diff options
| author | Stefan Lankes <slankes@eonerc.rwth-aachen.de> | 2019-11-13 00:24:37 +0100 |
|---|---|---|
| committer | Stefan Lankes <slankes@eonerc.rwth-aachen.de> | 2019-11-13 00:24:37 +0100 |
| commit | 88717319142e162ae2f48124d94f33d2c21bc2ce (patch) | |
| tree | e50a968f836c55007e9bc8f305d4f0f80aca042c /src/libsyntax/parse/lexer | |
| parent | 969b74144641bf1c8ae5aba0581f4b52a4c15bac (diff) | |
| parent | 4f03f4a989d1c8346c19dfb417a77c09b34408b8 (diff) | |
| download | rust-88717319142e162ae2f48124d94f33d2c21bc2ce.tar.gz rust-88717319142e162ae2f48124d94f33d2c21bc2ce.zip | |
Merge remote-tracking branch 'rust-lang/master' into hermit
Diffstat (limited to 'src/libsyntax/parse/lexer')
| -rw-r--r-- | src/libsyntax/parse/lexer/comments.rs | 254 | ||||
| -rw-r--r-- | src/libsyntax/parse/lexer/comments/tests.rs | 47 | ||||
| -rw-r--r-- | src/libsyntax/parse/lexer/mod.rs | 659 | ||||
| -rw-r--r-- | src/libsyntax/parse/lexer/tests.rs | 269 | ||||
| -rw-r--r-- | src/libsyntax/parse/lexer/tokentrees.rs | 279 | ||||
| -rw-r--r-- | src/libsyntax/parse/lexer/unicode_chars.rs | 392 |
6 files changed, 0 insertions, 1900 deletions
diff --git a/src/libsyntax/parse/lexer/comments.rs b/src/libsyntax/parse/lexer/comments.rs deleted file mode 100644 index ac79ce323bf..00000000000 --- a/src/libsyntax/parse/lexer/comments.rs +++ /dev/null @@ -1,254 +0,0 @@ -pub use CommentStyle::*; - -use crate::ast; -use crate::source_map::SourceMap; -use crate::parse::lexer::is_block_doc_comment; -use crate::parse::lexer::ParseSess; - -use syntax_pos::{BytePos, CharPos, Pos, FileName}; - -use std::usize; - -#[cfg(test)] -mod tests; - -#[derive(Clone, Copy, PartialEq, Debug)] -pub enum CommentStyle { - /// No code on either side of each line of the comment - Isolated, - /// Code exists to the left of the comment - Trailing, - /// Code before /* foo */ and after the comment - Mixed, - /// Just a manual blank line "\n\n", for layout - BlankLine, -} - -#[derive(Clone)] -pub struct Comment { - pub style: CommentStyle, - pub lines: Vec<String>, - pub pos: BytePos, -} - -fn is_doc_comment(s: &str) -> bool { - (s.starts_with("///") && super::is_doc_comment(s)) || s.starts_with("//!") || - (s.starts_with("/**") && is_block_doc_comment(s)) || s.starts_with("/*!") -} - -pub fn doc_comment_style(comment: &str) -> ast::AttrStyle { - assert!(is_doc_comment(comment)); - if comment.starts_with("//!") || comment.starts_with("/*!") { - ast::AttrStyle::Inner - } else { - ast::AttrStyle::Outer - } -} - -pub fn strip_doc_comment_decoration(comment: &str) -> String { - /// remove whitespace-only lines from the start/end of lines - fn vertical_trim(lines: Vec<String>) -> Vec<String> { - let mut i = 0; - let mut j = lines.len(); - // first line of all-stars should be omitted - if !lines.is_empty() && lines[0].chars().all(|c| c == '*') { - i += 1; - } - - while i < j && lines[i].trim().is_empty() { - i += 1; - } - // like the first, a last line of all stars should be omitted - if j > i && - lines[j - 1] - .chars() - .skip(1) - .all(|c| c == '*') { - j -= 1; - } - - while j > i && lines[j - 1].trim().is_empty() { - j -= 1; - } - - lines[i..j].to_vec() - } - - /// remove a "[ \t]*\*" block from each line, if possible - fn horizontal_trim(lines: Vec<String>) -> Vec<String> { - let mut i = usize::MAX; - let mut can_trim = true; - let mut first = true; - - for line in &lines { - for (j, c) in line.chars().enumerate() { - if j > i || !"* \t".contains(c) { - can_trim = false; - break; - } - if c == '*' { - if first { - i = j; - first = false; - } else if i != j { - can_trim = false; - } - break; - } - } - if i >= line.len() { - can_trim = false; - } - if !can_trim { - break; - } - } - - if can_trim { - lines.iter() - .map(|line| (&line[i + 1..line.len()]).to_string()) - .collect() - } else { - lines - } - } - - // one-line comments lose their prefix - const ONELINERS: &[&str] = &["///!", "///", "//!", "//"]; - - for prefix in ONELINERS { - if comment.starts_with(*prefix) { - return (&comment[prefix.len()..]).to_string(); - } - } - - if comment.starts_with("/*") { - let lines = comment[3..comment.len() - 2] - .lines() - .map(|s| s.to_string()) - .collect::<Vec<String>>(); - - let lines = vertical_trim(lines); - let lines = horizontal_trim(lines); - - return lines.join("\n"); - } - - panic!("not a doc-comment: {}", comment); -} - -/// Returns `None` if the first `col` chars of `s` contain a non-whitespace char. -/// Otherwise returns `Some(k)` where `k` is first char offset after that leading -/// whitespace. Note that `k` may be outside bounds of `s`. -fn all_whitespace(s: &str, col: CharPos) -> Option<usize> { - let mut idx = 0; - for (i, ch) in s.char_indices().take(col.to_usize()) { - if !ch.is_whitespace() { - return None; - } - idx = i + ch.len_utf8(); - } - Some(idx) -} - -fn trim_whitespace_prefix(s: &str, col: CharPos) -> &str { - let len = s.len(); - match all_whitespace(&s, col) { - Some(col) => if col < len { &s[col..] } else { "" }, - None => s, - } -} - -fn split_block_comment_into_lines( - text: &str, - col: CharPos, -) -> Vec<String> { - let mut res: Vec<String> = vec![]; - let mut lines = text.lines(); - // just push the first line - res.extend(lines.next().map(|it| it.to_string())); - // for other lines, strip common whitespace prefix - for line in lines { - res.push(trim_whitespace_prefix(line, col).to_string()) - } - res -} - -// it appears this function is called only from pprust... that's -// probably not a good thing. -crate fn gather_comments(sess: &ParseSess, path: FileName, src: String) -> Vec<Comment> { - let cm = SourceMap::new(sess.source_map().path_mapping().clone()); - let source_file = cm.new_source_file(path, src); - let text = (*source_file.src.as_ref().unwrap()).clone(); - - let text: &str = text.as_str(); - let start_bpos = source_file.start_pos; - let mut pos = 0; - let mut comments: Vec<Comment> = Vec::new(); - let mut code_to_the_left = false; - - if let Some(shebang_len) = rustc_lexer::strip_shebang(text) { - comments.push(Comment { - style: Isolated, - lines: vec![text[..shebang_len].to_string()], - pos: start_bpos, - }); - pos += shebang_len; - } - - for token in rustc_lexer::tokenize(&text[pos..]) { - let token_text = &text[pos..pos + token.len]; - match token.kind { - rustc_lexer::TokenKind::Whitespace => { - if let Some(mut idx) = token_text.find('\n') { - code_to_the_left = false; - while let Some(next_newline) = &token_text[idx + 1..].find('\n') { - idx = idx + 1 + next_newline; - comments.push(Comment { - style: BlankLine, - lines: vec![], - pos: start_bpos + BytePos((pos + idx) as u32), - }); - } - } - } - rustc_lexer::TokenKind::BlockComment { terminated: _ } => { - if !is_block_doc_comment(token_text) { - let code_to_the_right = match text[pos + token.len..].chars().next() { - Some('\r') | Some('\n') => false, - _ => true, - }; - let style = match (code_to_the_left, code_to_the_right) { - (true, true) | (false, true) => Mixed, - (false, false) => Isolated, - (true, false) => Trailing, - }; - - // Count the number of chars since the start of the line by rescanning. - let pos_in_file = start_bpos + BytePos(pos as u32); - let line_begin_in_file = source_file.line_begin_pos(pos_in_file); - let line_begin_pos = (line_begin_in_file - start_bpos).to_usize(); - let col = CharPos(text[line_begin_pos..pos].chars().count()); - - let lines = split_block_comment_into_lines(token_text, col); - comments.push(Comment { style, lines, pos: pos_in_file }) - } - } - rustc_lexer::TokenKind::LineComment => { - if !is_doc_comment(token_text) { - comments.push(Comment { - style: if code_to_the_left { Trailing } else { Isolated }, - lines: vec![token_text.to_string()], - pos: start_bpos + BytePos(pos as u32), - }) - } - } - _ => { - code_to_the_left = true; - } - } - pos += token.len; - } - - comments -} diff --git a/src/libsyntax/parse/lexer/comments/tests.rs b/src/libsyntax/parse/lexer/comments/tests.rs deleted file mode 100644 index f9cd69fb50d..00000000000 --- a/src/libsyntax/parse/lexer/comments/tests.rs +++ /dev/null @@ -1,47 +0,0 @@ -use super::*; - -#[test] -fn test_block_doc_comment_1() { - let comment = "/**\n * Test \n ** Test\n * Test\n*/"; - let stripped = strip_doc_comment_decoration(comment); - assert_eq!(stripped, " Test \n* Test\n Test"); -} - -#[test] -fn test_block_doc_comment_2() { - let comment = "/**\n * Test\n * Test\n*/"; - let stripped = strip_doc_comment_decoration(comment); - assert_eq!(stripped, " Test\n Test"); -} - -#[test] -fn test_block_doc_comment_3() { - let comment = "/**\n let a: *i32;\n *a = 5;\n*/"; - let stripped = strip_doc_comment_decoration(comment); - assert_eq!(stripped, " let a: *i32;\n *a = 5;"); -} - -#[test] -fn test_block_doc_comment_4() { - let comment = "/*******************\n test\n *********************/"; - let stripped = strip_doc_comment_decoration(comment); - assert_eq!(stripped, " test"); -} - -#[test] -fn test_line_doc_comment() { - let stripped = strip_doc_comment_decoration("/// test"); - assert_eq!(stripped, " test"); - let stripped = strip_doc_comment_decoration("///! test"); - assert_eq!(stripped, " test"); - let stripped = strip_doc_comment_decoration("// test"); - assert_eq!(stripped, " test"); - let stripped = strip_doc_comment_decoration("// test"); - assert_eq!(stripped, " test"); - let stripped = strip_doc_comment_decoration("///test"); - assert_eq!(stripped, "test"); - let stripped = strip_doc_comment_decoration("///!test"); - assert_eq!(stripped, "test"); - let stripped = strip_doc_comment_decoration("//test"); - assert_eq!(stripped, "test"); -} diff --git a/src/libsyntax/parse/lexer/mod.rs b/src/libsyntax/parse/lexer/mod.rs deleted file mode 100644 index 48cf117f65a..00000000000 --- a/src/libsyntax/parse/lexer/mod.rs +++ /dev/null @@ -1,659 +0,0 @@ -use crate::parse::token::{self, Token, TokenKind}; -use crate::sess::ParseSess; -use crate::symbol::{sym, Symbol}; -use crate::parse::unescape_error_reporting::{emit_unescape_error, push_escaped_char}; - -use errors::{FatalError, DiagnosticBuilder}; -use syntax_pos::{BytePos, Pos, Span}; -use rustc_lexer::Base; -use rustc_lexer::unescape; - -use std::char; -use std::convert::TryInto; -use rustc_data_structures::sync::Lrc; -use log::debug; - -#[cfg(test)] -mod tests; - -pub mod comments; -mod tokentrees; -mod unicode_chars; - -#[derive(Clone, Debug)] -pub struct UnmatchedBrace { - pub expected_delim: token::DelimToken, - pub found_delim: Option<token::DelimToken>, - pub found_span: Span, - pub unclosed_span: Option<Span>, - pub candidate_span: Option<Span>, -} - -pub struct StringReader<'a> { - sess: &'a ParseSess, - /// Initial position, read-only. - start_pos: BytePos, - /// The absolute offset within the source_map of the current character. - pos: BytePos, - /// Stop reading src at this index. - end_src_index: usize, - /// Source text to tokenize. - src: Lrc<String>, - override_span: Option<Span>, -} - -impl<'a> StringReader<'a> { - pub fn new(sess: &'a ParseSess, - source_file: Lrc<syntax_pos::SourceFile>, - override_span: Option<Span>) -> Self { - if source_file.src.is_none() { - sess.span_diagnostic.bug(&format!("cannot lex `source_file` without source: {}", - source_file.name)); - } - - let src = (*source_file.src.as_ref().unwrap()).clone(); - - StringReader { - sess, - start_pos: source_file.start_pos, - pos: source_file.start_pos, - end_src_index: src.len(), - src, - override_span, - } - } - - pub fn retokenize(sess: &'a ParseSess, mut span: Span) -> Self { - let begin = sess.source_map().lookup_byte_offset(span.lo()); - let end = sess.source_map().lookup_byte_offset(span.hi()); - - // Make the range zero-length if the span is invalid. - if span.lo() > span.hi() || begin.sf.start_pos != end.sf.start_pos { - span = span.shrink_to_lo(); - } - - let mut sr = StringReader::new(sess, begin.sf, None); - - // Seek the lexer to the right byte range. - sr.end_src_index = sr.src_index(span.hi()); - - sr - } - - - fn mk_sp(&self, lo: BytePos, hi: BytePos) -> Span { - self.override_span.unwrap_or_else(|| Span::with_root_ctxt(lo, hi)) - } - - /// Returns the next token, including trivia like whitespace or comments. - /// - /// `Err(())` means that some errors were encountered, which can be - /// retrieved using `buffer_fatal_errors`. - pub fn next_token(&mut self) -> Token { - let start_src_index = self.src_index(self.pos); - let text: &str = &self.src[start_src_index..self.end_src_index]; - - if text.is_empty() { - let span = self.mk_sp(self.pos, self.pos); - return Token::new(token::Eof, span); - } - - { - let is_beginning_of_file = self.pos == self.start_pos; - if is_beginning_of_file { - if let Some(shebang_len) = rustc_lexer::strip_shebang(text) { - let start = self.pos; - self.pos = self.pos + BytePos::from_usize(shebang_len); - - let sym = self.symbol_from(start + BytePos::from_usize("#!".len())); - let kind = token::Shebang(sym); - - let span = self.mk_sp(start, self.pos); - return Token::new(kind, span); - } - } - } - - let token = rustc_lexer::first_token(text); - - let start = self.pos; - self.pos = self.pos + BytePos::from_usize(token.len); - - debug!("try_next_token: {:?}({:?})", token.kind, self.str_from(start)); - - // This could use `?