Add proc-macro to derive CheckCodePrefix (#1656)

IMO a derive macro is a natural way to generate new code, and it reduces
the chance of merge conflicts.

ruff_macros/Cargo.toml

@@ -7,6 +7,7 @@ edition = "2021"
 proc-macro = true
 
 [dependencies]
+once_cell = { version = "1.17.0" }
 proc-macro2 = { version = "1.0.47" }
 quote = { version = "1.0.21" }
 syn = { version = "1.0.103", features = ["derive", "parsing"] }

ruff_macros/src/check_code_prefix.rs

@@ -0,0 +1,289 @@
+use std::collections::{BTreeMap, BTreeSet, HashMap};
+
+use once_cell::sync::Lazy;
+use proc_macro2::Span;
+use quote::quote;
+use syn::punctuated::Punctuated;
+use syn::token::Comma;
+use syn::{DataEnum, DeriveInput, Ident, Variant};
+
+const ALL: &str = "ALL";
+
+/// A hash map from deprecated `CheckCodePrefix` to latest `CheckCodePrefix`.
+pub static PREFIX_REDIRECTS: Lazy<HashMap<&'static str, &'static str>> = Lazy::new(|| {
+    HashMap::from_iter([
+        // TODO(charlie): Remove by 2023-01-01.
+        ("U001", "UP001"),
+        ("U003", "UP003"),
+        ("U004", "UP004"),
+        ("U005", "UP005"),
+        ("U006", "UP006"),
+        ("U007", "UP007"),
+        ("U008", "UP008"),
+        ("U009", "UP009"),
+        ("U010", "UP010"),
+        ("U011", "UP011"),
+        ("U012", "UP012"),
+        ("U013", "UP013"),
+        ("U014", "UP014"),
+        ("U015", "UP015"),
+        ("U016", "UP016"),
+        ("U017", "UP017"),
+        ("U019", "UP019"),
+        // TODO(charlie): Remove by 2023-02-01.
+        ("I252", "TID252"),
+        ("M001", "RUF100"),
+        // TODO(charlie): Remove by 2023-02-01.
+        ("PDV002", "PD002"),
+        ("PDV003", "PD003"),
+        ("PDV004", "PD004"),
+        ("PDV007", "PD007"),
+        ("PDV008", "PD008"),
+        ("PDV009", "PD009"),
+        ("PDV010", "PD010"),
+        ("PDV011", "PD011"),
+        ("PDV012", "PD012"),
+        ("PDV013", "PD013"),
+        ("PDV015", "PD015"),
+        ("PDV901", "PD901"),
+        // TODO(charlie): Remove by 2023-02-01.
+        ("R501", "RET501"),
+        ("R502", "RET502"),
+        ("R503", "RET503"),
+        ("R504", "RET504"),
+        ("R505", "RET505"),
+        ("R506", "RET506"),
+        ("R507", "RET507"),
+        ("R508", "RET508"),
+        ("IC001", "ICN001"),
+        ("IC002", "ICN001"),
+        ("IC003", "ICN001"),
+        ("IC004", "ICN001"),
+        // TODO(charlie): Remove by 2023-01-01.
+        ("U", "UP"),
+        ("U0", "UP0"),
+        ("U00", "UP00"),
+        ("U01", "UP01"),
+        // TODO(charlie): Remove by 2023-02-01.
+        ("I2", "TID2"),
+        ("I25", "TID25"),
+        ("M", "RUF100"),
+        ("M0", "RUF100"),
+        // TODO(charlie): Remove by 2023-02-01.
+        ("PDV", "PD"),
+        ("PDV0", "PD0"),
+        ("PDV01", "PD01"),
+        ("PDV9", "PD9"),
+        ("PDV90", "PD90"),
+        // TODO(charlie): Remove by 2023-02-01.
+        ("R", "RET"),
+        ("R5", "RET5"),
+        ("R50", "RET50"),
+        // TODO(charlie): Remove by 2023-02-01.
+        ("IC", "ICN"),
+        ("IC0", "ICN0"),
+    ])
+});
+
+pub fn derive_impl(input: DeriveInput) -> syn::Result<proc_macro2::TokenStream> {
