|
let mut func = func.clone();
|
|
let init_accum = init_accum.clone();
|
|
scope.spawn(async move {
|
|
#[cfg(feature = "trace")]
|
|
let _span = self.par_iter_span.enter();
|
|
let accum = init_accum();
|
|
self.query_unchecked_manual_with_ticks(world, last_run, this_run)
|
|
.iter_many_inner(batch)
|
|
.fold(accum, &mut func);
|
|
});
|
|
}
|
|
|
|
#[cfg(feature = "trace")]
|
|
let _span = self.par_iter_span.enter();
|
|
let accum = init_accum();
|
|
self.query_unchecked_manual_with_ticks(world, last_run, this_run)
|
|
.iter_many_inner(remainder)
|
|
.fold(accum, &mut func);
|
|
});
|
|
}
|
|
}
|
|
|
|
impl<D: QueryData, F: QueryFilter> QueryState<D, F> {
|
|
/// Returns a single immutable query result when there is exactly one entity matching
|
|
/// the query.
|
|
///
|
|
/// This can only be called for read-only queries,
|
|
/// see [`single_mut`](Self::single_mut) for write-queries.
|
|
///
|
|
/// # Panics
|
|
///
|
|
/// Panics if the number of query results is not exactly one. Use
|
|
/// [`get_single`](Self::get_single) to return a `Result` instead of panicking.
|
|
#[track_caller]
|
|
#[inline]
|
|
pub fn single<'w>(&mut self, world: &'w World) -> ROQueryItem<'w, D> {
|
|
self.query(world).single_inner()
|
|
}
|
|
|
|
/// Returns a single immutable query result when there is exactly one entity matching
|
|
/// the query.
|
|
///
|
|
/// This can only be called for read-only queries,
|
|
/// see [`get_single_mut`](Self::get_single_mut) for write-queries.
|
|
///
|
|
/// If the number of query results is not exactly one, a [`QuerySingleError`] is returned
|
|
/// instead.
|
|
#[inline]
|
|
pub fn get_single<'w>(
|
|
&mut self,
|
|
world: &'w World,
|
|
) -> Result<ROQueryItem<'w, D>, QuerySingleError> {
|
|
self.query(world).get_single_inner()
|
|
}
|
|
|
|
/// Returns a single mutable query result when there is exactly one entity matching
|
|
/// the query.
|
|
///
|
|
/// # Panics
|
|
///
|
|
/// Panics if the number of query results is not exactly one. Use
|
|
/// [`get_single_mut`](Self::get_single_mut) to return a `Result` instead of panicking.
|
|
#[track_caller]
|
|
#[inline]
|
|
pub fn single_mut<'w>(&mut self, world: &'w mut World) -> D::Item<'w> {
|
|
self.query_mut(world).single_inner()
|
|
}
|
|
|
|
/// Returns a single mutable query result when there is exactly one entity matching
|
|
/// the query.
|
|
///
|
|
/// If the number of query results is not exactly one, a [`QuerySingleError`] is returned
|
|
/// instead.
|
|
#[inline]
|
|
pub fn get_single_mut<'w>(
|
|
&mut self,
|
|
world: &'w mut World,
|
|
) -> Result<D::Item<'w>, QuerySingleError> {
|
|
self.query_mut(world).get_single_inner()
|
|
}
|
|
|
|
/// Returns a query result when there is exactly one entity matching the query.
|
|
///
|
|
/// If the number of query results is not exactly one, a [`QuerySingleError`] is returned
|
|
/// instead.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
#[inline]
|
|
pub unsafe fn get_single_unchecked<'w>(
|
|
&mut self,
|
|
world: UnsafeWorldCell<'w>,
|
|
) -> Result<D::Item<'w>, QuerySingleError> {
|
|
self.query_unchecked(world).get_single_inner()
|
|
}
|
|
|
|
/// Returns a query result when there is exactly one entity matching the query,
|
|
/// where the last change and the current change tick are given.