`, but that makes code significantly (10-20%) slower. - // https://github.com/rust-lang/rust/issues/37939 - let kind = self.cook_lexer_token(token.kind, start); - - let span = self.mk_sp(start, self.pos); - Token::new(kind, span) - } - - /// Report a fatal lexical error with a given span. - fn fatal_span(&self, sp: Span, m: &str) -> FatalError { - self.sess.span_diagnostic.span_fatal(sp, m) - } - - /// Report a lexical error with a given span. - fn err_span(&self, sp: Span, m: &str) { - self.sess.span_diagnostic.struct_span_err(sp, m).emit(); - } - - - /// Report a fatal error spanning [`from_pos`, `to_pos`). - fn fatal_span_(&self, from_pos: BytePos, to_pos: BytePos, m: &str) -> FatalError { - self.fatal_span(self.mk_sp(from_pos, to_pos), m) - } - - /// Report a lexical error spanning [`from_pos`, `to_pos`). - fn err_span_(&self, from_pos: BytePos, to_pos: BytePos, m: &str) { - self.err_span(self.mk_sp(from_pos, to_pos), m) - } - - fn struct_span_fatal(&self, from_pos: BytePos, to_pos: BytePos, m: &str) - -> DiagnosticBuilder<'a> - { - self.sess.span_diagnostic.struct_span_fatal(self.mk_sp(from_pos, to_pos), m) - } - - fn struct_fatal_span_char(&self, from_pos: BytePos, to_pos: BytePos, m: &str, c: char) - -> DiagnosticBuilder<'a> - { - let mut m = m.to_string(); - m.push_str(": "); - push_escaped_char(&mut m, c); - - self.sess.span_diagnostic.struct_span_fatal(self.mk_sp(from_pos, to_pos), &m[..]) - } - - /// Turns simple `rustc_lexer::TokenKind` enum into a rich - /// `libsyntax::TokenKind`. This turns strings into interned - /// symbols and runs additional validation. - fn cook_lexer_token( - &self, - token: rustc_lexer::TokenKind, - start: BytePos, - ) -> TokenKind { - match token { - rustc_lexer::TokenKind::LineComment => { - let string = self.str_from(start); - // comments with only more "/"s are not doc comments - let tok = if is_doc_comment(string) { - self.forbid_bare_cr(start, string, "bare CR not allowed in doc-comment"); - token::DocComment(Symbol::intern(string)) - } else { - token::Comment - }; - - tok - } - rustc_lexer::TokenKind::BlockComment { terminated } => { - let string = self.str_from(start); - // block comments starting with "/**" or "/*!" are doc-comments - // but comments with only "*"s between two "/"s are not - let is_doc_comment = is_block_doc_comment(string); - - if !terminated { - let msg = if is_doc_comment { - "unterminated block doc-comment" - } else { - "unterminated block comment" - }; - let last_bpos = self.pos; - self.fatal_span_(start, last_bpos, msg).raise(); - } - - let tok = if is_doc_comment { - self.forbid_bare_cr(start, - string, - "bare CR not allowed in block doc-comment"); - token::DocComment(Symbol::intern(string)) - } else { - token::Comment - }; - - tok - } - rustc_lexer::TokenKind::Whitespace => token::Whitespace, - rustc_lexer::TokenKind::Ident | rustc_lexer::TokenKind::RawIdent => { - let is_raw_ident = token == rustc_lexer::TokenKind::RawIdent; - let mut ident_start = start; - if is_raw_ident { - ident_start = ident_start + BytePos(2); - } - // FIXME: perform NFKC normalization here. (Issue #2253) - let sym = self.symbol_from(ident_start); - if is_raw_ident { - let span = self.mk_sp(start, self.pos); - if !sym.can_be_raw() { - self.err_span(span, &format!("`{}` cannot be a raw identifier", sym)); - } - self.sess.raw_identifier_spans.borrow_mut().push(span); - } - token::Ident(sym, is_raw_ident) - } - rustc_lexer::TokenKind::Literal { kind, suffix_start } => { - let suffix_start = start + BytePos(suffix_start as u32); - let (kind, symbol) = self.cook_lexer_literal(start, suffix_start, kind); - let suffix = if suffix_start < self.pos { - let string = self.str_from(suffix_start); - if string == "_" { - self.sess.span_diagnostic - .struct_span_warn(self.mk_sp(suffix_start, self.pos), - "underscore literal suffix is not allowed") - .warn("this was previously accepted by the compiler but is \ - being phased out; it will become a hard error in \ - a future release!") - .note("for more information, see issue #42326 \ - <https://github.com/rust-lang/rust/issues/42326>") - .emit(); - None - } else { - Some(Symbol::intern(string)) - } - } else { - None - }; - token::Literal(token::Lit { kind, symbol, suffix }) - } - rustc_lexer::TokenKind::Lifetime { starts_with_number } => { - // Include the leading `'` in the real identifier, for macro - // expansion purposes. See #12512 for the gory details of why - // this is necessary. - let lifetime_name = self.str_from(start); - if starts_with_number { - self.err_span_( - start, - self.pos, - "lifetimes cannot start with a number", - ); - } - let ident = Symbol::intern(lifetime_name); - token::Lifetime(ident) - } - rustc_lexer::TokenKind::Semi => token::Semi, - rustc_lexer::TokenKind::Comma => token::Comma, - rustc_lexer::TokenKind::Dot => token::Dot, - rustc_lexer::TokenKind::OpenParen => token::OpenDelim(token::Paren), - rustc_lexer::TokenKind::CloseParen => token::CloseDelim(token::Paren), - rustc_lexer::TokenKind::OpenBrace => token::OpenDelim(token::Brace), - rustc_lexer::TokenKind::CloseBrace => token::CloseDelim(token::Brace), - rustc_lexer::TokenKind::OpenBracket => token::OpenDelim(token::Bracket), - rustc_lexer::TokenKind::CloseBracket => token::CloseDelim(token::Bracket), - rustc_lexer::TokenKind::At => token::At, - rustc_lexer::TokenKind::Pound => token::Pound, - rustc_lexer::TokenKind::Tilde => token::Tilde, - rustc_lexer::TokenKind::Question => token::Question, - rustc_lexer::TokenKind::Colon => token::Colon, - rustc_lexer::TokenKind::Dollar => token::Dollar, - rustc_lexer::TokenKind::Eq => token::Eq, - rustc_lexer::TokenKind::Not => token::Not, - rustc_lexer::TokenKind::Lt => token::Lt, - rustc_lexer::TokenKind::Gt => token::Gt, - rustc_lexer::TokenKind::Minus => token::BinOp(token::Minus), - rustc_lexer::TokenKind::And => token::BinOp(token::And), - rustc_lexer::TokenKind::Or => token::BinOp(token::Or), - rustc_lexer::TokenKind::Plus => token::BinOp(token::Plus), - rustc_lexer::TokenKind::Star => token::BinOp(token::Star), - rustc_lexer::TokenKind::Slash => token::BinOp(token::Slash), - rustc_lexer::TokenKind::Caret => token::BinOp(token::Caret), - rustc_lexer::TokenKind::Percent => token::BinOp(token::Percent), - - rustc_lexer::TokenKind::Unknown => { - let c = self.str_from(start).chars().next().unwrap(); - let mut err = self.struct_fatal_span_char(start, - self.pos, - "unknown start of token", - c); - // FIXME: the lexer could be used to turn the ASCII version of unicode homoglyphs, - // instead of keeping a table in `check_for_substitution`into the token. Ideally, - // this should be inside `rustc_lexer`. However, we should first remove compound - // tokens like `<<` from `rustc_lexer`, and then add fancier error recovery to it, - // as there will be less overall work to do this way. - let token = unicode_chars::check_for_substitution(self, start, c, &mut