peepmatic: Introduce the peepmatic-macro crate

This crate provides the derive macros used by `peepmatic`, notable AST-related derives that enumerate child AST nodes, and operator-related derives that provide helpers for type checking.
2020-05-01 15:33:55 -07:00
parent c82326a1ae
commit 0f03a97475
6 changed files with 679 additions and 0 deletions
--- a/cranelift/peepmatic/crates/macro/Cargo.toml
+++ b/cranelift/peepmatic/crates/macro/Cargo.toml
@@ -0,0 +1,15 @@
+[package]
+name = "peepmatic-macro"
+version = "0.1.0"
+authors = ["Nick Fitzgerald <fitzgen@gmail.com>"]
+edition = "2018"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+proc-macro2 = "1.0.9"
+quote = "1.0.3"
+syn = { version = "1.0.16", features = ['extra-traits'] }
+
+[lib]
+proc_macro = true
--- a/cranelift/peepmatic/crates/macro/src/child_nodes.rs
+++ b/cranelift/peepmatic/crates/macro/src/child_nodes.rs
@@ -0,0 +1,110 @@
+use quote::quote;
+use syn::DeriveInput;
+use syn::{parse_quote, GenericParam, Generics, Result};
+
+pub fn derive_child_nodes(input: &DeriveInput) -> Result<impl quote::ToTokens> {
+    let children = get_child_nodes(&input.data)?;
+    let name = &input.ident;
+    let generics = add_trait_bounds(input.generics.clone());
+    let (impl_generics, ty_generics, where_clause) = generics.split_for_impl();
+
+    Ok(quote! {
+        impl #impl_generics ChildNodes<'a, 'a> for #name #ty_generics #where_clause {
+            fn child_nodes(&'a self, children: &mut impl Extend<DynAstRef<'a>>) {
+                #children
+            }
+        }
+    })
+}
+
+fn get_child_nodes(data: &syn::Data) -> Result<impl quote::ToTokens> {
+    match data {
+        syn::Data::Struct(s) => {
+            let mut fields = vec![];
+
+            match &s.fields {
+                syn::Fields::Named(n) => {
+                    for f in n.named.iter() {
+                        let opts = crate::PeepmaticOpts::from_attrs(&mut f.attrs.clone())?;
+
+                        if opts.skip_child {
+                            continue;
+                        }
+
+                        let field_name = f.ident.as_ref().unwrap();
+                        if opts.flatten {
+                            fields.push(quote! {
+                                self.#field_name.iter().map(DynAstRef::from)
+                            });
+                        } else {
+                            fields.push(quote! {
+                                std::iter::once(DynAstRef::from(&self.#field_name))
+                            });
+                        }
+                    }
+                }
+                syn::Fields::Unnamed(u) => {
+                    for (i, f) in u.unnamed.iter().enumerate() {
+                        let opts = crate::PeepmaticOpts::from_attrs(&mut f.attrs.clone())?;
+                        if opts.skip_child {
+                            continue;
+                        }
+                        if opts.flatten {
+                            return Err(syn::Error::new(
+                                u.paren_token.span,
+                                "#[peepmatic(flatten)] is only allowed with named fields",
+                            ));
+                        }
+                        fields.push(quote! {
+                            std::iter::once(DynAstRef::from(&self.#i))
+                        });
+                    }
+                }
+                syn::Fields::Unit => {}
+            }
+
+            Ok(match fields.as_slice() {
+                [] => quote! { let _ = children; },
+                [f, rest @ ..] => {
+                    let rest = rest.iter().map(|f| {
+                        quote! {
+                            .chain(#f)
+                        }
+                    });