|
|
///
|
|
/// If the number of query results is not exactly one, a [`QuerySingleError`] is returned
|
|
/// instead.
|
|
///
|
|
/// # Safety
|
|
///
|
|
/// This does not check for mutable query correctness. To be safe, make sure mutable queries
|
|
/// have unique access to the components they query.
|
|
/// This does not validate that `world.id()` matches `self.world_id`. Calling this on a `world`
|
|
/// with a mismatched [`WorldId`] is unsound.
|
|
#[inline]
|
|
pub unsafe fn get_single_unchecked_manual<'w>(
|
|
&self,
|
|
world: UnsafeWorldCell<'w>,
|
|
last_run: Tick,
|
|
this_run: Tick,
|
|
) -> Result<D::Item<'w>, QuerySingleError> {
|
|
// SAFETY:
|
|
// - The caller ensured we have the correct access to the world.
|
|
// - The caller ensured that the world matches.
|
|
self.query_unchecked_manual_with_ticks(world, last_run, this_run)
|
|
.get_single_inner()
|
|
}
|
|
}
|
|
|
|
impl<D: QueryData, F: QueryFilter> From<QueryBuilder<'_, D, F>> for QueryState<D, F> {
|
|
fn from(mut value: QueryBuilder<D, F>) -> Self {
|
|
QueryState::from_builder(&mut value)
|
|
}
|
|
}
|
|
|
|
#[cfg(test)]
|
|
mod tests {
|
|
use crate::{
|
|
component::Component, entity_disabling::DefaultQueryFilters, prelude::*,
|
|
world::FilteredEntityRef,
|
|
};
|
|
|
|
#[test]
|
|
#[should_panic]
|
|
fn right_world_get() {
|
|
let mut world_1 = World::new();
|
|
let world_2 = World::new();
|
|
|
|
let mut query_state = world_1.query::<Entity>();
|
|
let _panics = query_state.get(&world_2, Entity::from_raw(0));
|
|
}
|
|
|
|
#[test]
|
|
#[should_panic]
|
|
fn right_world_get_many() {
|
|
let mut world_1 = World::new();
|
|
let world_2 = World::new();
|
|
|
|
let mut query_state = world_1.query::<Entity>();
|
|
let _panics = query_state.get_many(&world_2, []);
|
|
}
|
|
|
|
#[test]
|
|
#[should_panic]
|
|
fn right_world_get_many_mut() {
|
|
let mut world_1 = World::new();
|
|
let mut world_2 = World::new();
|
|
|
|
let mut query_state = world_1.query::<Entity>();
|
|
let _panics = query_state.get_many_mut(&mut world_2, []);
|
|
}
|
|
|
|
#[derive(Component, PartialEq, Debug)]
|
|
struct A(usize);
|
|
|
|
#[derive(Component, PartialEq, Debug)]
|
|
struct B(usize);
|
|
|
|
#[derive(Component, PartialEq, Debug)]
|
|
struct C(usize);
|
|
|
|
#[test]
|
|
fn can_transmute_to_more_general() {
|
|
let mut world = World::new();
|
|
world.spawn((A(1), B(0)));
|
|
|
|
let query_state = world.query::<(&A, &B)>();
|
|
let mut new_query_state = query_state.transmute::<&A>(&world);
|
|
assert_eq!(new_query_state.iter(&world).len(), 1);
|
|
let a = new_query_state.single(&world);
|
|
|
|
assert_eq!(a.0, 1);
|
|
}
|
|
|
|
#[test]
|
|
fn cannot_get_data_not_in_original_query() {
|
|
let mut world = World::new();
|
|
world.spawn((A(0), B(0)));
|
|
world.spawn((A(1), B(0), C(0)));
|
|
|
|
let query_state = world.query_filtered::<(&A, &B), Without<C>>();
|
|
let mut new_query_state = query_state.transmute::<&A>(&world);
|
|
// even though we change the query to not have Without<C>, we do not get the component with C.