err) - .unwrap_or_else(|| token::Unknown(self.symbol_from(start))); - err.emit(); - token - } - } - } - - fn cook_lexer_literal( - &self, - start: BytePos, - suffix_start: BytePos, - kind: rustc_lexer::LiteralKind - ) -> (token::LitKind, Symbol) { - match kind { - rustc_lexer::LiteralKind::Char { terminated } => { - if !terminated { - self.fatal_span_(start, suffix_start, - "unterminated character literal".into()) - .raise() - } - let content_start = start + BytePos(1); - let content_end = suffix_start - BytePos(1); - self.validate_char_escape(content_start, content_end); - let id = self.symbol_from_to(content_start, content_end); - (token::Char, id) - }, - rustc_lexer::LiteralKind::Byte { terminated } => { - if !terminated { - self.fatal_span_(start + BytePos(1), suffix_start, - "unterminated byte constant".into()) - .raise() - } - let content_start = start + BytePos(2); - let content_end = suffix_start - BytePos(1); - self.validate_byte_escape(content_start, content_end); - let id = self.symbol_from_to(content_start, content_end); - (token::Byte, id) - }, - rustc_lexer::LiteralKind::Str { terminated } => { - if !terminated { - self.fatal_span_(start, suffix_start, - "unterminated double quote string".into()) - .raise() - } - let content_start = start + BytePos(1); - let content_end = suffix_start - BytePos(1); - self.validate_str_escape(content_start, content_end); - let id = self.symbol_from_to(content_start, content_end); - (token::Str, id) - } - rustc_lexer::LiteralKind::ByteStr { terminated } => { - if !terminated { - self.fatal_span_(start + BytePos(1), suffix_start, - "unterminated double quote byte string".into()) - .raise() - } - let content_start = start + BytePos(2); - let content_end = suffix_start - BytePos(1); - self.validate_byte_str_escape(content_start, content_end); - let id = self.symbol_from_to(content_start, content_end); - (token::ByteStr, id) - } - rustc_lexer::LiteralKind::RawStr { n_hashes, started, terminated } => { - if !started { - self.report_non_started_raw_string(start); - } - if !terminated { - self.report_unterminated_raw_string(start, n_hashes) - } - let n_hashes: u16 = self.restrict_n_hashes(start, n_hashes); - let n = u32::from(n_hashes); - let content_start = start + BytePos(2 + n); - let content_end = suffix_start - BytePos(1 + n); - self.validate_raw_str_escape(content_start, content_end); - let id = self.symbol_from_to(content_start, content_end); - (token::StrRaw(n_hashes), id) - } - rustc_lexer::LiteralKind::RawByteStr { n_hashes, started, terminated } => { - if !started { - self.report_non_started_raw_string(start); - } - if !terminated { - self.report_unterminated_raw_string(start, n_hashes) - } - let n_hashes: u16 = self.restrict_n_hashes(start, n_hashes); - let n = u32::from(n_hashes); - let content_start = start + BytePos(3 + n); - let content_end = suffix_start - BytePos(1 + n); - self.validate_raw_byte_str_escape(content_start, content_end); - let id = self.symbol_from_to(content_start, content_end); - (token::ByteStrRaw(n_hashes), id) - } - rustc_lexer::LiteralKind::Int { base, empty_int } => { - if empty_int { - self.err_span_(start, suffix_start, "no valid digits found for number"); - (token::Integer, sym::integer(0)) - } else { - self.validate_int_literal(base, start, suffix_start); - (token::Integer, self.symbol_from_to(start, suffix_start)) - } - }, - rustc_lexer::LiteralKind::Float { base, empty_exponent } => { - if empty_exponent { - let mut err = self.struct_span_fatal( - start, self.pos, - "expected at least one digit in exponent" - ); - err.emit(); - } - - match base { - Base::Hexadecimal => { - self.err_span_(start, suffix_start, - "hexadecimal float literal is not supported") - } - Base::Octal => { - self.err_span_(start, suffix_start, - "octal float literal is not supported") - } - Base::Binary => { - self.err_span_(start, suffix_start, - "binary float literal is not supported") - } - _ => () - } - - let id = self.symbol_from_to(start, suffix_start); - (token::Float, id) - }, - } - } - - #[inline] - fn src_index(&self, pos: BytePos) -> usize { - (pos - self.start_pos).to_usize() - } - - /// Slice of the source text from `start` up to but excluding `self.pos`, - /// meaning the slice does not include the character `self.ch`. - fn str_from(&self, start: BytePos) -> &str - { - self.str_from_to(start, self.pos) - } - - /// Creates a Symbol from a given offset to the current offset. - fn symbol_from(&self, start: BytePos) -> Symbol { - debug!("taking an ident from {:?} to {:?}", start, self.pos); - Symbol::intern(self.str_from(start)) - } - - /// As symbol_from, with an explicit endpoint. - fn symbol_from_to(&self, start: BytePos, end: BytePos) -> Symbol { - debug!("taking an ident from {:?} to {:?}", start, end); - Symbol::intern(self.str_from_to(start, end)) - } - - /// Slice of the source text spanning from `start` up to but excluding `end`. - fn str_from_to(&self, start: BytePos, end: BytePos) -> &str - { - &self.src[self.src_index(start)..self.src_index(end)] - } - - fn forbid_bare_cr(&self, start: BytePos, s: &str, errmsg: &str) { - let mut idx = 0; - loop { - idx = match s[idx..].find('\r') { - None => break, - Some(it) => idx + it + 1 - }; - self.err_span_(start + BytePos(idx as u32 - 1), - start + BytePos(idx as u32), - errmsg); - } - } - - fn report_non_started_raw_string(&self, start: BytePos) -> ! { - let bad_char = self.str_from(start).chars().last().unwrap(); - self - .struct_fatal_span_char( - start, - self.pos, - "found invalid character; only `#` is allowed \ - in raw string delimitation", - bad_char, - ) - .emit(); - FatalError.raise() - } - - fn report_unterminated_raw_string(&self, start: BytePos, n_hashes: usize) -> ! { - let mut err = self.struct_span_fatal( - start, start, - "unterminated raw string", - ); - err.span_label( - self.mk_sp(start, start), - "unterminated raw string", - ); - - if n_hashes > 0 { - err.note(&format!