+    let DeriveInput { ident, data, .. } = input;
+    let syn::Data::Enum(DataEnum { variants, .. }) = data else {
+        return Err(syn::Error::new(
+            ident.span(),
+            "Can only derive `CheckCodePrefix` from enums.",
+        ));
+    };
+
+    let prefix_ident = Ident::new(&format!("{ident}Prefix"), ident.span());
+    let prefix = expand(&ident, &prefix_ident, &variants);
+    let expanded = quote! {
+        #[derive(PartialEq, Eq, PartialOrd, Ord)]
+        pub enum SuffixLength {
+            None,
+            Zero,
+            One,
+            Two,
+            Three,
+            Four,
+        }
+
+        #prefix
+    };
+    Ok(expanded)
+}
+
+fn expand(
+    ident: &Ident,
+    prefix_ident: &Ident,
+    variants: &Punctuated<Variant, Comma>,
+) -> proc_macro2::TokenStream {
+    // Build up a map from prefix to matching CheckCodes.
+    let mut prefix_to_codes: BTreeMap<Ident, BTreeSet<String>> = BTreeMap::default();
+    for variant in variants {
+        let span = variant.ident.span();
+        let code_str = variant.ident.to_string();
+        let code_prefix_len = code_str
+            .chars()
+            .take_while(|char| char.is_alphabetic())
+            .count();
+        let code_suffix_len = code_str.len() - code_prefix_len;
+        for i in 0..=code_suffix_len {
+            let prefix = code_str[..code_prefix_len + i].to_string();
+            prefix_to_codes
+                .entry(Ident::new(&prefix, span))
+                .or_default()
+                .insert(code_str.clone());
+        }
+        prefix_to_codes
+            .entry(Ident::new(ALL, span))
+            .or_default()
+            .insert(code_str.clone());
+    }
+
+    // Add any prefix aliases (e.g., "U" to "UP").
+    for (alias, check_code) in PREFIX_REDIRECTS.iter() {
+        prefix_to_codes.insert(
+            Ident::new(alias, Span::call_site()),
+            prefix_to_codes
+                .get(&Ident::new(check_code, Span::call_site()))
+                .unwrap_or_else(|| panic!("Unknown CheckCode: {alias:?}"))
+                .clone(),
+        );
+    }
+
+    let prefix_variants = prefix_to_codes.keys().map(|prefix| {
+        quote! {
+            #prefix
+        }
+    });
+
+    let prefix_impl = generate_impls(ident, prefix_ident, &prefix_to_codes);
+
+    let prefix_redirects = PREFIX_REDIRECTS.iter().map(|(alias, check_code)| {
+        let code = Ident::new(check_code, Span::call_site());
+        quote! {
+            (#alias, #prefix_ident::#code)
+        }
+    });
+
+    quote! {
+        #[derive(
+            ::strum_macros::EnumString,
+            ::strum_macros::AsRefStr,
+            Debug,
+            PartialEq,
+            Eq,
+            PartialOrd,
+            Ord,
+            Clone,
+            ::serde::Serialize,
+            ::serde::Deserialize,
+            ::schemars::JsonSchema,
+        )]
+        pub enum #prefix_ident {
+            #(#prefix_variants,)*
+        }
+
+        #prefix_impl
+
+        /// A hash map from deprecated `CheckCodePrefix` to latest `CheckCodePrefix`.
+        pub static PREFIX_REDIRECTS: ::once_cell::sync::Lazy<::rustc_hash::FxHashMap<&'static str, #prefix_ident>> = ::once_cell::sync::Lazy::new(|| {