|
|
let a = new_query_state.single(&world);
|
|
|
|
assert_eq!(a.0, 0);
|
|
}
|
|
|
|
#[test]
|
|
fn can_transmute_empty_tuple() {
|
|
let mut world = World::new();
|
|
world.register_component::<A>();
|
|
let entity = world.spawn(A(10)).id();
|
|
|
|
let q = world.query::<()>();
|
|
let mut q = q.transmute::<Entity>(&world);
|
|
assert_eq!(q.single(&world), entity);
|
|
}
|
|
|
|
#[test]
|
|
fn can_transmute_immut_fetch() {
|
|
let mut world = World::new();
|
|
world.spawn(A(10));
|
|
|
|
let q = world.query::<&A>();
|
|
let mut new_q = q.transmute::<Ref<A>>(&world);
|
|
assert!(new_q.single(&world).is_added());
|
|
|
|
let q = world.query::<Ref<A>>();
|
|
let _ = q.transmute::<&A>(&world);
|
|
}
|
|
|
|
#[test]
|
|
fn can_transmute_mut_fetch() {
|
|
let mut world = World::new();
|
|
world.spawn(A(0));
|
|
|
|
let q = world.query::<&mut A>();
|
|
let _ = q.transmute::<Ref<A>>(&world);
|
|
let _ = q.transmute::<&A>(&world);
|
|
}
|
|
|
|
#[test]
|
|
fn can_transmute_entity_mut() {
|
|
let mut world = World::new();
|
|
world.spawn(A(0));
|
|
|
|
let q: QueryState<EntityMut<'_>> = world.query::<EntityMut>();
|
|
let _ = q.transmute::<EntityRef>(&world);
|
|
}
|
|
|
|
#[test]
|
|
fn can_generalize_with_option() {
|
|
let mut world = World::new();
|
|
world.spawn((A(0), B(0)));
|
|
|
|
let query_state = world.query::<(Option<&A>, &B)>();
|
|
let _ = query_state.transmute::<Option<&A>>(&world);
|
|
let _ = query_state.transmute::<&B>(&world);
|
|
}
|
|
|
|
#[test]
|
|
#[should_panic(
|
|
expected = "Transmuted state for ((&bevy_ecs::query::state::tests::A, &bevy_ecs::query::state::tests::B), ()) attempts to access terms that are not allowed by original state (&bevy_ecs::query::state::tests::A, ())."
|
|
)]
|
|
fn cannot_transmute_to_include_data_not_in_original_query() {
|
|
let mut world = World::new();
|
|
world.register_component::<A>();
|
|
world.register_component::<B>();
|
|
world.spawn(A(0));
|
|
|
|
let query_state = world.query::<&A>();
|
|
let mut _new_query_state = query_state.transmute::<(&A, &B)>(&world);
|
|
}
|
|
|
|
#[test]
|
|
#[should_panic(
|
|
expected = "Transmuted state for (&mut bevy_ecs::query::state::tests::A, ()) attempts to access terms that are not allowed by original state (&bevy_ecs::query::state::tests::A, ())."
|
|
)]
|
|
fn cannot_transmute_immut_to_mut() {
|
|
let mut world = World::new();
|
|
world.spawn(A(0));
|
|
|
|
let query_state = world.query::<&A>();
|
|
let mut _new_query_state = query_state.transmute::<&mut A>(&world);
|
|
}
|
|
|
|
#[test]
|
|
#[should_panic(
|
|
expected = "Transmuted state for (&bevy_ecs::query::state::tests::A, ()) attempts to access terms that are not allowed by original state (core::option::Option<&bevy_ecs::query::state::tests::A>, ())."