("this raw string should be terminated with `\"{}`", - "#".repeat(n_hashes as usize))); - } - - err.emit(); - FatalError.raise() - } - - fn restrict_n_hashes(&self, start: BytePos, n_hashes: usize) -> u16 { - match n_hashes.try_into() { - Ok(n_hashes) => n_hashes, - Err(_) => { - self.fatal_span_(start, - self.pos, - "too many `#` symbols: raw strings may be \ - delimited by up to 65535 `#` symbols").raise(); - } - } - } - - fn validate_char_escape(&self, content_start: BytePos, content_end: BytePos) { - let lit = self.str_from_to(content_start, content_end); - if let Err((off, err)) = unescape::unescape_char(lit) { - emit_unescape_error( - &self.sess.span_diagnostic, - lit, - self.mk_sp(content_start - BytePos(1), content_end + BytePos(1)), - unescape::Mode::Char, - 0..off, - err, - ) - } - } - - fn validate_byte_escape(&self, content_start: BytePos, content_end: BytePos) { - let lit = self.str_from_to(content_start, content_end); - if let Err((off, err)) = unescape::unescape_byte(lit) { - emit_unescape_error( - &self.sess.span_diagnostic, - lit, - self.mk_sp(content_start - BytePos(1), content_end + BytePos(1)), - unescape::Mode::Byte, - 0..off, - err, - ) - } - } - - fn validate_str_escape(&self, content_start: BytePos, content_end: BytePos) { - let lit = self.str_from_to(content_start, content_end); - unescape::unescape_str(lit, &mut |range, c| { - if let Err(err) = c { - emit_unescape_error( - &self.sess.span_diagnostic, - lit, - self.mk_sp(content_start - BytePos(1), content_end + BytePos(1)), - unescape::Mode::Str, - range, - err, - ) - } - }) - } - - fn validate_raw_str_escape(&self, content_start: BytePos, content_end: BytePos) { - let lit = self.str_from_to(content_start, content_end); - unescape::unescape_raw_str(lit, &mut |range, c| { - if let Err(err) = c { - emit_unescape_error( - &self.sess.span_diagnostic, - lit, - self.mk_sp(content_start - BytePos(1), content_end + BytePos(1)), - unescape::Mode::Str, - range, - err, - ) - } - }) - } - - fn validate_raw_byte_str_escape(&self, content_start: BytePos, content_end: BytePos) { - let lit = self.str_from_to(content_start, content_end); - unescape::unescape_raw_byte_str(lit, &mut |range, c| { - if let Err(err) = c { - emit_unescape_error( - &self.sess.span_diagnostic, - lit, - self.mk_sp(content_start - BytePos(1), content_end + BytePos(1)), - unescape::Mode::ByteStr, - range, - err, - ) - } - }) - } - - fn validate_byte_str_escape(&self, content_start: BytePos, content_end: BytePos) { - let lit = self.str_from_to(content_start, content_end); - unescape::unescape_byte_str(lit, &mut |range, c| { - if let Err(err) = c { - emit_unescape_error( - &self.sess.span_diagnostic, - lit, - self.mk_sp(content_start - BytePos(1), content_end + BytePos(1)), - unescape::Mode::ByteStr, - range, - err, - ) - } - }) - } - - fn validate_int_literal(&self, base: Base, content_start: BytePos, content_end: BytePos) { - let base = match base { - Base::Binary => 2, - Base::Octal => 8, - _ => return, - }; - let s = self.str_from_to(content_start + BytePos(2), content_end); - for (idx, c) in s.char_indices() { - let idx = idx as u32; - if c != '_' && c.to_digit(base).is_none() { - let lo = content_start + BytePos(2 + idx); - let hi = content_start + BytePos(2 + idx + c.len_utf8() as u32); - self.err_span_(lo, hi, - &format!("invalid digit for a base {} literal", base)); - - } - } - } -} - -fn is_doc_comment(s: &str) -> bool { - let res = (s.starts_with("///") && *s.as_bytes().get(3).unwrap_or(&b' ') != b'/') || - s.starts_with("//!"); - debug!("is {:?} a doc comment? {}", s, res); - res -} - -fn is_block_doc_comment(s: &str) -> bool { - // Prevent `/**/` from being parsed as a doc comment - let res = ((s.starts_with("/**") && *s.as_bytes().get(3).unwrap_or(&b' ') != b'*') || - s.starts_with("/*!")) && s.len() >= 5; - debug!("is {:?} a doc comment? {}", s, res); - res -} diff --git a/src/libsyntax/parse/lexer/tests.rs b/src/libsyntax/parse/lexer/tests.rs deleted file mode 100644 index de301b1fc49..00000000000 --- a/src/libsyntax/parse/lexer/tests.rs +++ /dev/null @@ -1,269 +0,0 @@ -use super::*; - -use crate::symbol::Symbol; -use crate::source_map::{SourceMap, FilePathMapping}; -use crate::parse::token; -use crate::with_default_globals; - -use errors::{Handler, emitter::EmitterWriter}; -use std::io; -use std::path::PathBuf; -use syntax_pos::{BytePos, Span}; - -fn mk_sess(sm: Lrc<SourceMap>) -> ParseSess { - let emitter = EmitterWriter::new( - Box::new(io::sink()), - Some(sm.clone()), - false, - false, - false, - None, - false, - ); - ParseSess::with_span_handler(Handler::with_emitter(true, None, Box::new(emitter)), sm) -} - -// Creates a string reader for the given string. -fn setup<'a>(sm: &SourceMap, - sess: &'a ParseSess, - teststr: String) - -> StringReader<'a> { - let sf = sm.new_source_file(PathBuf::from(teststr.clone()).into(), teststr); - StringReader::new(sess, sf, None) -} - -#[test] -fn t1() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - let mut string_reader = setup( - &sm, - &sh, - "/* my source file */ fn main() { println!(\"zebra\"); }\n".to_string(), - ); - assert_eq!(string_reader.next_token(), token::Comment); - assert_eq!(string_reader.next_token(), token::Whitespace); - let tok1 = string_reader.next_token(); - let tok2 = Token::new( - mk_ident("fn"), - Span::with_root_ctxt(BytePos(21), BytePos(23)), - ); - assert_eq!(tok1.kind, tok2.kind); - assert_eq!(tok1.span, tok2.span); - assert_eq!(string_reader.next_token(), token::Whitespace); - // Read another token. - let tok3 = string_reader.next_token(); - assert_eq!(string_reader.pos.clone(), BytePos(28)); - let tok4 = Token::new( - mk_ident("main"), - Span::with_root_ctxt(BytePos(24), BytePos(28)), - ); - assert_eq!(tok3.kind, tok4.kind); - assert_eq!(tok3.span, tok4.span); - - assert_eq!(string_reader.next_token(), token::OpenDelim(token::Paren)); - assert_eq!(string_reader.pos.clone(), BytePos(29)) - }) -} - -// Checks that the given reader produces the desired stream -// of tokens (stop checking after exhausting `expected`). -fn check_tokenization(mut string_reader: StringReader<'_>, expected: Vec<TokenKind>) { - for expected_tok in &expected { - assert_eq!(&string_reader.next_token(), expected_tok); - } -} - -// Makes the identifier by looking up the string in the interner. -fn mk_ident(id: &str) -> TokenKind { - token::Ident(Symbol::intern(id), false) -} - -fn mk_lit(kind: token::LitKind, symbol: &str, suffix: Option<&str>) -> TokenKind { - TokenKind::lit(kind, Symbol::intern(symbol), suffix.map(Symbol::intern)) -} - -#[test] -fn doublecolon_parsing() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - check_tokenization( - setup(&sm, &sh, "a b".to_string()), - vec![mk_ident("a"), token::Whitespace, mk_ident("b")], - ); - }) -} - -#[test] -fn doublecolon_parsing_2() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - check_tokenization( - setup(&sm, &sh, "a::b".to_string()), - vec![mk_ident("a"), token::Colon, token::Colon, mk_ident("b")], - ); - }) -} - -#[test] -fn doublecolon_parsing_3() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - check_tokenization( - setup(&sm, &sh, "a ::b".to_string()), - vec![mk_ident("a"), token::Whitespace, token::Colon, token::Colon, mk_ident("b")], - ); - }) -} - -#[test] -fn doublecolon_parsing_4() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - check_tokenization( - setup(&sm, &sh, "a:: b".to_string()), - vec![mk_ident("a"), token::Colon, token::Colon, token::Whitespace, mk_ident("b")], - ); - }) -} - -#[test] -fn character_a() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - assert_eq!