+            ::rustc_hash::FxHashMap::from_iter([
+                #(#prefix_redirects),*
+            ])
+        });
+    }
+}
+
+fn generate_impls(
+    ident: &Ident,
+    prefix_ident: &Ident,
+    prefix_to_codes: &BTreeMap<Ident, BTreeSet<String>>,
+) -> proc_macro2::TokenStream {
+    let codes_match_arms = prefix_to_codes.iter().map(|(prefix, codes)| {
+        let codes = codes.iter().map(|code| {
+            let code = Ident::new(code, Span::call_site());
+            quote! {
+                #ident::#code
+            }
+        });
+        let prefix_str = prefix.to_string();
+        if let Some(target) = PREFIX_REDIRECTS.get(prefix_str.as_str()) {
+            quote! {
+                #prefix_ident::#prefix => {
+                    crate::one_time_warning!(
+                        "{}{} {}",
+                        "warning".yellow().bold(),
+                        ":".bold(),
+                        format!("`{}` has been remapped to `{}`", #prefix_str, #target).bold()
+                    );
+                    vec![#(#codes),*]
+                }
+            }
+        } else {
+            quote! {
+                #prefix_ident::#prefix => vec![#(#codes),*],
+            }
+        }
+    });
+
+    let specificity_match_arms = prefix_to_codes.keys().map(|prefix| {
+        if *prefix == ALL {
+            quote! {
+                #prefix_ident::#prefix => SuffixLength::None,
+            }
+        } else {
+            let num_numeric = prefix
+                .to_string()
+                .chars()
+                .filter(|char| char.is_numeric())
+                .count();
+            let suffix_len = match num_numeric {
+                0 => quote! { SuffixLength::Zero },
+                1 => quote! { SuffixLength::One },
+                2 => quote! { SuffixLength::Two },
+                3 => quote! { SuffixLength::Three },
+                4 => quote! { SuffixLength::Four },
+                _ => panic!("Invalid prefix: {prefix}"),
+            };
+            quote! {
+                #prefix_ident::#prefix => #suffix_len,
+            }
+        }
+    });
+
+    let categories = prefix_to_codes.keys().map(|prefix| {
+        let prefix_str = prefix.to_string();
+        if prefix_str.chars().all(char::is_alphabetic)
+            && !PREFIX_REDIRECTS.contains_key(&prefix_str.as_str())
+        {
+            quote! {
+                #prefix_ident::#prefix,
+            }
+        } else {
+            quote! {}
+        }
+    });
+
+    quote! {
+        impl #prefix_ident {
+            pub fn codes(&self) -> Vec<#ident> {
+                use colored::Colorize;
+
+                #[allow(clippy::match_same_arms)]
+                match self {
+                    #(#codes_match_arms)*
+                }
+            }
+
+            pub fn specificity(&self) -> SuffixLength {
+                #[allow(clippy::match_same_arms)]
+                match self {
+                    #(#specificity_match_arms)*
+                }
+            }
+        }
+
+        pub const CATEGORIES: &[#prefix_ident] = &[#(#categories)*];
+    }
+}

