`b6f9e8309dbc`

source

xtask/src/sourcegen/kinds.rs9.4 KiBsource history
1#[derive(Debug)]2pub struct KindsSrc {3	/// Key - how this token appears in ungrammar4	defined_tokens: IndexMap<String, TokenKind>,5	defined_node_names: HashSet<String>,6	pub nodes: Vec<String>,7}89#[derive(Debug, Clone)]10pub enum TokenKind {11	/// May exist in token tree, but never in source code12	Meta { grammar_name: String, name: String },13	/// Specific parsing/lexing errors may be emitted as this type of kind14	Error {15		grammar_name: String,16		name: String,17		#[allow(dead_code)]18		/// Is this error returned by lexer directly, or from lex.rs19		is_lexer_error: bool,20		regex: Option<String>,21		priority: Option<u32>,22		description: String,23	},24	/// Keyword - literal match of token25	Keyword {26		/// How this keyword appears in grammar/code, should be same as Kinds key27		code: String,28		name: String,29	},30	/// Literal - something defined by user, i.e strings, identifiers, smth31	Literal {32		/// How this keyword appears in grammar, should be same as Kinds key33		grammar_name: String,34		name: String,35		/// Regex for Logos lexer36		regex: String,37		/// Path to custom lexer38		lexer: Option<String>,39	},40}4142impl TokenKind {43	pub fn grammar_name(&self) -> &str {44		match self {45			Self::Keyword { code, .. } => code,46			Self::Literal { grammar_name, .. }47			| Self::Meta { grammar_name, .. }48			| Self::Error { grammar_name, .. } => grammar_name,49		}50	}51	/// How this keyword should appear in kinds enum, screaming snake cased52	pub fn name(&self) -> &str {53		match self {54			Self::Keyword { name, .. }55			| Self::Literal { name, .. }56			| Self::Meta { name, .. }57			| Self::Error { name, .. } => name,58		}59	}60	pub fn expand_kind(&self, lexer: bool) -> TokenStream {61		let name = format_ident!("{}", self.name());62		let attr = match self {63			Self::Keyword { code, .. } => quote! {#[token(#code)]},64			Self::Literal { regex, lexer, .. } => {65				let lexer = lexer66					.as_deref()67					.map(TokenStream::from_str)68					.map(|r| r.expect("path is correct"));69				quote! {#[regex(#regex, #lexer)]}70			}71			Self::Error {72				regex, priority, ..73			} if regex.is_some() => {74				let priority = priority.map(|p| quote! {, priority = #p});75				quote! {#[regex(#regex #priority)]}76			}77			_ => quote! {},78		};79		let attr = if lexer {80			attr81		} else {82			quote! {}83		};84		quote! {85			#attr86			#name87		}88	}89	pub fn expand_t_macros(&self) -> Option<TokenStream> {90		match self {91			Self::Keyword { code, name } => {92				let code = escape_token_macro(code);93				let name = format_ident!("{name}");94				Some(quote! {95					[#code] => {$crate::SyntaxKind::#name}96				})97			}98			// Meta items should not appear in T![_]99			_ => None,100		}101	}102103	/// How this token should be referenced in code104	/// Keywords are referenced with `T![_]` macro,105	/// and literals are referenced directly by name106	pub fn reference(&self) -> TokenStream {107		if let Self::Keyword { code, .. } = self {108			let code = escape_token_macro(code);109			quote! {T![#code]}110		} else {111			let name = self.name();112			let ident = format_ident!("{name}");113			quote! {#ident}114		}115	}116117	pub fn display_name(&self) -> String {118		match self {119			Self::Keyword { code, .. } => format!