( - setup(&sm, &sh, "'a'".to_string()).next_token(), - mk_lit(token::Char, "a", None), - ); - }) -} - -#[test] -fn character_space() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - assert_eq!( - setup(&sm, &sh, "' '".to_string()).next_token(), - mk_lit(token::Char, " ", None), - ); - }) -} - -#[test] -fn character_escaped() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - assert_eq!( - setup(&sm, &sh, "'\\n'".to_string()).next_token(), - mk_lit(token::Char, "\\n", None), - ); - }) -} - -#[test] -fn lifetime_name() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - assert_eq!( - setup(&sm, &sh, "'abc".to_string()).next_token(), - token::Lifetime(Symbol::intern("'abc")), - ); - }) -} - -#[test] -fn raw_string() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - assert_eq!( - setup(&sm, &sh, "r###\"\"#a\\b\x00c\"\"###".to_string()).next_token(), - mk_lit(token::StrRaw(3), "\"#a\\b\x00c\"", None), - ); - }) -} - -#[test] -fn literal_suffixes() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - macro_rules! test { - ($input: expr, $tok_type: ident, $tok_contents: expr) => {{ - assert_eq!( - setup(&sm, &sh, format!("{}suffix", $input)).next_token(), - mk_lit(token::$tok_type, $tok_contents, Some("suffix")), - ); - // with a whitespace separator - assert_eq!( - setup(&sm, &sh, format!("{} suffix", $input)).next_token(), - mk_lit(token::$tok_type, $tok_contents, None), - ); - }} - } - - test!("'a'", Char, "a"); - test!("b'a'", Byte, "a"); - test!("\"a\"", Str, "a"); - test!("b\"a\"", ByteStr, "a"); - test!("1234", Integer, "1234"); - test!("0b101", Integer, "0b101"); - test!("0xABC", Integer, "0xABC"); - test!("1.0", Float, "1.0"); - test!("1.0e10", Float, "1.0e10"); - - assert_eq!( - setup(&sm, &sh, "2us".to_string()).next_token(), - mk_lit(token::Integer, "2", Some("us")), - ); - assert_eq!( - setup(&sm, &sh, "r###\"raw\"###suffix".to_string()).next_token(), - mk_lit(token::StrRaw(3), "raw", Some("suffix")), - ); - assert_eq!( - setup(&sm, &sh, "br###\"raw\"###suffix".to_string()).next_token(), - mk_lit(token::ByteStrRaw(3), "raw", Some("suffix")), - ); - }) -} - -#[test] -fn line_doc_comments() { - assert!(is_doc_comment("///")); - assert!(is_doc_comment("/// blah")); - assert!(!is_doc_comment("////")); -} - -#[test] -fn nested_block_comments() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - let mut lexer = setup(&sm, &sh, "/* /* */ */'a'".to_string()); - assert_eq!(lexer.next_token(), token::Comment); - assert_eq!(lexer.next_token(), mk_lit(token::Char, "a", None)); - }) -} - -#[test] -fn crlf_comments() { - with_default_globals(|| { - let sm = Lrc::new(SourceMap::new(FilePathMapping::empty())); - let sh = mk_sess(sm.clone()); - let mut lexer = setup(&sm, &sh, "// test\r\n/// test\r\n".to_string()); - let comment = lexer.next_token(); - assert_eq!(comment.kind, token::Comment); - assert_eq!((comment.span.lo(), comment.span.hi()), (BytePos(0), BytePos(7))); - assert_eq!(lexer.next_token(), token::Whitespace); - assert_eq!(lexer.next_token(), token::DocComment(Symbol::intern("/// test"))); - }) -} diff --git a/src/libsyntax/parse/lexer/tokentrees.rs b/src/libsyntax/parse/lexer/tokentrees.rs deleted file mode 100644 index de8ac2c71e8..00000000000 --- a/src/libsyntax/parse/lexer/tokentrees.rs +++ /dev/null @@ -1,279 +0,0 @@ -use rustc_data_structures::fx::FxHashMap; -use syntax_pos::Span; - -use super::{StringReader, UnmatchedBrace}; - -use crate::print::pprust::token_to_string; -use crate::parse::token::{self, Token}; -use crate::parse::PResult; -use crate::tokenstream::{DelimSpan, IsJoint::{self, *}, TokenStream, TokenTree, TreeAndJoint}; - -impl<'a> StringReader<'a> { - crate fn into_token_trees(self) -> (PResult<'a, TokenStream>, Vec<UnmatchedBrace>) { - let mut tt_reader = TokenTreesReader { - string_reader: self, - token: Token::dummy(), - joint_to_prev: Joint, - open_braces: Vec::new(), - unmatched_braces: Vec::new(), - matching_delim_spans: Vec::new(), - last_unclosed_found_span: None, - last_delim_empty_block_spans: FxHashMap::default() - }; - let res = tt_reader.parse_all_token_trees(); - (res, tt_reader.unmatched_braces) - } -} - -struct TokenTreesReader<'a> { - string_reader: StringReader<'a>, - token: Token, - joint_to_prev: IsJoint, - /// Stack of open delimiters and their spans. Used for error message. - open_braces: Vec<(token::DelimToken, Span)>, - unmatched_braces: Vec<UnmatchedBrace>, - /// The type and spans for all braces - /// - /// Used only for error recovery when arriving to EOF with mismatched braces. - matching_delim_spans: Vec<(token::DelimToken, Span, Span)>, - last_unclosed_found_span: Option<Span>, - last_delim_empty_block_spans: FxHashMap<token::DelimToken, Span> -} - -impl<'a> TokenTreesReader<'a> { - // Parse a stream of tokens into a list of `TokenTree`s, up to an `Eof`. - fn parse_all_token_trees(&mut self) -> PResult<'a, TokenStream> { - let mut buf = TokenStreamBuilder::default(); - - self.real_token(); - while self.token != token::Eof { - buf.push(self.parse_token_tree()?); - } - - Ok(buf.into_token_stream()) - } - - // Parse a stream of tokens into a list of `TokenTree`s, up to a `CloseDelim`. - fn parse_token_trees_until_close_delim(&mut self) -> TokenStream { - let mut buf = TokenStreamBuilder::default(); - loop { - if let token::CloseDelim(..) = self.token.kind { - return buf.into_token_stream(); - } - - match self.parse_token_tree() { - Ok(tree) => buf.push(tree), - Err(mut e) => { - e.emit(); - return buf.into_token_stream(); - } - } - } - } - - fn parse_token_tree(&mut self) -> PResult<'a, TreeAndJoint> { - let sm = self.string_reader.sess.source_map(); - match self.token.kind { - token::Eof => { - let msg = "this file contains an un-closed delimiter"; - let mut err = self.string_reader.sess.span_diagnostic - .struct_span_err(self.token.span, msg); - for &(_, sp) in &self.open_braces { - err.span_label(sp, "un-closed delimiter"); - self.unmatched_braces.push(UnmatchedBrace { - expected_delim: token::DelimToken::Brace, - found_delim: None, - found_span: self.token.span, - unclosed_span: Some(sp), - candidate_span: None, - }); - } - - if let Some((delim, _)) = self.open_braces.last() { - if let Some((_, open_sp, close_sp)) = self.matching_delim_spans.iter() - .filter(|(d, open_sp, close_sp)| { - if let Some(close_padding) = sm.span_to_margin(*close_sp) { - if let Some(open_padding) = sm.span_to_margin(*open_sp) { - return delim == d && close_padding != open_padding; - } - } - false - }).next() // these are in reverse order as they get inserted on close, but - { // we want the last open/first close - err.span_label( - *open_sp, - "this