ruff_macros/src/config.rs

@@ -0,0 +1,203 @@
+use quote::{quote, quote_spanned};
+use syn::parse::{Parse, ParseStream};
+use syn::spanned::Spanned;
+use syn::token::Comma;
+use syn::{
+    AngleBracketedGenericArguments, Attribute, Data, DataStruct, DeriveInput, Field, Fields, Lit,
+    LitStr, Path, PathArguments, PathSegment, Token, Type, TypePath,
+};
+
+pub fn derive_impl(input: DeriveInput) -> syn::Result<proc_macro2::TokenStream> {
+    let DeriveInput { ident, data, .. } = input;
+
+    match data {
+        Data::Struct(DataStruct {
+            fields: Fields::Named(fields),
+            ..
+        }) => {
+            let mut output = vec![];
+
+            for field in fields.named.iter() {
+                let docs: Vec<&Attribute> = field
+                    .attrs
+                    .iter()
+                    .filter(|attr| attr.path.is_ident("doc"))
+                    .collect();
+
+                if docs.is_empty() {
+                    return Err(syn::Error::new(
+                        field.span(),
+                        "Missing documentation for field",
+                    ));
+                }
+
+                if let Some(attr) = field.attrs.iter().find(|attr| attr.path.is_ident("option")) {
+                    output.push(handle_option(field, attr, docs)?);
+                };
+
+                if field
+                    .attrs
+                    .iter()
+                    .any(|attr| attr.path.is_ident("option_group"))
+                {
+                    output.push(handle_option_group(field)?);
+                };
+            }
+
+            Ok(quote! {
+              use crate::settings::options_base::{OptionEntry, OptionField, OptionGroup, ConfigurationOptions};
+
+              #[automatically_derived]
+              impl ConfigurationOptions for #ident {
+                  fn get_available_options() -> Vec<OptionEntry> {
+                      vec![#(#output),*]
+                  }
+              }
+            })
+        }
+        _ => Err(syn::Error::new(
+            ident.span(),
+            "Can only derive ConfigurationOptions from structs with named fields.",
+        )),
+    }
+}
+
+/// For a field with type `Option<Foobar>` where `Foobar` itself is a struct
+/// deriving `ConfigurationOptions`, create code that calls retrieves options
+/// from that group: `Foobar::get_available_options()`
+fn handle_option_group(field: &Field) -> syn::Result<proc_macro2::TokenStream> {
+    let ident = field
+        .ident
+        .as_ref()
+        .expect("Expected to handle named fields");
+
+    match &field.ty {
+        Type::Path(TypePath {
+            path: Path { segments, .. },
+            ..
+        }) => match segments.first() {
+            Some(PathSegment {
+                ident: type_ident,
+                arguments:
+                    PathArguments::AngleBracketed(AngleBracketedGenericArguments { args, .. }),
+                ..
+            }) if type_ident == "Option" => {
+                let path = &args[0];
+                let kebab_name = LitStr::new(&ident.to_string().replace('_', "-"), ident.span());
+
+                Ok(quote_spanned!(
+                    ident.span() => OptionEntry::Group(OptionGroup {
+                        name: #kebab_name,
+                        fields: #path::get_available_options(),
+                    })
+                ))
+            }
+            _ => Err(syn::Error::new(
+                ident.span(),
+                "Expected `Option<_>`  as type.",
+            )),
+        },
+        _ => Err(syn::Error::new(ident.span(), "Expected type.")),
+    }
+}
+
+/// Parse a `doc` attribute into it a string literal.
+fn parse_doc(doc: &Attribute) -> syn::Result<String> {
+    let doc = doc
+        .parse_meta()
+        .map_err(|e| syn::Error::new(doc.span(), e))?;
+
+    match doc {
+        syn::Meta::NameValue(syn::MetaNameValue {
+            lit: Lit::Str(lit_str),
+            ..
+        }) => Ok(lit_str.value()),
+        _ => Err(syn::Error::new(doc.span(), "Expected doc attribute.")),
+    }
+}
+
+/// Parse an `#[option(doc="...", default="...", value_type="...",
+/// example="...")]` attribute and return data in the form of an `OptionField`.
+fn handle_option(
+    field: &Field,
+    attr: &Attribute,
+    docs: Vec<&Attribute>,
+) -> syn::Result<proc_macro2::TokenStream> {
+    // Convert the list of `doc` attributes into a single string.
+    let doc = textwrap::dedent(
+        &docs
+            .into_iter()
+            .map(parse_doc)
+            .collect::<syn::Result<Vec<_>>>()?
+            .join("\n"),
+    )
+    .trim_matches('\n')
+    .to_string();
+
+    let ident = field
+        .ident
+        .as_ref()
+        .expect("Expected to handle named fields");
+
+    let FieldAttributes {
+        default,
+        value_type,
+        example,
+        ..
+    } = attr.parse_args::<FieldAttributes>()?;
+    let kebab_name = LitStr::new(&ident.to_string().replace('_', "-"), ident.span());
+
+    Ok(quote_spanned!(
+        ident.span() => OptionEntry::Field(OptionField {
+            name: #kebab_name,
+            doc: &#doc,
+            default: &#default,
+            value_type: &#value_type,
+            example: &#example,
+        })
+    ))
+}
+
+#[derive(Debug)]
+struct FieldAttributes {
+    default: String,
+    value_type: String,
+    example: String,
+}
+
+impl Parse for FieldAttributes {
+    fn parse(input: ParseStream) -> syn::Result<Self> {
+        let default = _parse_key_value(input, "default")?;
+        input.parse::<Comma>()?;
+        let value_type = _parse_key_value(input, "value_type")?;
+        input.parse::<Comma>()?;
+        let example = _parse_key_value(input, "example")?;
+        if !input.is_empty() {
+            input.parse::<Comma>()?;
+        }
+
+        Ok(FieldAttributes {
+            default,
+            value_type,
+            example: textwrap::dedent(&example).trim_matches('\n').to_string(),
+        })
+    }
+}
+
+fn _parse_key_value(input: ParseStream, name: &str) -> syn::Result<String> {
+    let ident: proc_macro2::Ident = input.parse()?;
+    if ident != name {
+        return Err(syn::Error::new(
+            ident.span(),
+            format!("Expected `{name}` name"),
+        ));
+    }
+
+    input.parse::<Token![=]>()?;
+    let value: Lit = input.parse()?;
+
+    match &value {
+        Lit::Str(v) => Ok(v.value()),
+        _ => Err(syn::Error::new(value.span(), "Expected literal string")),
+    }
+}

