bevy/crates/bevy_transform/src
Carter Anderson d8fa57bd7b
Switch ChildOf back to tuple struct (#18672)
# Objective

In #17905 we swapped to a named field on `ChildOf` to help resolve
variable naming ambiguity of child vs parent (ex: `child_of.parent`
clearly reads as "I am accessing the parent of the child_of
relationship", whereas `child_of.0` is less clear).

Unfortunately this has the side effect of making initialization less
ideal. `ChildOf { parent }` reads just as well as `ChildOf(parent)`, but
`ChildOf { parent: root }` doesn't read nearly as well as
`ChildOf(root)`.

## Solution

Move back to `ChildOf(pub Entity)` but add a `child_of.parent()`
function and use it for all accesses. The downside here is that users
are no longer "forced" to access the parent field with `parent`
nomenclature, but I think this strikes the right balance.

Take a look at the diff. I think the results provide strong evidence for
this change. Initialization has the benefit of reading much better _and_
of taking up significantly less space, as many lines go from 3 to 1, and
we're cutting out a bunch of syntax in some cases.

Sadly I do think this should land in 0.16 as the cost of doing this
_after_ the relationships migration is high.
2025-04-02 00:10:10 +00:00
..
components Transform Propagation Optimization: Static Subtree Marking (#18589) 2025-03-30 02:43:39 +00:00
commands.rs Parent -> ChildOf (#17427) 2025-01-20 22:13:29 +00:00
helper.rs Switch ChildOf back to tuple struct (#18672) 2025-04-02 00:10:10 +00:00
lib.rs Move #![warn(clippy::allow_attributes, clippy::allow_attributes_without_reason)] to the workspace Cargo.toml (#17374) 2025-01-15 01:14:58 +00:00
plugins.rs Transform Propagation Optimization: Static Subtree Marking (#18589) 2025-03-30 02:43:39 +00:00
systems.rs Switch ChildOf back to tuple struct (#18672) 2025-04-02 00:10:10 +00:00
traits.rs Conversions for Isometry3d ⟷ Transform/GlobalTransform (#14478) 2024-07-25 20:23:32 +00:00