("'{code}'"),120			Self::Literal { name, .. } => match name.as_str() {121				"FLOAT" => "number".to_owned(),122				"IDENT" => "identifier".to_owned(),123				"STRING_DOUBLE" | "STRING_SINGLE" | "STRING_DOUBLE_VERBATIM"124				| "STRING_SINGLE_VERBATIM" | "STRING_BLOCK" => "string".to_owned(),125				"WHITESPACE" => "whitespace".to_owned(),126				"SINGLE_LINE_SLASH_COMMENT" | "SINGLE_LINE_HASH_COMMENT"127				| "MULTI_LINE_COMMENT" => "comment".to_owned(),128				_ => name.to_lowercase(),129			},130			Self::Meta { name, .. } => name.to_lowercase(),131			Self::Error { description, .. } => description.clone(),132		}133	}134135	pub fn method_name(&self) -> Ident {136		match self {137			Self::Keyword { name, .. } => {138				format_ident!("{}_token", name.to_lowercase())139			}140			Self::Literal { name, .. } => {141				format_ident!("{}_lit", name.to_lowercase())142			}143			Self::Meta { name, .. } => format_ident!("{}_meta", name.to_lowercase()),144			Self::Error { name, .. } => format_ident!("{}_error", name.to_lowercase()),145		}146	}147}148149#[macro_export]150macro_rules! define_kinds {151	($into:ident = lit($name:literal) => $regex:literal $(, $lexer:literal)? $(; $($rest:tt)*)?) => {{152		$into.define_token(TokenKind::Literal {153			grammar_name: format!("LIT_{}!", $name),154			name: $name.to_owned(),155			regex: $regex.to_owned(),156			lexer: None $(.or_else(|| Some($lexer.to_string())))?,157		});158		$(define_kinds!($into = $($rest)*))?159	}};160	($into:ident = error($name:literal, $desc:literal $(, priority = $priority:literal)? $(, lexer = $lexer:literal)?) $(=> $regex:literal)? $(; $($rest:tt)*)?) => {{161		{162			let regex = None$(.or(Some($regex.to_owned())))?;163			let priority = None$(.or(Some($priority)))?;164			$into.define_token(TokenKind::Error {165				grammar_name: format!("ERROR_{}!", $name),166				name: format!("ERROR_{}", $name),167				is_lexer_error: false $(|| $lexer)? || regex.is_some() || priority.is_some(),168				regex,169				priority,170				description: $desc.to_owned(),171			});172		}173		$(define_kinds!($into = $($rest)*))?174	}};175	($into:ident = $tok:literal => $name:literal $(; $($rest:tt)*)?) => {{176		$into.define_token(TokenKind::Keyword {177			code: format!("{}", $tok),178			name: $name.to_owned(),179		});180		$(define_kinds!($into = $($rest)*))?181	}};182	($into:ident =) => {{}}183}184use std::{collections::HashSet, str::FromStr};185186use indexmap::IndexMap;187use proc_macro2::{Ident, TokenStream};188use quote::{format_ident, quote};189190use super::escape_token_macro;191192impl KindsSrc {193	pub fn new() -> Self {194		Self {195			defined_tokens: IndexMap::new(),196			defined_node_names: HashSet::new(),197			nodes: Vec::new(),198		}199	}200	pub fn define_token(&mut self, token: TokenKind) {201		assert!(202			self.defined_node_names.insert(token.name().to_owned()),203			"node name already defined: {}",204			token.name()205		);206		assert!(207			self.defined_tokens208				.insert(token.grammar_name().to_owned(), token.clone())209				.is_none(),210			"token already defined: {}",211			token.grammar_name()212		);213	}214	pub fn define_node(&mut self, node: &str) {215		assert!