delimiter might not be properly closed...", - ); - err.span_label( - *close_sp, - "...as it matches this but it has different indentation", - ); - } - } - Err(err) - }, - token::OpenDelim(delim) => { - // The span for beginning of the delimited section - let pre_span = self.token.span; - - // Parse the open delimiter. - self.open_braces.push((delim, self.token.span)); - self.real_token(); - - // Parse the token trees within the delimiters. - // We stop at any delimiter so we can try to recover if the user - // uses an incorrect delimiter. - let tts = self.parse_token_trees_until_close_delim(); - - // Expand to cover the entire delimited token tree - let delim_span = DelimSpan::from_pair(pre_span, self.token.span); - - match self.token.kind { - // Correct delimiter. - token::CloseDelim(d) if d == delim => { - let (open_brace, open_brace_span) = self.open_braces.pop().unwrap(); - let close_brace_span = self.token.span; - - if tts.is_empty() { - let empty_block_span = open_brace_span.to(close_brace_span); - self.last_delim_empty_block_spans.insert(delim, empty_block_span); - } - - if self.open_braces.len() == 0 { - // Clear up these spans to avoid suggesting them as we've found - // properly matched delimiters so far for an entire block. - self.matching_delim_spans.clear(); - } else { - self.matching_delim_spans.push( - (open_brace, open_brace_span, close_brace_span), - ); - } - // Parse the close delimiter. - self.real_token(); - } - // Incorrect delimiter. - token::CloseDelim(other) => { - let mut unclosed_delimiter = None; - let mut candidate = None; - if self.last_unclosed_found_span != Some(self.token.span) { - // do not complain about the same unclosed delimiter multiple times - self.last_unclosed_found_span = Some(self.token.span); - // This is a conservative error: only report the last unclosed - // delimiter. The previous unclosed delimiters could actually be - // closed! The parser just hasn't gotten to them yet. - if let Some(&(_, sp)) = self.open_braces.last() { - unclosed_delimiter = Some(sp); - }; - if let Some(current_padding) = sm.span_to_margin(self.token.span) { - for (brace, brace_span) in &self.open_braces { - if let Some(padding) = sm.span_to_margin(*brace_span) { - // high likelihood of these two corresponding - if current_padding == padding && brace == &other { - candidate = Some(*brace_span); - } - } - } - } - let (tok, _) = self.open_braces.pop().unwrap(); - self.unmatched_braces.push(UnmatchedBrace { - expected_delim: tok, - found_delim: Some(other), - found_span: self.token.span, - unclosed_span: unclosed_delimiter, - candidate_span: candidate, - }); - } else { - self.open_braces.pop(); - } - - // If the incorrect delimiter matches an earlier opening - // delimiter, then don't consume it (it can be used to - // close the earlier one). Otherwise, consume it. - // E.g., we try to recover from: - // fn foo() { - // bar(baz( - // } // Incorrect delimiter but matches the earlier `{` - if !self.open_braces.iter().any(|&(b, _)| b == other) { - self.real_token(); - } - } - token::Eof => { - // Silently recover, the EOF token will be seen again - // and an error emitted then. Thus we don't pop from - // self.open_braces here. - }, - _ => {} - } - - Ok(TokenTree::Delimited( - delim_span, - delim, - tts.into() - ).into()) - }, - token::CloseDelim(delim) => { - // An unexpected closing delimiter (i.e., there is no - // matching opening delimiter). - let token_str = token_to_string(&self.token); - let msg = format!("unexpected close delimiter: `{}`", token_str); - let mut err = self.string_reader.sess.span_diagnostic - .struct_span_err(self.token.span, &msg); - - if let Some(span) = self.last_delim_empty_block_spans.remove(&delim) { - err.span_label( - span, - "this block is empty, you might have not meant to close it" - ); - } - err.span_label(self.token.span, "unexpected close delimiter"); - Err(err) - }, - _ => { - let tt = TokenTree::Token(self.token.take()); - self.real_token(); - let is_joint = self.joint_to_prev == Joint && self.token.is_op(); - Ok((tt, if is_joint { Joint } else { NonJoint })) - } - } - } - - fn real_token(&mut self) { - self.joint_to_prev = Joint; - loop { - let token = self.string_reader.next_token(); - match token.kind { - token::Whitespace | token::Comment | token::Shebang(_) | token::Unknown(_) => { - self.joint_to_prev = NonJoint; - } - _ => { - self.token = token; - return; - } - } - } - } -} - -#[derive(Default)] -struct TokenStreamBuilder { - buf: Vec<TreeAndJoint>, -} - -impl TokenStreamBuilder { - fn push(&mut self, (tree, joint): TreeAndJoint) { - if let Some((TokenTree::Token(prev_token), Joint)) = self.buf.last() { - if let TokenTree::Token(token) = &tree { - if let Some(glued) = prev_token.glue(token) { - self.buf.pop(); - self.buf.push((TokenTree::Token(glued), joint)); - return; - } - } - } - self.buf.push((tree, joint)) - } - - fn into_token_stream(self) -> TokenStream { - TokenStream::new(self.buf) - } -} diff --git a/src/libsyntax/parse/lexer/unicode_chars.rs b/src/libsyntax/parse/lexer/unicode_chars.rs deleted file mode 100644 index 525b4215aff..00000000000 --- a/src/libsyntax/parse/lexer/unicode_chars.rs +++ /dev/null @@ -1,392 +0,0 @@ -// Characters and their corresponding confusables were collected from -// http://www.unicode.org/Public/security/10.0.0/confusables.txt - -use super::StringReader; -use errors::{Applicability, DiagnosticBuilder}; -use syntax_pos::{BytePos, Pos, Span, symbol::kw}; -use crate::parse::token; - -#[rustfmt::skip] // for line breaks -const UNICODE_ARRAY: &[(char, &str, char)] = &[ - (' ', "Line Separator", ' '), - (' ', "Paragraph Separator", ' '), - (' ', "Ogham Space mark", ' '), - (' ', "En Quad", ' '), - (' ', "Em Quad", ' '), - (' ', "En Space", ' '), - (' ', "Em Space", ' '), - (' ', "Three-Per-Em Space", ' '), - (' ', "Four-Per-Em Space", ' '), - (' ', "Six-Per-Em Space", ' '), - (' ', "Punctuation Space", ' '), - (' ', "Thin Space", ' '), - (' ', "Hair Space", ' '), - (' ', "Medium Mathematical Space", ' '), - (' ', "No-Break Space", ' '), - (' ', "Figure Space", ' '), - (' ', "Narrow No-Break Space", ' '), - (' ', "Ideographic Space", ' '), - - ('ߺ', "Nko Lajanyalan", '_'), - ('﹍', "Dashed Low Line", '_'), - ('﹎', "Centreline Low Line", '_'), - ('﹏', "Wavy Low Line", '_'), - ('_', "Fullwidth Low Line", '_'), - - ('‐', "Hyphen", '-'), - ('‑', "Non-Breaking Hyphen", '-'), - ('‒', "Figure Dash", '-'), - ('–', "En Dash", '-'), - ('—', "Em Dash", '-'), - ('﹘', "Small Em Dash", '-'), - ('۔', "Arabic Full Stop", '-'), - ('⁃', "Hyphen Bullet", '-'), - ('˗', "Modifier Letter Minus Sign", '-'), - ('−', "Minus Sign", '-'), - ('➖', "Heavy Minus Sign", '-'), - ('Ⲻ', "Coptic Letter Dialect-P