+                    quote! {
+                        children.extend( #f #( #rest )* );
+                    }
+                }
+            })
+        }
+        syn::Data::Enum(e) => {
+            let mut match_arms = vec![];
+            for v in e.variants.iter() {
+                match v.fields {
+                    syn::Fields::Unnamed(ref u) if u.unnamed.len() == 1 => {
+                        let variant = &v.ident;
+                        match_arms.push(quote! {
+                            Self::#variant(x) => children.extend(Some(x.into())),
+                        });
+                    }
+                    _ => panic!("#[derive(ChildNodes)] only supports enums whose variants all ahve a single unnamed field")
+                }
+            }
+            Ok(quote! {
+                match self {
+                    #( #match_arms )*
+                }
+            })
+        }
+        syn::Data::Union(_) => panic!("#[derive(ChildNodes)] is not supported on unions"),
+    }
+}
+
+fn add_trait_bounds(mut generics: Generics) -> Generics {
+    for param in &mut generics.params {
+        if let GenericParam::Type(type_param) = param {
+            type_param.bounds.push(parse_quote!(ChildNodes<'a, 'a>));
+        }
+    }
+    generics
+}
--- a/cranelift/peepmatic/crates/macro/src/into_dyn_ast_ref.rs
+++ b/cranelift/peepmatic/crates/macro/src/into_dyn_ast_ref.rs
@@ -0,0 +1,23 @@
+use quote::quote;
+use syn::DeriveInput;
+use syn::Result;
+
+pub fn derive_into_dyn_ast_ref(input: &DeriveInput) -> Result<impl quote::ToTokens> {
+    let ty = &input.ident;
+
+    let opts = crate::PeepmaticOpts::from_attrs(&mut input.attrs.clone())?;
+    if opts.no_into_dyn_node {
+        return Ok(quote! {});
+    }
+
+    let (impl_generics, ty_generics, where_clause) = input.generics.split_for_impl();
+
+    Ok(quote! {
+        impl #impl_generics From<&'a #ty #ty_generics> for DynAstRef<'a> #where_clause {
+            #[inline]
+            fn from(x: &'a #ty #ty_generics) -> Self {
+                Self::#ty(x)
+            }
+        }
+    })
+}
--- a/cranelift/peepmatic/crates/macro/src/lib.rs
+++ b/cranelift/peepmatic/crates/macro/src/lib.rs
@@ -0,0 +1,156 @@
+extern crate proc_macro;
+
+use crate::proc_macro::TokenStream;
+use proc_macro2::Span;
+use quote::quote;
+use syn::parse::{Parse, ParseStream};
+use syn::punctuated::Punctuated;
+use syn::DeriveInput;
+use syn::Error;
+use syn::{parse_macro_input, Ident, Result};
+
+mod child_nodes;
+mod into_dyn_ast_ref;
+mod operator;
+mod span;
+
+#[proc_macro_derive(PeepmaticOperator, attributes(peepmatic))]
+pub fn operator(input: TokenStream) -> TokenStream {
+    operator::derive_operator(input)
+}
+
+#[proc_macro_derive(Ast, attributes(peepmatic))]
+pub fn derive_ast(input: TokenStream) -> TokenStream {
+    let input = parse_macro_input!(input as DeriveInput);
+
+    let span_impl = match span::derive_span(&input) {
+        Ok(s) => s,
+        Err(e) => return e.to_compile_error().into(),
+    };
+
+    let child_nodes_impl = match child_nodes::derive_child_nodes(&input) {
+        Ok(c) => c,
+        Err(e) => return e.to_compile_error().into(),
+    };
+
+    let into_dyn_ast_ref_impl = match into_dyn_ast_ref::derive_into_dyn_ast_ref(&input) {
+        Ok(n) => n,
+        Err(e) => return e.to_compile_error().into(),
+    };
+
+    let expanded = quote! {
+        #span_impl
+        #child_nodes_impl
+        #into_dyn_ast_ref_impl
+    };
+
+    TokenStream::from(expanded)
+}
+
+#[derive(Default)]
+pub(crate) struct PeepmaticOpts {
+    // `ChildNodes` options.
+    skip_child: bool,
+    flatten: bool,
+
+    // `From<&'a Self> for DynAstRef<'a>` options.
+    no_into_dyn_node: bool,
+
+    // Peepmatic operator options.