|
|
)]
|
|
fn cannot_transmute_option_to_immut() {
|
|
let mut world = World::new();
|
|
world.spawn(C(0));
|
|
|
|
let query_state = world.query::<Option<&A>>();
|
|
let mut new_query_state = query_state.transmute::<&A>(&world);
|
|
let x = new_query_state.single(&world);
|
|
assert_eq!(x.0, 1234);
|
|
}
|
|
|
|
#[test]
|
|
#[should_panic(
|
|
expected = "Transmuted state for (&bevy_ecs::query::state::tests::A, ()) attempts to access terms that are not allowed by original state (bevy_ecs::world::entity_ref::EntityRef, ())."
|
|
)]
|
|
fn cannot_transmute_entity_ref() {
|
|
let mut world = World::new();
|
|
world.register_component::<A>();
|
|
|
|
let q = world.query::<EntityRef>();
|
|
let _ = q.transmute::<&A>(&world);
|
|
}
|
|
|
|
#[test]
|
|
fn can_transmute_filtered_entity() {
|
|
let mut world = World::new();
|
|
let entity = world.spawn((A(0), B(1))).id();
|
|
let query =
|
|
QueryState::<(Entity, &A, &B)>::new(&mut world).transmute::<FilteredEntityRef>(&world);
|
|
|
|
let mut query = query;
|
|
// Our result is completely untyped
|
|
let entity_ref = query.single(&world);
|
|
|
|
assert_eq!(entity, entity_ref.id());
|
|
assert_eq!(0, entity_ref.get::<A>().unwrap().0);
|
|
assert_eq!(1, entity_ref.get::<B>().unwrap().0);
|
|
}
|
|
|
|
#[test]
|
|
fn can_transmute_added() {
|
|
let mut world = World::new();
|
|
let entity_a = world.spawn(A(0)).id();
|
|
|
|
let mut query = QueryState::<(Entity, &A, Has<B>)>::new(&mut world)
|
|
.transmute_filtered::<(Entity, Has<B>), Added<A>>(&world);
|
|
|
|
assert_eq!((entity_a, false), query.single(&world));
|
|
|
|
world.clear_trackers();
|
|
|
|
let entity_b = world.spawn((A(0), B(0))).id();
|
|
assert_eq!((entity_b, true), query.single(&world));
|
|
|
|
world.clear_trackers();
|
|
|
|
assert!(query.get_single(&world).is_err());
|
|
}
|
|
|
|
#[test]
|
|
fn can_transmute_changed() {
|
|
let mut world = World::new();
|
|
let entity_a = world.spawn(A(0)).id();
|
|
|
|
let mut detection_query = QueryState::<(Entity, &A)>::new(&mut world)
|
|
.transmute_filtered::<Entity, Changed<A>>(&world);
|
|
|
|
let mut change_query = QueryState::<&mut A>::new(&mut world);
|
|
assert_eq!(entity_a, detection_query.single(&world));
|
|
|
|
world.clear_trackers();
|
|
|
|
assert!(detection_query.get_single(&world).is_err());
|
|
|
|
change_query.single_mut(&mut world).0 = 1;
|
|
|
|
assert_eq!(entity_a, detection_query.single(&world));
|
|
}
|
|
|
|
#[test]
|
|
#[should_panic(
|
|
expected = "Transmuted state for (bevy_ecs::entity::Entity, bevy_ecs::query::filter::Changed<bevy_ecs::query::state::tests::B>) attempts to access terms that are not allowed by original state (&bevy_ecs::query::state::tests::A, ())."
