Consider the following:
struct Complex<T> {
something: T,
#[compute(field_bound)]
something_else: Generic<Self, T>,
}
That will generate:
impl<T> ToComputedValue for Complex<T>
where
T: ToComputedValue,
Generic<Self, T>: ToComputedValue<ComputedValue = Generic<Self, <T as ToComputedValue>::ComputedValue>>,
{
// ...
}
That last clause is obviously incorrect. map_type_params correctly maps
the T, but it should know also about Self.
Ideally we could just do the same as for T and do:
<Self as ToComputedValue>::ComputedValue
But that doesn't quite work, because we are in that implementation of
the trait, and the compiler rightfully complains about we don't yet
knowing the computed type. So we need to pass it explicitly, which is
simple enough, if a bit annoying.
Differential Revision: https://phabricator.services.mozilla.com/D83816
This does not (yet) upgrade ./rust-toolchain
The warnings:
* dead_code "field is never read"
* redundant_semicolons "unnecessary trailing semicolon"
* non_fmt_panic "panic message is not a string literal, this is no longer accepted in Rust 2021"
* unstable_name_collisions "a method with this name may be added to the standard library in the future"
* legacy_derive_helpers "derive helper attribute is used before it is introduced" https://github.com/rust-lang/rust/issues/79202
We have this optimization where, for non-generic structs, we generate just a
clone / move as the ToComputedValue / ToResolvedValue implementation.
This moves the optimization a bit further down, and refines it so that we still
generate all the relevant where clauses that make it sound, that is, that all
the ToComputedValue implementations of the fields return the same type.
Otherwise this wouldn't be sound and the type would need to become generic.
We add an escape hatch (no_field_bound) for fields that need to be cloned but
which don't implement the trait. This is right now only for the RefPtr<> in the
shared font-family list, and a piece of code in PaintWorklet which looks kinda
fishy, and probably should be fixed (but we don't ship it in Firefox and there's
a pre-existing FIXME for servo, so I punted on it for now).
The other thing this patch does is adding a bunch of ToComputedValue /
ToResolvedValue implementations that are trivial and were missing.
Differential Revision: https://phabricator.services.mozilla.com/D67913
No need to build the match body if there are no generics. This should hopefully
save some work too.
Depends on D26289
Differential Revision: https://phabricator.services.mozilla.com/D26290
Otherwise, deriving ToComputedValue and ToAnimatedValue in structs or enums with
other where clauses just doesn't work.
Differential Revision: https://phabricator.services.mozilla.com/D21859
We assume that types such as `<Self as ToComputedValue>::ToComputedValue == Self`
just construct a new value that is just a clone of the original one without any
additional code.
For the traits we derive which methods don't depend on associated types (i.e.
all of them but ToAnimatedValue and ToComputedValue), we now add trait bounds
for the actual field types directly, instead of bounding the type parameters.
For now, only impls for types like in style::values::generics can be derived.
This also needed a few ToComputedValueAsSpecified impls that I would like to
replace by some #[to_computed_value(clone)] attribute, but I think it is ok
to keep it like this for now.