+    immediates_paren: syn::token::Paren,
+    immediates: Vec<syn::Ident>,
+    params_paren: syn::token::Paren,
+    params: Vec<syn::Ident>,
+    result: Option<syn::Ident>,
+}
+
+impl Parse for PeepmaticOpts {
+    fn parse(input: ParseStream) -> Result<Self> {
+        enum Attr {
+            Immediates(syn::token::Paren, Vec<syn::Ident>),
+            Params(syn::token::Paren, Vec<syn::Ident>),
+            Result(syn::Ident),
+            NoIntoDynNode,
+            SkipChild,
+            Flatten,
+        }
+
+        let attrs = Punctuated::<_, syn::token::Comma>::parse_terminated(input)?;
+        let mut ret = PeepmaticOpts::default();
+        for attr in attrs {
+            match attr {
+                Attr::Immediates(paren, imms) => {
+                    ret.immediates_paren = paren;
+                    ret.immediates = imms;
+                }
+                Attr::Params(paren, ps) => {
+                    ret.params_paren = paren;
+                    ret.params = ps;
+                }
+                Attr::Result(r) => ret.result = Some(r),
+                Attr::NoIntoDynNode => ret.no_into_dyn_node = true,
+                Attr::SkipChild => ret.skip_child = true,
+                Attr::Flatten => ret.flatten = true,
+            }
+        }
+
+        return Ok(ret);
+
+        impl Parse for Attr {
+            fn parse(input: ParseStream) -> Result<Self> {
+                let attr: Ident = input.parse()?;
+                if attr == "immediates" {
+                    let inner;
+                    let paren = syn::parenthesized!(inner in input);
+                    let imms = Punctuated::<_, syn::token::Comma>::parse_terminated(&inner)?;
+                    return Ok(Attr::Immediates(paren, imms.into_iter().collect()));
+                }
+                if attr == "params" {
+                    let inner;
+                    let paren = syn::parenthesized!(inner in input);
+                    let params = Punctuated::<_, syn::token::Comma>::parse_terminated(&inner)?;
+                    return Ok(Attr::Params(paren, params.into_iter().collect()));
+                }
+                if attr == "result" {
+                    let inner;
+                    syn::parenthesized!(inner in input);
+                    return Ok(Attr::Result(syn::Ident::parse(&inner)?));
+                }
+                if attr == "skip_child" {
+                    return Ok(Attr::SkipChild);
+                }
+                if attr == "no_into_dyn_node" {
+                    return Ok(Attr::NoIntoDynNode);
+                }
+                if attr == "flatten" {
+                    return Ok(Attr::Flatten);
+                }
+                return Err(Error::new(attr.span(), "unexpected attribute"));
+            }
+        }
+    }
+}
+
+fn peepmatic_attrs(attrs: &mut Vec<syn::Attribute>) -> TokenStream {
+    let mut ret = proc_macro2::TokenStream::new();
+    let ident = syn::Path::from(syn::Ident::new("peepmatic", Span::call_site()));
+    for i in (0..attrs.len()).rev() {
+        if attrs[i].path != ident {
+            continue;
+        }
+        let attr = attrs.remove(i);
+        let group = match attr.tokens.into_iter().next().unwrap() {
+            proc_macro2::TokenTree::Group(g) => g,
+            _ => panic!("#[peepmatic(...)] expected"),
+        };
+        ret.extend(group.stream());
+        ret.extend(quote! { , });
+    }
+    return ret.into();
+}
+
+impl PeepmaticOpts {
+    pub(crate) fn from_attrs(attrs: &mut Vec<syn::Attribute>) -> syn::Result<Self> {
+        syn::parse(peepmatic_attrs(attrs))
+    }
+}
--- a/cranelift/peepmatic/crates/macro/src/operator.rs
+++ b/cranelift/peepmatic/crates/macro/src/operator.rs
@@ -0,0 +1,325 @@
+//! Implementation of the `#[peepmatic]` macro for the `Operator` AST node.
+
+use crate::proc_macro::TokenStream;
+use crate::PeepmaticOpts;
+use proc_macro2::{Ident, Span};
+use quote::quote;
+use syn::DeriveInput;
+use syn::Error;
+use syn::{parse_macro_input, Result};
+
+pub fn derive_operator(input: TokenStream) -> TokenStream {
+    let input = parse_macro_input!(input as DeriveInput);
+
+    let variants = match get_enum_variants(&input) {
+        Ok(v) => v,
+        Err(e) => return e.to_compile_error().into(),
+    };
+
+    let arity = match create_arity(&variants) {
+        Ok(a) => a,
+        Err(e) => return e.to_compile_error().into(),
+    };
+
+    let num_operators = variants.len();
+    let type_methods = create_type_methods(&variants);
+    let parse_impl = create_parse_impl(&input.ident, &variants);
+    let display_impl = create_display_impl(&input.ident, &variants);
+    let try_from_u32_impl = create_try_from_u32_impl(&input.ident, &variants);
+    let ident = &input.ident;
+
+    let expanded = quote! {
+        impl #ident {
+            #arity
+            #type_methods
+
+            /// Get the total number of different operators.
