Rust AST
Examples of TokenStreams and DerivedInput objects in Rust using Syn crate.
Rust Compilation Process
The rust compilation process looks roughly like this. A lexical parser does the tokenization, then rustc goes through a couple of intermediary representations that do type solving and optimizations. Finally the IR is passed onto LLVM to generate output binaries.
More rustc details here
Macros
When you start working with Rust Macros you will likely need to interact with AST (abstract syntax tree) objects. You can use the Syn crate to parse an incoming TokenStream into a DerivedInput and then the Quote crate to turn it back into a TokenStream. These crates pass around TokenStream from Proc_Macro2 crate which adds extra functionality while we're still operating on the AST.
In this post I show two types (a struct and an enum) as well as their TokenStream and DeriveInput information. You can convert between the two like this:
let ast: &syn::DeriveInput = syn::parse(input).unwrap();
Struct
struct Vector3 {
x: f32,
y: f32,
z: f32
}Vector3 struct expands to the below Derived Input. Ident members give us access to the names of things struct, Vector3, x, y, z.
It's also worth noting that structs use Brace { and } and can be seen as delimiter: Brace. Tuples are similar but instead use delimiter: Parenthesis ( and ).
TokenStream [
Ident {
ident: "struct",
span: #0 bytes(143..149)
},
Ident {
ident: "Vector3",
span: #0 bytes(150..157)
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "x",
span: #0 bytes(164..165)
},
Punct {
ch: ':',
spacing: Alone,
span: #0 bytes(165..166)
}, Ident {
ident: "f32",
span: #0 bytes(167..170)
},
Punct {
ch: ',',
spacing: Alone,
span: #0 bytes(170..171)
},
Ident {
ident: "y",
span: #0 bytes(176..177)
},
Punct {
ch: ':',
spacing: Alone,
span: #0 bytes(177..178)
},
Ident {
ident: "f32",
span: #0 bytes(179..182)
},
Punct {
ch: ',',
spacing: Alone,
span: #0 bytes(182..183)
},
Ident {
ident: "z",
span: #0 bytes(188..189)
},
Punct {
ch: ':',
spacing: Alone,
span: #0 bytes(189..190)
},
Ident {
ident: "f32",
span: #0 bytes(191..194)
}
],
span: #0 bytes(158..196)
}
]DeriveInput {
attrs: [],
vis: Inherited,
ident: Ident {
ident: "Vector3",
span: #0 bytes(150..157)
},
generics: Generics {
lt_token: None,
params: [],
gt_token: None,
where_clause: None
},
data: Struct(DataStruct {
struct_token: Struct,
fields: Named(FieldsNamed {
brace_token: Brace,
named: [
Field {
attrs: [],
vis: Inherited,
ident: Some(Ident {
ident: "x",
span: #0 bytes(164..165)
}),
colon_token: Some(Colon),
ty: Path(TypePath {
qself: None,
path: Path {
leading_colon: None,
segments: [
PathSegment {
ident: Ident {
ident: "f32",
span: #0 bytes(167..170)
},
arguments: None
}
]
}
})
},
Comma,
Field {
attrs: [],
vis: Inherited,
ident: Some(Ident {
ident: "y",
span: #0 bytes(176..177)
}),
colon_token: Some(Colon),
ty: Path(TypePath {
qself: None,
path: Path {
leading_colon: None,
segments: [
PathSegment {
ident: Ident {
ident: "f32",
span: #0 bytes(179..182)
},
arguments: None
}
]
}
})
},
Comma,
Field {
attrs: [],
vis: Inherited,
ident: Some(Ident {
ident: "z",
span: #0 bytes(188..189)
}),
colon_token: Some(Colon),
ty: Path(TypePath {
qself: None,
path: Path {
leading_colon: None,
segments: [
PathSegment {
ident: Ident {
ident: "f32",
span: #0 bytes(191..194)
},
arguments: None
}
]
}
})
}
]
}),
semi_token: None
})
}Enum
enum MyEnum {
VariantA,
VariantB(i32),
VariantC{x: i32, y: i32, z:i32},
VariantD(f32, f32)
}This one expands quite a bit, but the key piece of information here is data.variants which provides access to each of the Enums Variants.