(216			self.defined_node_names.insert(node.to_owned()),217			"node name already defined: {node}"218		);219		self.nodes.push(node.to_string());220	}221	pub fn token(&self, tok: &str) -> Option<&TokenKind> {222		self.defined_tokens.get(tok)223	}224	pub fn is_token(&self, tok: &str) -> bool {225		self.defined_tokens.contains_key(tok)226	}227	pub fn tokens(&self) -> impl Iterator<Item = &TokenKind> {228		self.defined_tokens.iter().map(|(_, v)| v)229	}230}231232pub fn jsonnet_kinds() -> KindsSrc {233	let mut kinds = KindsSrc::new();234	define_kinds![kinds =235		"||" => "OR";236		"??" => "NULL_COAELSE";237		"&&" => "AND";238		"|" => "BIT_OR";239		"^" => "BIT_XOR";240		"&" => "BIT_AND";241		"==" => "EQ";242		"!=" => "NE";243		"<" => "LT";244		">" => "GT";245		"<=" => "LE";246		">=" => "GE";247		"<<" => "LHS";248		">>" => "RHS";249		"+" => "PLUS";250		"-" => "MINUS";251		"*" => "MUL";252		"/" => "DIV";253		"%" => "MODULO";254		"!" => "NOT";255		"~" => "BIT_NOT";256		"[" => "L_BRACK";257		"]" => "R_BRACK";258		"(" => "L_PAREN";259		")" => "R_PAREN";260		"{" => "L_BRACE";261		"}" => "R_BRACE";262		":" => "COLON";263		";" => "SEMI";264		"." => "DOT";265		"..." => "DOTDOTDOT";266		"," => "COMMA";267		"$" => "DOLLAR";268		"=" => "ASSIGN";269		"?" => "QUESTION_MARK";270		// Literals271		lit("FLOAT") => r"(?:0|[1-9][0-9]*(?:_[0-9]+)*)(?:\.[0-9]+(?:_[0-9]+)*)?(?:[eE][+-]?[0-9]+(?:_[0-9]+)*)?";272		error("FLOAT_JUNK_AFTER_POINT", "junk after decimal point in number literal") => r"(?:0|[1-9][0-9]*(?:_[0-9]+)*)\.[^0-9]";273		error("FLOAT_JUNK_AFTER_EXPONENT", "junk after exponent in number literal") => r"(?:0|[1-9][0-9]*(?:_[0-9]+)*)(?:\.[0-9]+(?:_[0-9]+)*)?[eE][^+\-0-9]";274		error("FLOAT_JUNK_AFTER_EXPONENT_SIGN", "junk after exponent sign in number literal") => r"(?:0|[1-9][0-9]*(?:_[0-9]+)*)(?:\.[0-9]+(?:_[0-9]+)*)?[eE][+-][^0-9]";275		lit("STRING_DOUBLE") => "\"(?s:[^\"\\\\]|\\\\.)*\"";276		error("STRING_DOUBLE_UNTERMINATED", "unterminated double-quoted string") => "\"(?s:[^\"\\\\]|\\\\.)*";277		lit("STRING_SINGLE") => "'(?s:[^'\\\\]|\\\\.)*'";278		error("STRING_SINGLE_UNTERMINATED", "unterminated single-quoted string") => "'(?s:[^'\\\\]|\\\\.)*";279		lit("STRING_DOUBLE_VERBATIM") => "@\"(?:[^\"]|\"\")*\"";280		error("STRING_DOUBLE_VERBATIM_UNTERMINATED", "unterminated verbatim double-quoted string") => "@\"(?:[^\"]|\"\")*";281		lit("STRING_SINGLE_VERBATIM") => "@'(?:[^']|'')*'";282		error("STRING_SINGLE_VERBATIM_UNTERMINATED", "unterminated verbatim single-quoted string") => "@'(?:[^']|'')*";283		error("STRING_VERBATIM_MISSING_QUOTES", "verbatim string missing opening quotes") => "@[^\"'\\s]\\S+";284		lit("STRING_BLOCK") => r"\|\|\|", "crate::string_block::lex_str_block_test";285		error("STRING_BLOCK_UNEXPECTED_END", "unexpected end of text block", lexer = true);286		error("STRING_BLOCK_MISSING_NEW_LINE", "text block requires new line after |||", lexer = true);287		error("STRING_BLOCK_MISSING_TERMINATION", "unterminated text block", lexer = true);288		error("STRING_BLOCK_MISSING_INDENT", "text block first line must be indented", lexer = true);289		lit("IDENT") => r"[_a-zA-Z][_a-zA-Z0-9]*";290		lit("WHITESPACE") => r"[ \t\n\r]+";291		lit("SINGLE_LINE_SLASH_COMMENT") => r"//[^\r\n]*?(\r\n|\n)?";292		lit("SINGLE_LINE_HASH_COMMENT") => r"#[^\r\n]*?(\r\n|\n)?";293		lit("MULTI_LINE_COMMENT") => r"/\*([^*]|\*[^/])*\*/";294		error("COMMENT_TOO_SHORT", "comment too short") => r"/\*/";295		error("COMMENT_UNTERMINATED", "unterminated multi-line comment") =>  r"/\*([^*/]|\*[^/])+";296		error("NO_OPERATOR", "expected operator");297		error("MISSING_TOKEN", "missing token");298		error("UNEXPECTED_TOKEN", "unexpected token");299		error("CUSTOM", "error");300	];301	kinds302}