1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220
// Copyright (C) 2023 Parity Technologies (UK) Ltd. (admin@parity.io)
// This file is a part of the scale-decode crate.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
/*!
`parity-scale-codec` provides a `Decode` trait which allows bytes to be scale decoded into types based on the shape of those types.
This crate builds on this, and allows bytes to be decoded into types based on [`scale_info`] type information, rather than the shape
of the target type. At a high level, this crate just aims to do the reverse of the `scale-encode` crate.
This crate exposes four traits:
- A [`visitor::Visitor`] trait which when implemented on some type, can be used in conjunction with [`visitor::decode_with_visitor`]
to decode SCALE encoded bytes based on some type information into some arbitrary type.
- An [`IntoVisitor`] trait which can be used to obtain the [`visitor::Visitor`] implementation for some type.
- A [`DecodeAsType`] trait which is implemented for types which implement [`IntoVisitor`], and provides a high level interface for
decoding SCALE encoded bytes into some type with the help of a type ID and [`scale_info::PortableRegistry`].
- A [`DecodeAsFields`] trait which when implemented on some type, describes how SCALE encoded bytes can be decoded
into it with the help of a slice of [`PortableField`]'s or [`PortableFieldId`]'s and type registry describing the
shape of the encoded bytes. This is generally only implemented for tuples and structs, since we need a set of fields
to map to the provided slices.
Implementations for many built-in types are also provided for each trait, and the [`macro@DecodeAsType`] macro can be used to
generate the relevant impls on new struct and enum types such that they get a [`DecodeAsType`] impl.
The [`DecodeAsType`] and [`DecodeAsFields`] traits are basically the mirror of `scale-encode`'s `EncodeAsType` and `EncodeAsFields`
traits in terms of their interface.
# Motivation
By de-coupling the shape of a type from how bytes are decoded into it, we make it much more likely that the decoding will succeed,
and are no longer reliant on types having a precise layout in order to be decoded into correctly. Some examples of this follow.
```rust
use codec::Encode;
use scale_decode::DecodeAsType;
use scale_info::{PortableRegistry, TypeInfo};
use std::fmt::Debug;
// We normally expect to have type information to hand, but for our examples
// we construct type info from any type that implements `TypeInfo`.
fn get_type_info<T: TypeInfo + 'static>() -> (u32, PortableRegistry) {
let m = scale_info::MetaType::new::<T>();
let mut types = scale_info::Registry::new();
let ty = types.register_type(&m);
let portable_registry: PortableRegistry = types.into();
(ty.id(), portable_registry)
}
// Encode the left value statically.
// Decode those bytes into the right type via `DecodeAsType`.
// Assert that the decoded bytes are identical to the right value.
fn assert_decodes_to<A, B>(a: A, b: B)
where
A: Encode + TypeInfo + 'static,
B: DecodeAsType + PartialEq + Debug,
{
let (type_id, types) = get_type_info::<A>();
let a_bytes = a.encode();
let new_b = B::decode_as_type(&mut &*a_bytes, type_id, &types).unwrap();
assert_eq!(b, new_b);
}
// Start simple; a u8 can DecodeAsType into a u64 and vice versa. Numbers will all
// try to convert into the desired output size, failing if this isn't possible:
assert_decodes_to(123u64, 123u8);
assert_decodes_to(123u8, 123u64);
// Compact decoding is also handled "under the hood" by DecodeAsType, so no "compact"
// annotations are needed on values.
assert_decodes_to(codec::Compact(123u64), 123u64);
// Enum variants are lined up by variant name, so no explicit "index" annotation are
// needed either; DecodeAsType will take care of it.
#[derive(Encode, TypeInfo)]
enum Foo {
#[codec(index = 10)]
Something(u64),
}
#[derive(DecodeAsType, PartialEq, Debug)]
enum FooTarget {
Something(u128),
}
assert_decodes_to(Foo::Something(123), FooTarget::Something(123));
// DecodeAsType will skip annotated fields and not look for them in the encoded bytes.
// #[codec(skip)] and #[decode_as_type(skip)] both work.
#[derive(Encode, TypeInfo)]
struct Bar {
a: bool,
}
#[derive(DecodeAsType, PartialEq, Debug)]
struct BarTarget {
a: bool,
#[decode_as_type(skip)]
b: String,
}
assert_decodes_to(
Bar { a: true },
BarTarget { a: true, b: String::new() },
);
// DecodeAsType impls will generally skip through any newtype wrappers.