ruff_macros/src/lib.rs

@@ -11,215 +11,25 @@
     clippy::too_many_lines
 )]
 
-use quote::{quote, quote_spanned};
-use syn::parse::{Parse, ParseStream};
-use syn::spanned::Spanned;
-use syn::token::Comma;
-use syn::{
-    parse_macro_input, AngleBracketedGenericArguments, Attribute, Data, DataStruct, DeriveInput,
-    Field, Fields, Lit, LitStr, Path, PathArguments, PathSegment, Token, Type, TypePath,
-};
+use syn::{parse_macro_input, DeriveInput};
+
+mod check_code_prefix;
+mod config;
 
 #[proc_macro_derive(ConfigurationOptions, attributes(option, doc, option_group))]
-pub fn derive(input: proc_macro::TokenStream) -> proc_macro::TokenStream {
+pub fn derive_config(input: proc_macro::TokenStream) -> proc_macro::TokenStream {
     let input = parse_macro_input!(input as DeriveInput);
 
-    derive_impl(input)
+    config::derive_impl(input)
         .unwrap_or_else(syn::Error::into_compile_error)
         .into()
 }
 
-fn derive_impl(input: DeriveInput) -> syn::Result<proc_macro2::TokenStream> {
-    let DeriveInput { ident, data, .. } = input;
+#[proc_macro_derive(CheckCodePrefix)]
+pub fn derive_check_code_prefix(input: proc_macro::TokenStream) -> proc_macro::TokenStream {
+    let input = parse_macro_input!(input as DeriveInput);
 