+            pub const fn num_operators() -> usize {
+                #num_operators
+            }
+        }
+
+        #display_impl
+        #try_from_u32_impl
+        #parse_impl
+    };
+
+    // eprintln!("{}", expanded);
+    TokenStream::from(expanded)
+}
+
+fn get_enum_variants(input: &DeriveInput) -> Result<Vec<OperatorVariant>> {
+    let en = match &input.data {
+        syn::Data::Enum(en) => en,
+        syn::Data::Struct(_) => {
+            panic!("can only put #[peepmatic] on an enum; found it on a struct")
+        }
+        syn::Data::Union(_) => panic!("can only put #[peepmatic] on an enum; found it on a union"),
+    };
+    en.variants
+        .iter()
+        .cloned()
+        .map(|mut variant| {
+            Ok(OperatorVariant {
+                opts: PeepmaticOpts::from_attrs(&mut variant.attrs)?,
+                syn: variant,
+            })
+        })
+        .collect()
+}
+
+struct OperatorVariant {
+    syn: syn::Variant,
+    opts: PeepmaticOpts,
+}
+
+fn create_arity(variants: &[OperatorVariant]) -> Result<impl quote::ToTokens> {
+    let mut imm_arities = vec![];
+    let mut params_arities = vec![];
+
+    for v in variants {
+        let variant = &v.syn.ident;
+
+        let imm_arity = v.opts.immediates.len();
+        if imm_arity > std::u8::MAX as usize {
+            return Err(Error::new(
+                v.opts.immediates_paren.span,
+                "cannot have more than u8::MAX immediates",
+            ));
+        }
+        let imm_arity = imm_arity as u8;
+
+        imm_arities.push(quote! {
+            Self::#variant => #imm_arity,
+        });
+
+        let params_arity = v.opts.params.len();
+        if params_arity > std::u8::MAX as usize {
+            return Err(Error::new(
+                v.opts.params_paren.span,
+                "cannot have more than u8::MAX params",
+            ));
+        }
+        let params_arity = params_arity as u8;
+
+        params_arities.push(quote! {
+            Self::#variant => #params_arity,
+        });
+    }
+
+    Ok(quote! {
+        /// Get the number of immediates that this operator has.
+        pub fn immediates_arity(&self) -> u8 {
+            match *self {
+                #( #imm_arities )*
+            }
+        }
+
+        /// Get the number of parameters that this operator takes.
+        pub fn params_arity(&self) -> u8 {
+            match *self {
+                #( #params_arities )*
+            }
+        }
+    })
+}
+
+fn create_type_methods(variants: &[OperatorVariant]) -> impl quote::ToTokens {
+    let mut result_types = vec![];
+    let mut imm_types = vec![];
+    let mut param_types = vec![];
+
+    for v in variants {
+        let variant = &v.syn.ident;
+
+        let result_ty = v.opts.result.as_ref().unwrap_or_else(|| {
+            panic!(
+                "must define #[peepmatic(result(..))] on operator `{}`",
+                variant
+            )
+        });
+        result_types.push(quote! {
+            Self::#variant => {
+                context.#result_ty(span)
+            }
+        });
+
+        let imm_tys = match &v.opts.immediates[..] {
+            [] => quote! {},
+            [ty, rest @ ..] => {
+                let rest = rest.iter().map(|ty| {
+                    quote! { .chain(::std::iter::once(context.#ty(span))) }
+                });
+                quote! {
+                    types.extend(::std::iter::once(context.#ty(span))#( #rest )*);
+                }
+            }
+        };
+        imm_types.push(quote! {
+            Self::#variant => {
+                #imm_tys
+            }
+        });
+
+        let param_tys = match &v.opts.params[..] {
+            [] => quote! {},
+            [ty, rest @ ..] => {
+                let rest = rest.iter().map(|ty| {
+                    quote! { .chain(::std::iter::once(context.#ty(span))) }
+                });
+                quote! {
+                    types.extend(::std::iter::once(context.#ty(span))#( #rest )*);
+                }
+            }
+        };
+        param_types.push(quote! {
+            Self::#variant => {
+                #param_tys
+            }
+        });
+    }
+
+    quote! {
+        /// Get the result type of this operator.
+        #[cfg(feature = "construct")]
+        pub fn result_type<'a, C>(
+            &self,
+            context: &mut C,
+            span: wast::Span,
+        ) -> C::TypeVariable
+        where
+            C: 'a + TypingContext<'a>,
+        {
+            match *self {
+                #( #result_types )*
+            }
+        }
+
+        /// Get the immediate types of this operator.
+        #[cfg(feature = "construct")]
+        pub fn immediate_types<'a, C>(
+            &self,
+            context: &mut C,
+            span: wast::Span,
+            types: &mut impl Extend<C::TypeVariable>,
+        )
+        where
+            C: 'a + TypingContext<'a>,
+        {
+            match *self {
+                #( #imm_types )*
+            }
+        }
+
+        /// Get the parameter types of this operator.