|
|
)]
|
|
fn cannot_transmute_changed_without_access() {
|
|
let mut world = World::new();
|
|
world.register_component::<A>();
|
|
world.register_component::<B>();
|
|
let query = QueryState::<&A>::new(&mut world);
|
|
let _new_query = query.transmute_filtered::<Entity, Changed<B>>(&world);
|
|
}
|
|
|
|
// Regression test for #14629
|
|
#[test]
|
|
#[should_panic]
|
|
fn transmute_with_different_world() {
|
|
let mut world = World::new();
|
|
world.spawn((A(1), B(2)));
|
|
|
|
let mut world2 = World::new();
|
|
world2.register_component::<B>();
|
|
|
|
world.query::<(&A, &B)>().transmute::<&B>(&world2);
|
|
}
|
|
|
|
/// Regression test for issue #14528
|
|
#[test]
|
|
fn transmute_from_sparse_to_dense() {
|
|
#[derive(Component)]
|
|
struct Dense;
|
|
|
|
#[derive(Component)]
|
|
#[component(storage = "SparseSet")]
|
|
struct Sparse;
|
|
|
|
let mut world = World::new();
|
|
|
|
world.spawn(Dense);
|
|
world.spawn((Dense, Sparse));
|
|
|
|
let mut query = world
|
|
.query_filtered::<&Dense, With<Sparse>>()
|
|
.transmute::<&Dense>(&world);
|
|
|
|
let matched = query.iter(&world).count();
|
|
assert_eq!(matched, 1);
|
|
}
|
|
#[test]
|
|
fn transmute_from_dense_to_sparse() {
|
|
#[derive(Component)]
|
|
struct Dense;
|
|
|
|
#[derive(Component)]
|
|
#[component(storage = "SparseSet")]
|
|
struct Sparse;
|
|
|
|
let mut world = World::new();
|
|
|
|
world.spawn(Dense);
|
|
world.spawn((Dense, Sparse));
|
|
|
|
let mut query = world
|
|
.query::<&Dense>()
|
|
.transmute_filtered::<&Dense, With<Sparse>>(&world);
|
|
|
|
// Note: `transmute_filtered` is supposed to keep the same matched tables/archetypes,
|
|
// so it doesn't actually filter out those entities without `Sparse` and the iteration
|
|
// remains dense.
|
|
let matched = query.iter(&world).count();
|
|
assert_eq!(matched, 2);
|
|
}
|
|
|
|
#[test]
|
|
fn join() {
|
|
let mut world = World::new();
|
|
world.spawn(A(0));
|
|
world.spawn(B(1));
|
|
let entity_ab = world.spawn((A(2), B(3))).id();
|
|
world.spawn((A(4), B(5), C(6)));
|
|
|
|
let query_1 = QueryState::<&A, Without<C>>::new(&mut world);
|
|
let query_2 = QueryState::<&B, Without<C>>::new(&mut world);
|
|
let mut new_query: QueryState<Entity, ()> = query_1.join_filtered(&world, &query_2);
|
|
|
|
assert_eq!(new_query.single(&world), entity_ab);
|
|
}
|
|
|
|
#[test]
|
|
fn join_with_get() {
|
|
let mut world = World::new();
|
|
world.spawn(A(0));
|
|
world.spawn(B(1));
|
|
let entity_ab = world.spawn((A(2), B(3))).id();
|
|
let entity_abc = world.spawn((A(4), B(5), C(6))).id();
|
|
|
|
let query_1 = QueryState::<&A>::new(&mut world);
|
|
let query_2 = QueryState::<&B, Without<C>>::new(&mut world);
|
|
let mut new_query: QueryState<Entity, ()> = query_1.join_filtered(&world, &query_2);
|
|
|
|
assert!(new_query.get(&world, entity_ab).is_ok());
|
|
// should not be able to get entity with c.
|
|
assert!(new_query.get(&world, entity_abc).is_err());
|
|
}
|
|
|
|
#[test]
|
|
#[should_panic(expected = "Joined state for (&bevy_ecs::query::state::tests::C, ()) \
|
|
attempts to access terms that are not allowed by state \
|
|
(&bevy_ecs::query::state::tests::A, ()) joined with (&bevy_ecs::query::state::tests::B, ()).")]