#[derive(DecodeAsType, Encode, TypeInfo, PartialEq, Debug)]
struct Wrapper {
value: u64
}
assert_decodes_to(
(Wrapper { value: 123 },),
123u64
);
assert_decodes_to(
123u64,
(123,)
);
// Things like arrays and sequences are generally interchangeable despite the
// encoding format being slightly different:
assert_decodes_to([1u8,2,3,4,5], vec![1u64,2,3,4,5]);
assert_decodes_to(vec![1u64,2,3,4,5], [1u8,2,3,4,5]);
```
If this high level interface isn't suitable, you can implement your own [`visitor::Visitor`]'s. These can support zero-copy decoding
(unlike the higher level [`DecodeAsType`] interface), and generally the Visitor construction and execution is zero alloc, allowing
for efficient type based decoding.
*/
#![deny(missing_docs)]
mod impls;
mod utils;
pub mod error;
pub mod visitor;
pub use crate::error::Error;
pub use visitor::Visitor;
#[cfg(feature = "derive")]
pub use scale_decode_derive::DecodeAsType;
// Used in trait definitions.
pub use scale_info::PortableRegistry;
/// A description of a single field in a tuple or struct type.
pub type PortableField = scale_info::Field<scale_info::form::PortableForm>;
/// A type ID used to represent tuple fields.
pub type PortableFieldId = scale_info::interner::UntrackedSymbol<std::any::TypeId>;
/// This trait is implemented for any type `T` where `T` implements [`IntoVisitor`] and the errors returned
/// from this [`Visitor`] can be converted into [`Error`]. It's essentially a convenience wrapper around
/// [`visitor::decode_with_visitor`] that mirrors `scale-encode`'s `EncodeAsType`.
pub trait DecodeAsType: Sized {
/// Given some input bytes, a `type_id`, and type registry, attempt to decode said bytes into
/// `Self`. Implementations should modify the `&mut` reference to the bytes such that any bytes
/// not used in the course of decoding are still pointed to after decoding is complete.
fn decode_as_type(
input: &mut &[u8],
type_id: u32,
types: &PortableRegistry,
) -> Result<Self, Error>;
}
impl<T> DecodeAsType for T
where
T: IntoVisitor,
Error: From<<T::Visitor as Visitor>::Error>,
{
fn decode_as_type(
input: &mut &[u8],
type_id: u32,
types: &scale_info::PortableRegistry,
) -> Result<Self, Error> {
let res = visitor::decode_with_visitor(input, type_id, types, T::into_visitor())?;
Ok(res)
}
}
/// This is similar to [`DecodeAsType`], except that it's instead implemented for types that can be given a list of
/// fields denoting the type being decoded from and attempt to do this decoding. This is generally implemented just
/// for tuple and struct types, and is automatically implemented via the [`macro@DecodeAsType`] macro.
pub trait DecodeAsFields: Sized {
/// Given some bytes and some fields denoting their structure, attempt to decode.
fn decode_as_fields(
input: &mut &[u8],
fields: &[PortableField],
types: &PortableRegistry,
) -> Result<Self, Error>;
/// Given some bytes and some field IDs denoting their structure, attempt to decode.
fn decode_as_field_ids(
input: &mut &[u8],
field_ids: &[PortableFieldId],
types: &PortableRegistry,
) -> Result<Self, Error> {
// [TODO jsdw]: It would be good to use a more efficient data structure
// here to avoid allocating with smaller numbers of fields.
let fields: Vec<PortableField> =
field_ids.iter().map(|f| PortableField::new(None, *f, None, Vec::new())).collect();
Self::decode_as_fields(input, &fields, types)
}
}
/// This trait can be implemented on any type that has an associated [`Visitor`] responsible for decoding
/// SCALE encoded bytes to it. If you implement this on some type and the [`Visitor`] that you return has
/// an error type that converts into [`Error`], then you'll also get a [`DecodeAsType`] implementation for free.
pub trait IntoVisitor {
/// The visitor type used to decode SCALE encoded bytes to `Self`.
type Visitor: for<'scale, 'info> visitor::Visitor<Value<'scale, 'info> = Self>;
/// A means of obtaining this visitor.
fn into_visitor() -> Self::Visitor;
}