-    match data {
-        Data::Struct(DataStruct {
-            fields: Fields::Named(fields),
-            ..
-        }) => {
-            let mut output = vec![];
-
-            for field in fields.named.iter() {
-                let docs: Vec<&Attribute> = field
-                    .attrs
-                    .iter()
-                    .filter(|attr| attr.path.is_ident("doc"))
-                    .collect();
-
-                if docs.is_empty() {
-                    return Err(syn::Error::new(
-                        field.span(),
-                        "Missing documentation for field",
-                    ));
-                }
-
-                if let Some(attr) = field.attrs.iter().find(|attr| attr.path.is_ident("option")) {
-                    output.push(handle_option(field, attr, docs)?);
-                };
-
-                if field
-                    .attrs
-                    .iter()
-                    .any(|attr| attr.path.is_ident("option_group"))
-                {
-                    output.push(handle_option_group(field)?);
-                };
-            }
-
-            Ok(quote! {
-              use crate::settings::options_base::{OptionEntry, OptionField, OptionGroup, ConfigurationOptions};
-
-              #[automatically_derived]
-              impl ConfigurationOptions for #ident {
-                  fn get_available_options() -> Vec<OptionEntry> {
-                      vec![#(#output),*]
-                  }
-              }
-            })
-        }
-        _ => Err(syn::Error::new(
-            ident.span(),
-            "Can only derive ConfigurationOptions from structs with named fields.",
-        )),
-    }
-}
-
-/// For a field with type `Option<Foobar>` where `Foobar` itself is a struct
-/// deriving `ConfigurationOptions`, create code that calls retrieves options
-/// from that group: `Foobar::get_available_options()`
-fn handle_option_group(field: &Field) -> syn::Result<proc_macro2::TokenStream> {
-    let ident = field
-        .ident
-        .as_ref()
-        .expect("Expected to handle named fields");
-
-    match &field.ty {
-        Type::Path(TypePath {
-            path: Path { segments, .. },
-            ..
-        }) => match segments.first() {
-            Some(PathSegment {
-                ident: type_ident,
-                arguments:
-                    PathArguments::AngleBracketed(AngleBracketedGenericArguments { args, .. }),
-                ..
-            }) if type_ident == "Option" => {
-                let path = &args[0];
-                let kebab_name = LitStr::new(&ident.to_string().replace('_', "-"), ident.span());
-
-                Ok(quote_spanned!(
-                    ident.span() => OptionEntry::Group(OptionGroup {
-                        name: #kebab_name,
-                        fields: #path::get_available_options(),
-                    })
-                ))
-            }
-            _ => Err(syn::Error::new(
-                ident.span(),
-                "Expected `Option<_>`  as type.",
-            )),
-        },
-        _ => Err(syn::Error::new(ident.span(), "Expected type.")),
-    }
-}
-
-/// Parse a `doc` attribute into it a string literal.
-fn parse_doc(doc: &Attribute) -> syn::Result<String> {
-    let doc = doc
-        .parse_meta()
-        .map_err(|e| syn::Error::new(doc.span(), e))?;
-
-    match doc {
-        syn::Meta::NameValue(syn::MetaNameValue {
-            lit: Lit::Str(lit_str),
-            ..
-        }) => Ok(lit_str.value()),
-        _ => Err(syn::Error::new(doc.span(), "Expected doc attribute.")),
-    }
-}
-
-/// Parse an `#[option(doc="...", default="...", value_type="...",
-/// example="...")]` attribute and return data in the form of an `OptionField`.
-fn handle_option(
-    field: &Field,
-    attr: &Attribute,
-    docs: Vec<&Attribute>,
-) -> syn::Result<proc_macro2::TokenStream> {
-    // Convert the list of `doc` attributes into a single string.
-    let doc = textwrap::dedent(
-        &docs
-            .into_iter()
-            .map(parse_doc)
-            .collect::<syn::Result<Vec<_>>>()?
-            .join("\n"),
-    )
-    .trim_matches('\n')
-    .to_string();
-
-    let ident = field
-        .ident
-        .as_ref()
-        .expect("Expected to handle named fields");
-
-    let FieldAttributes {
-        default,
-        value_type,
-        example,
-        ..
-    } = attr.parse_args::<FieldAttributes>()?;
-    let kebab_name = LitStr::new(&ident.to_string().replace('_', "-"), ident.span());
-
-    Ok(quote_spanned!(
-        ident.span() => OptionEntry::Field(OptionField {
-            name: #kebab_name,
-            doc: &#doc,
-            default: &#default,
-            value_type: &#value_type,
-            example: &#example,
-        })
-    ))
-}
-
-#[derive(Debug)]
-struct FieldAttributes {
-    default: String,
-    value_type: String,
-    example: String,
-}
-
-impl Parse for FieldAttributes {
-    fn parse(input: ParseStream) -> syn::Result<Self> {
-        let default = _parse_key_value(input, "default")?;
-        input.parse::<Comma>()?;
-        let value_type = _parse_key_value(input, "value_type")?;
-        input.parse::<Comma>()?;
-        let example = _parse_key_value(input, "example")?;
-        if !input.is_empty() {
-            input.parse::<Comma>()?;
-        }
-
-        Ok(FieldAttributes {
-            default,
-            value_type,
-            example: textwrap::dedent(&example).trim_matches('\n').to_string(),
-        })
-    }
-}
-
-fn _parse_key_value(input: ParseStream, name: &str) -> syn::Result<String> {
-    let ident: proc_macro2::Ident = input.parse()?;
-    if ident != name {
-        return Err(syn::Error::new(
-            ident.span(),
-            format!("Expected `{name}` name"),
-        ));
-    }
-
-    input.parse::<Token![=]>()?;
-    let value: Lit = input.parse()?;
-
-    match &value {
-        Lit::Str(v) => Ok(v.value()),
-        _ => Err(syn::Error::new(value.span(), "Expected literal string")),
-    }
+    check_code_prefix::derive_impl(input)
+        .unwrap_or_else(syn::Error::into_compile_error)
+        .into()
 }