|
|
fn cannot_join_wrong_fetch() {
|
|
let mut world = World::new();
|
|
world.register_component::<C>();
|
|
let query_1 = QueryState::<&A>::new(&mut world);
|
|
let query_2 = QueryState::<&B>::new(&mut world);
|
|
let _query: QueryState<&C> = query_1.join(&world, &query_2);
|
|
}
|
|
|
|
#[test]
|
|
#[should_panic(
|
|
expected = "Joined state for (bevy_ecs::entity::Entity, bevy_ecs::query::filter::Changed<bevy_ecs::query::state::tests::C>) \
|
|
attempts to access terms that are not allowed by state \
|
|
(&bevy_ecs::query::state::tests::A, bevy_ecs::query::filter::Without<bevy_ecs::query::state::tests::C>) \
|
|
joined with (&bevy_ecs::query::state::tests::B, bevy_ecs::query::filter::Without<bevy_ecs::query::state::tests::C>)."
|
|
)]
|
|
fn cannot_join_wrong_filter() {
|
|
let mut world = World::new();
|
|
let query_1 = QueryState::<&A, Without<C>>::new(&mut world);
|
|
let query_2 = QueryState::<&B, Without<C>>::new(&mut world);
|
|
let _: QueryState<Entity, Changed<C>> = query_1.join_filtered(&world, &query_2);
|
|
}
|
|
|
|
#[test]
|
|
fn query_respects_default_filters() {
|
|
let mut world = World::new();
|
|
world.spawn((A(0), B(0)));
|
|
world.spawn((B(0), C(0)));
|
|
world.spawn(C(0));
|
|
|
|
let mut df = DefaultQueryFilters::empty();
|
|
df.register_disabling_component(world.register_component::<C>());
|
|
world.insert_resource(df);
|
|
|
|
// Without<C> only matches the first entity
|
|
let mut query = QueryState::<()>::new(&mut world);
|
|
assert_eq!(1, query.iter(&world).count());
|
|
|
|
// With<C> matches the last two entities
|
|
let mut query = QueryState::<(), With<C>>::new(&mut world);
|
|
assert_eq!(2, query.iter(&world).count());
|
|
|
|
// Has should bypass the filter entirely
|
|
let mut query = QueryState::<Has<C>>::new(&mut world);
|
|
assert_eq!(3, query.iter(&world).count());
|
|
|
|
// Other filters should still be respected
|
|
let mut query = QueryState::<Has<C>, Without<B>>::new(&mut world);
|
|
assert_eq!(1, query.iter(&world).count());
|
|
}
|
|
|
|
#[derive(Component)]
|
|
struct Table;
|
|
|
|
#[derive(Component)]
|
|
#[component(storage = "SparseSet")]
|
|
struct Sparse;
|
|
|
|
#[test]
|
|
fn query_default_filters_updates_is_dense() {
|
|
let mut world = World::new();
|
|
world.spawn((Table, Sparse));
|
|
world.spawn(Table);
|
|
world.spawn(Sparse);
|
|
|
|
let mut query = QueryState::<()>::new(&mut world);
|
|
// There are no sparse components involved thus the query is dense
|
|
assert!(query.is_dense);
|
|
assert_eq!(3, query.iter(&world).count());
|
|
|
|
let mut df = DefaultQueryFilters::empty();
|
|
df.register_disabling_component(world.register_component::<Sparse>());
|
|
world.insert_resource(df);
|
|
|
|
let mut query = QueryState::<()>::new(&mut world);
|
|
// The query doesn't ask for sparse components, but the default filters adds
|
|
// a sparse components thus it is NOT dense
|
|
assert!(!query.is_dense);
|
|
assert_eq!(1, query.iter(&world).count());
|
|
|
|
let mut df = DefaultQueryFilters::empty();
|
|
df.register_disabling_component(world.register_component::<Table>());
|
|
world.insert_resource(df);
|
|
|
|
let mut query = QueryState::<()>::new(&mut world);
|
|
// If the filter is instead a table components, the query can still be dense
|
|
assert!(query.is_dense);
|
|
assert_eq!(1, query.iter(&world).count());
|
|
|
|
let mut query = QueryState::<&Sparse>::new(&mut world);
|
|
// But only if the original query was dense
|
|
assert!(!query.is_dense);
|
|
assert_eq!(1, query.iter(&world).count());
|
|
}
|
|
}
|