git.delta.rocks / jrsonnet / refs/commits / b3f009bb7b2e

difftreelog

source

crates/jrsonnet-rowan-parser/src/string_block.rs5.8 KiBsourcehistory
1#[derive(Clone, Copy, Debug, PartialEq, Eq)]2pub enum StringBlockError {3	UnexpectedEnd,4	MissingNewLine,5	MissingTermination,6	MissingIndent,7}89use logos::Lexer;10use StringBlockError::*;1112use crate::SyntaxKind;1314pub(crate) fn lex_str_block_test<'d>(lex: &mut Lexer<'d, SyntaxKind>) {15	let _ = lex_str_block(lex);16}1718pub(crate) struct Context<'a> {19	source: &'a str,20	index: usize,21}2223impl<'a> Context<'a> {24	fn rest(&self) -> &'a str {25		&self.source[self.index..]26	}2728	fn next(&mut self) -> Option<char> {29		if self.index == self.source.len() {30			return None;31		}3233		match self.rest().chars().next() {34			None => None,35			Some(c) => {36				self.index += c.len_utf8();37				Some(c)38			}39		}40	}4142	fn peek(&self) -> Option<char> {43		if self.index == self.source.len() {44			return None;45		}4647		self.rest().chars().next()48	}4950	fn eat_if(&mut self, f: impl Fn(char) -> bool) -> usize {51		if self.peek().map(f).unwrap_or(false) {52			self.index += 1;53			return 1;54		}55		056	}5758	fn eat_while(&mut self, f: impl Fn(char) -> bool) -> usize {59		if self.index == self.source.len() {60			return 0;61		}6263		let next_char = self.rest().char_indices().find(|(_, c)| !f(*c));6465		match next_char {66			None => {67				let diff = self.source.len() - self.index;68				self.index = self.source.len();69				diff70			}71			Some((idx, _)) => {72				self.index += idx;73				idx74			}75		}76	}7778	fn skip(&mut self, len: usize) {79		self.index = match self.index + len {80			n if n > self.source.len() => self.source.len(),81			n => n,82		};83	}84}8586// Check that b has at least the same whitespace prefix as a and returns the87// amount of this whitespace, otherwise returns 0.  If a has no whitespace88// prefix than return 0.89fn check_whitespace(a: &str, b: &str) -> usize {90	let a = a.as_bytes();91	let b = b.as_bytes();9293	for i in 0..a.len() {94		if a[i] != b' ' && a[i] != b'\t' {95			// a has run out of whitespace and b matched up to this point. Return result.96			return i;97		}9899		if i >= b.len() {100			// We ran off the edge of b while a still has whitespace. Return 0 as failure.101			return 0;102		}103104		if a[i] != b[i] {105			// a has whitespace but b does not. Return 0 as failure.106			return 0;107		}108	}109110	// We ran off the end of a and b kept up111	a.len()112}113114pub(crate) trait StrBlockLexCtx<'d> {115	fn remainder(&self) -> &'d str;116	fn eat_error(&mut self, ctx: &Context<'d>);117	fn bump_pos(&mut self, s: usize);118	fn mark_truncating(&mut self);119	fn mark_line(&mut self, line: &'d str);120}121122impl<'d> StrBlockLexCtx<'d> for Lexer<'d, SyntaxKind> {123	fn remainder(&self) -> &'d str {124		self.remainder()125	}126	fn eat_error(&mut self, ctx: &Context<'d>) {127		let end_index = ctx128			.rest()129			.find("|||")130			.map_or_else(|| ctx.rest().len(), |v| v + 3);131		self.bump(ctx.index + end_index);132	}133	fn bump_pos(&mut self, s: usize) {134		self.bump(s);135	}136	fn mark_truncating(&mut self) {137		// Lexer test doesn't collect anything138	}139	fn mark_line(&mut self, _line: &'d str) {140		// Lexer test doesn't collect anything141	}142}143144pub fn collect_lexed_str_block<'s>(145	input: &'s str,146) -> Result<CollectStrBlock<'s>, StringBlockError> {147	let mut collect = CollectStrBlock {148		truncate: false,149		lines: vec![],150		input,151		offset: 0,152	};153	lex_str_block(&mut collect)?;154	Ok(collect)155}156157pub struct CollectStrBlock<'s> {158	pub truncate: bool,159	pub lines: Vec<&'s str>,160	input: &'s str,161	offset: usize,162}163164impl<'d> StrBlockLexCtx<'d> for CollectStrBlock<'d> {165	fn remainder(&self) -> &'d str {166		self.input167	}168169	fn eat_error(&mut self, _ctx: &Context<'d>) {170		// Error will be returned, no need to record it here171	}172173	fn bump_pos(&mut self, s: usize) {174		self.offset += s;175	}176177	fn mark_truncating(&mut self) {178		self.truncate = true;179	}180181	fn mark_line(&mut self, line: &'d str) {182		self.lines.push(line)183	}184}185186pub(crate) fn lex_str_block<'a>(lex: &mut impl StrBlockLexCtx<'a>) -> Result<(), StringBlockError> {187	// debug_assert_eq!(lex.slice(), "|||");188	let mut ctx = Context::<'a> {189		source: lex.remainder(),190		index: 0,191	};192193	if ctx.eat_if(|v| v == '-') != 0 {194		lex.mark_truncating();195	}196197	// Skip whitespaces198	ctx.eat_while(|r| r == ' ' || r == '\t' || r == '\r');199200	// Skip \n201	match ctx.next() {202		Some('\n') => (),203		None => {204			lex.eat_error(&ctx);205			return Err(UnexpectedEnd);206		}207		// Text block requires new line after |||.208		Some(_) => {209			lex.eat_error(&ctx);210			return Err(MissingNewLine);211		}212	}213214	// Process leading blank lines before calculating string block indent215	while ctx.peek() == Some('\n') {216		ctx.next();217	}218219	let mut num_whitespace = check_whitespace(ctx.rest(), ctx.rest());220	let str_block_indent = &ctx.rest()[..num_whitespace];221222	if num_whitespace == 0 {223		// Text block's first line must start with whitespace224		lex.eat_error(&ctx);225		return Err(MissingIndent);226	}227228	loop {229		debug_assert_ne!(num_whitespace, 0, "Unexpected value for num_whitespace");230		ctx.skip(num_whitespace);231232		let line_start = ctx.index;233		let mut line_size = 0;234		loop {235			match ctx.next() {236				None => {237					lex.eat_error(&ctx);238					return Err(UnexpectedEnd);239				}240				Some('\n') => {241					lex.mark_line(&ctx.source[line_start..line_start + line_size]);242					break;243				}244				Some(c) => {245					line_size += c.len_utf8();246				}247			}248		}249250		// Skip any blank lines251		while ctx.peek() == Some('\n') {252			lex.mark_line("");253			ctx.next();254		}255256		// Look at the next line257		num_whitespace = check_whitespace(str_block_indent, ctx.rest());258		if num_whitespace == 0 {259			// End of the text block260			// let mut term_indent = String::with_capacity(num_whitespace);261			while let Some(' ' | '\t') = ctx.peek() {262				// term_indent.push(263				ctx.next().unwrap();264				// );265			}266267			if !ctx.rest().starts_with("|||") {268				if ctx.rest().is_empty() {269					lex.bump_pos(ctx.index);270					return Err(UnexpectedEnd);271				}272				lex.eat_error(&ctx);273				return Err(MissingTermination);274			}275276			// Skip '|||'277			ctx.skip(3);278			break;279		}280	}281282	lex.bump_pos(ctx.index);283	Ok(())284}