+        #[cfg(feature = "construct")]
+        pub fn param_types<'a, C>(
+            &self,
+            context: &mut C,
+            span: wast::Span,
+            types: &mut impl Extend<C::TypeVariable>,
+        )
+        where
+            C: 'a + TypingContext<'a>,
+        {
+            match *self {
+                #( #param_types )*
+            }
+        }
+    }
+}
+
+fn snake_case(s: &str) -> String {
+    let mut t = String::with_capacity(s.len() + 1);
+    for (i, ch) in s.chars().enumerate() {
+        if i != 0 && ch.is_uppercase() {
+            t.push('_');
+        }
+        t.extend(ch.to_lowercase());
+    }
+    t
+}
+
+fn create_parse_impl(ident: &syn::Ident, variants: &[OperatorVariant]) -> impl quote::ToTokens {
+    let token_defs = variants.iter().map(|v| {
+        let tok = snake_case(&v.syn.ident.to_string());
+        let tok = Ident::new(&tok, Span::call_site());
+        quote! {
+            wast::custom_keyword!(#tok);
+        }
+    });
+
+    let parses = variants.iter().map(|v| {
+        let tok = snake_case(&v.syn.ident.to_string());
+        let tok = Ident::new(&tok, Span::call_site());
+        let ident = &v.syn.ident;
+        quote! {
+            if p.peek::<#tok>() {
+                p.parse::<#tok>()?;
+                return Ok(Self::#ident);
+            }
+        }
+    });
+
+    let expected = format!("expected {}", ident);
+
+    quote! {
+        #[cfg(feature = "construct")]
+        impl<'a> wast::parser::Parse<'a> for #ident {
+            fn parse(p: wast::parser::Parser<'a>) -> wast::parser::Result<Self> {
+                #( #token_defs )*
+
+                #( #parses )*
+
+                Err(p.error(#expected))
+            }
+        }
+    }
+}
+
+fn create_display_impl(ident: &syn::Ident, variants: &[OperatorVariant]) -> impl quote::ToTokens {
+    let displays = variants.iter().map(|v| {
+        let variant = &v.syn.ident;
+        let snake = snake_case(&v.syn.ident.to_string());
+        quote! {
+            Self::#variant => write!(f, #snake),
+        }
+    });
+
+    quote! {
+        impl std::fmt::Display for #ident {
+            fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
+                match self {
+                    #( #displays )*
+                }
+            }
+        }
+    }
+}
+
+fn create_try_from_u32_impl(
+    ident: &syn::Ident,
+    variants: &[OperatorVariant],
+) -> impl quote::ToTokens {
+    let matches = variants.iter().map(|v| {
+        let variant = &v.syn.ident;
+        quote! {
+            x if Self::#variant as u32 == x => Ok(Self::#variant),
+        }
+    });
+
+    let error_msg = format!("value is not an `{}`", ident);
+
+    quote! {
+        impl std::convert::TryFrom<u32> for #ident {
+            type Error = &'static str;
+
+            fn try_from(value: u32) -> Result<Self, Self::Error> {
+                match value {
+                    #( #matches )*
+                    _ => Err(#error_msg)
+                }
+            }
+        }
+    }
+}
--- a/cranelift/peepmatic/crates/macro/src/span.rs
+++ b/cranelift/peepmatic/crates/macro/src/span.rs
@@ -0,0 +1,50 @@
+use quote::quote;
+use syn::DeriveInput;
+use syn::{parse_quote, GenericParam, Generics, Result};
+
+pub fn derive_span(input: &DeriveInput) -> Result<impl quote::ToTokens> {
+    let ty = &input.ident;
+
+    let body = match &input.data {
+        syn::Data::Struct(_) => quote! { self.span },
+        syn::Data::Enum(e) => {
+            let variants = e.variants.iter().map(|v| match v.fields {
+                syn::Fields::Unnamed(ref fields) if fields.unnamed.len() == 1 => {
+                    let variant = &v.ident;
+                    quote! { #ty::#variant(x) => x.span(), }
+                }
+                _ => panic!(
+                    "derive(Ast) on enums only supports variants with a single, unnamed field"
+                ),
+            });
+            quote! {
+                match self {
+                    #( #variants )*
+                }
+            }
+        }
+        syn::Data::Union(_) => panic!("derive(Ast) can only be used with structs and enums, not unions"),
+    };
+
+    let generics = add_span_trait_bounds(input.generics.clone());
+    let (impl_generics, ty_generics, where_clause) = generics.split_for_impl();
+
+    Ok(quote! {
+        impl #impl_generics Span for #ty #ty_generics #where_clause {
+            #[inline]
+            fn span(&self) -> wast::Span {
+                #body
+            }
+        }
+    })
+}
+
+// Add a bound `T: Span` to every type parameter `T`.
+fn add_span_trait_bounds(mut generics: Generics) -> Generics {
+    for param in &mut generics.params {
+        if let GenericParam::Type(ref mut type_param) = *param {
+            type_param.bounds.push(parse_quote!(Span));
+        }
+    }
+    generics
+}