TokenStream [
Ident {
ident: "enum",
span: #0 bytes(875..879)
},
Ident {
ident: "MyEnum",
span: #0 bytes(880..886)
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "VariantA",
span: #0 bytes(893..901)
},
Punct {
ch: ',',
spacing: Alone,
span: #0 bytes(901..902)
},
Ident {
ident: "VariantB",
span: #0 bytes(907..915)
},
Group {
delimiter: Parenthesis,
stream: TokenStream [
Ident {
ident: "i32",
span: #0 bytes(916..919)
}
],
span: #0 bytes(915..920)
},
Punct {
ch: ',',
spacing: Alone,
span: #0 bytes(920..921)
},
Ident {
ident: "VariantC",
span: #0 bytes(926..934)
},
Group {
delimiter: Brace,
stream: TokenStream [
Ident {
ident: "x",
span: #0 bytes(935..936)
},
Punct {
ch: ':',
spacing: Alone,
span: #0 bytes(936..937)
},
Ident {
ident: "i32",
span: #0 bytes(938..941)
},
Punct {
ch: ',',
spacing: Alone,
span: #0 bytes(941..942)
},
Ident {
ident: "y",
span: #0 bytes(943..944)
},
Punct {
ch: ':',
spacing: Alone,
span: #0 bytes(944..945)
},
Ident {
ident: "i32",
span: #0 bytes(946..949)
},
Punct {
ch: ',',
spacing: Alone,
span: #0 bytes(949..950)
},
Ident {
ident: "z",
span: #0 bytes(951..952)
},
Punct {
ch: ':',
spacing: Alone,
span: #0 bytes(952..953)
},
Ident {
ident: "i32",
span: #0 bytes(953..956)
}
],
span: #0 bytes(934..957)
},
Punct {
ch: ',',
spacing: Alone,
span: #0 bytes(957..958)
},
Ident {
ident: "VariantD",
span: #0 bytes(963..971)
},
Group {
delimiter: Parenthesis,
stream: TokenStream [
Ident {
ident: "f32",
span: #0 bytes(972..975)
},
Punct {
ch: ',',
spacing: Alone,
span: #0 bytes(975..976)
},
Ident {
ident: "f32",
span: #0 bytes(977..980)
}
],
span: #0 bytes(971..981)
}
],
span: #0 bytes(887..983)
}
]DeriveInput {
attrs: [],
vis: Inherited,
ident: Ident {
ident: "MyEnum",
span: #0 bytes(880..886)
},
generics: Generics {
lt_token: None,
params: [],
gt_token: None,
where_clause: None
},
data: Enum(DataEnum {
enum_token: Enum,
brace_token: Brace,
variants: [
Variant {
attrs: [],
ident: Ident {
ident: "VariantA",
span: #0 bytes(893..901)
},
fields: Unit,
discriminant: None
},
Comma,
Variant {
attrs: [],
ident: Ident {
ident: "VariantB",
span: #0 bytes(907..915)
},
fields: Unnamed(FieldsUnnamed {
paren_token: Paren,
unnamed: [
Field {
attrs: [],
vis: Inherited,
ident: None,
colon_token: None,
ty: Path(TypePath {
qself: None,
path: Path {
leading_colon: None,
segments: [
PathSegment {
ident: Ident {
ident: "i32",
span: #0 bytes(916..919)
},
arguments: None
}
]
}
})
}
]
}),
discriminant: None
},
Comma,
Variant {
attrs: [],
ident: Ident {
ident: "VariantC",
span: #0 bytes(926..934)
},
fields: Named(FieldsNamed {
brace_token: Brace,
named: [
Field {
attrs: [],
vis: Inherited,
ident: Some(Ident {
ident: "x",
span: #0 bytes(935..936)
}),
colon_token: Some(Colon),
ty: Path(TypePath {
qself: None,
path: Path {
leading_colon: None,
segments: [
PathSegment {
ident: Ident {
ident: "i32",
span: #0 bytes(938..941)
},
arguments: None
}
]
}
})
},
Comma,
Field {
attrs: [],
vis: Inherited,
ident: Some(Ident {
ident: "y",
span: #0 bytes(943..944)
}),
colon_token: Some(Colon),
ty: Path(TypePath {
qself: None,
path: Path {
leading_colon: None,
segments: [
PathSegment {
ident: Ident {
ident: "i32",
span: #0 bytes(946..949)
},
arguments: None
}
]
}
})
},
Comma,
Field {
attrs: [],
vis: Inherited,
ident: Some(Ident {
ident: "z",
span: #0 bytes(951..952)
}),
colon_token: Some(Colon),
ty: Path(TypePath {
qself: None,
path: Path {
leading_colon: None,
segments: [
PathSegment {
ident: Ident {
ident: "i32",
span: #0 bytes(953..956)
},
arguments: None
}
]
}
})
}
]
}),
discriminant: None
},
Comma,
Variant {
attrs: [],
ident: Ident {
ident: "VariantD",
span: #0 bytes(963..971)
},
fields: Unnamed(FieldsUnnamed {
paren_token: Paren,
unnamed: [
Field {
attrs: [],
vis: Inherited,
ident: None,
colon_token: None,
ty: Path(TypePath {
qself: None,
path: Path {
leading_colon: None,
segments: [
PathSegment {
ident: Ident {
ident: "f32",
span: #0 bytes(972..975)
},
arguments: None
}
]
}
})
},
Comma,
Field {
attrs: [],
vis: Inherited,
ident: None,
colon_token: None,
ty: Path(TypePath {
qself: None,
path: Path {
leading_colon: None,
segments: [PathSegment {
ident: Ident {
ident: "f32",
span: #0 bytes(977..980)
},
arguments: None
}
]
}
})
}
]
}),
discriminant: None
}
]
})
}