Ni", '-'), - ('ー', "Katakana-Hiragana Prolonged Sound Mark", '-'), - ('-', "Fullwidth Hyphen-Minus", '-'), - ('―', "Horizontal Bar", '-'), - ('─', "Box Drawings Light Horizontal", '-'), - ('━', "Box Drawings Heavy Horizontal", '-'), - ('㇐', "CJK Stroke H", '-'), - ('ꟷ', "Latin Epigraphic Letter Sideways I", '-'), - ('ᅳ', "Hangul Jungseong Eu", '-'), - ('ㅡ', "Hangul Letter Eu", '-'), - ('一', "CJK Unified Ideograph-4E00", '-'), - ('⼀', "Kangxi Radical One", '-'), - - ('؍', "Arabic Date Separator", ','), - ('٫', "Arabic Decimal Separator", ','), - ('‚', "Single Low-9 Quotation Mark", ','), - ('¸', "Cedilla", ','), - ('ꓹ', "Lisu Letter Tone Na Po", ','), - (',', "Fullwidth Comma", ','), - - (';', "Greek Question Mark", ';'), - (';', "Fullwidth Semicolon", ';'), - ('︔', "Presentation Form For Vertical Semicolon", ';'), - - ('ः', "Devanagari Sign Visarga", ':'), - ('ઃ', "Gujarati Sign Visarga", ':'), - (':', "Fullwidth Colon", ':'), - ('։', "Armenian Full Stop", ':'), - ('܃', "Syriac Supralinear Colon", ':'), - ('܄', "Syriac Sublinear Colon", ':'), - ('᛬', "Runic Multiple Punctuation", ':'), - ('︰', "Presentation Form For Vertical Two Dot Leader", ':'), - ('᠃', "Mongolian Full Stop", ':'), - ('᠉', "Mongolian Manchu Full Stop", ':'), - ('⁚', "Two Dot Punctuation", ':'), - ('׃', "Hebrew Punctuation Sof Pasuq", ':'), - ('˸', "Modifier Letter Raised Colon", ':'), - ('꞉', "Modifier Letter Colon", ':'), - ('∶', "Ratio", ':'), - ('ː', "Modifier Letter Triangular Colon", ':'), - ('ꓽ', "Lisu Letter Tone Mya Jeu", ':'), - ('︓', "Presentation Form For Vertical Colon", ':'), - - ('!', "Fullwidth Exclamation Mark", '!'), - ('ǃ', "Latin Letter Retroflex Click", '!'), - ('ⵑ', "Tifinagh Letter Tuareg Yang", '!'), - ('︕', "Presentation Form For Vertical Exclamation Mark", '!'), - - ('ʔ', "Latin Letter Glottal Stop", '?'), - ('Ɂ', "Latin Capital Letter Glottal Stop", '?'), - ('ॽ', "Devanagari Letter Glottal Stop", '?'), - ('Ꭾ', "Cherokee Letter He", '?'), - ('ꛫ', "Bamum Letter Ntuu", '?'), - ('?', "Fullwidth Question Mark", '?'), - ('︖', "Presentation Form For Vertical Question Mark", '?'), - - ('𝅭', "Musical Symbol Combining Augmentation Dot", '.'), - ('․', "One Dot Leader", '.'), - ('܁', "Syriac Supralinear Full Stop", '.'), - ('܂', "Syriac Sublinear Full Stop", '.'), - ('꘎', "Vai Full Stop", '.'), - ('𐩐', "Kharoshthi Punctuation Dot", '.'), - ('٠', "Arabic-Indic Digit Zero", '.'), - ('۰', "Extended Arabic-Indic Digit Zero", '.'), - ('ꓸ', "Lisu Letter Tone Mya Ti", '.'), - ('·', "Middle Dot", '.'), - ('・', "Katakana Middle Dot", '.'), - ('・', "Halfwidth Katakana Middle Dot", '.'), - ('᛫', "Runic Single Punctuation", '.'), - ('·', "Greek Ano Teleia", '.'), - ('⸱', "Word Separator Middle Dot", '.'), - ('𐄁', "Aegean Word Separator Dot", '.'), - ('•', "Bullet", '.'), - ('‧', "Hyphenation Point", '.'), - ('∙', "Bullet Operator", '.'), - ('⋅', "Dot Operator", '.'), - ('ꞏ', "Latin Letter Sinological Dot", '.'), - ('ᐧ', "Canadian Syllabics Final Middle Dot", '.'), - ('ᐧ', "Canadian Syllabics Final Middle Dot", '.'), - ('.', "Fullwidth Full Stop", '.'), - ('。', "Ideographic Full Stop", '.'), - ('︒', "Presentation Form For Vertical Ideographic Full Stop", '.'), - - ('՝', "Armenian Comma", '\''), - (''', "Fullwidth Apostrophe", '\''), - ('‘', "Left Single Quotation Mark", '\''), - ('’', "Right Single Quotation Mark", '\''), - ('‛', "Single High-Reversed-9 Quotation Mark", '\''), - ('′', "Prime", '\''), - ('‵', "Reversed Prime", '\''), - ('՚', "Armenian Apostrophe", '\''), - ('׳', "Hebrew Punctuation Geresh", '\''), - ('`', "Grave Accent", '\''), - ('`', "Greek Varia", '\''), - ('`', "Fullwidth Grave Accent", '\''), - ('´', "Acute Accent", '\''), - ('΄', "Greek Tonos", '\''), - ('´', "Greek Oxia", '\''), - ('᾽', "Greek Koronis", '\''), - ('᾿', "Greek Psili", '\''), - ('῾', "Greek Dasia", '\''), - ('ʹ', "Modifier Letter Prime", '\''), - ('ʹ', "Greek Numeral Sign", '\''), - ('ˈ', "Modifier Letter Vertical Line", '\''), - ('ˊ', "Modifier Letter Acute Accent", '\''), - ('ˋ', "Modifier Letter Grave Accent", '\''), - ('˴', "Modifier Letter Middle Grave Accent", '\''), - ('ʻ', "Modifier Letter Turned Comma", '\''), - ('ʽ', "Modifier Letter Reversed Comma", '\''), - ('ʼ', "Modifier Letter Apostrophe", '\''), - ('ʾ', "Modifier Letter Right Half Ring", '\''), - ('ꞌ', "Latin Small Letter Saltillo", '\''), - ('י', "Hebrew Letter Yod", '\''), - ('ߴ', "Nko High Tone Apostrophe", '\''), - ('ߵ', "Nko Low Tone Apostrophe", '\''), - ('ᑊ', "Canadian Syllabics West-Cree P", '\''), - ('ᛌ', "Runic Letter Short-Twig-Sol S", '\''), - ('𖽑', "Miao Sign Aspiration", '\''), - ('𖽒', "Miao Sign Reformed Voicing", '\''), - - ('᳓', "Vedic Sign Nihshvasa", '"'), - ('"', "Fullwidth Quotation Mark", '"'), - ('“', "Left Double Quotation Mark", '"'), - ('”', "Right Double Quotation Mark", '"'), - ('‟', "Double High-Reversed-9 Quotation Mark", '"'), - ('″', "Double Prime", '"'), - ('‶', "Reversed Double Prime", '"'), - ('〃', "Ditto Mark", '"'), - ('״', "Hebrew Punctuation Gershayim", '"'), - ('˝', "Double Acute Accent", '"'), - ('ʺ', "Modifier Letter Double Prime", '"'), - ('˶', "Modifier Letter Middle Double Acute Accent", '"'), - ('˵', "Modifier Letter Middle Double Grave Accent", '"'), - ('ˮ', "Modifier Letter Double Apostrophe", '"'), - ('ײ', "Hebrew Ligature Yiddish Double Yod", '"'), - ('❞', "Heavy Double Comma Quotation Mark Ornament", '"'), - ('❝', "Heavy Double Turned Comma Quotation Mark Ornament", '"'), - - ('(', "Fullwidth Left Parenthesis", '('), - ('❨', "Medium Left Parenthesis Ornament", '('), - ('﴾', "Ornate Left Parenthesis", '('), - - (')', "Fullwidth Right Parenthesis", ')'), - ('❩', "Medium Right Parenthesis Ornament", ')'), - ('﴿', "Ornate Right Parenthesis", ')'), - - ('[', "Fullwidth Left Square Bracket", '['), - ('❲', "Light Left Tortoise Shell Bracket Ornament", '['), - ('「', "Left Corner Bracket", '['), - ('『', "Left White Corner Bracket", '['), - ('【', "Left Black Lenticular Bracket", '['), - ('〔', "Left Tortoise Shell Bracket", '['), - ('〖', "Left White Lenticular Bracket", '['), - ('〘', "Left White Tortoise Shell Bracket", '['), - ('〚', "Left White Square Bracket", '['), - - (']', "Fullwidth Right Square Bracket", ']'), - ('❳', "Light Right Tortoise Shell Bracket Ornament", ']'), - ('」', "Right Corner Bracket", ']'), - ('』', "Right White Corner Bracket", ']'), - ('】', "Right Black Lenticular Bracket", ']'), - ('〕', "Right Tortoise Shell Bracket", ']'), - ('〗', "Right White Lenticular Bracket", ']'), - ('〙', "Right White Tortoise Shell Bracket", ']'), - ('〛', "Right White Square Bracket", ']'), - - ('❴', "Medium Left Curly Bracket Ornament", '{'), - ('𝄔', "Musical Symbol Brace", '{'), - ('{', "Fullwidth Left Curly Bracket", '{'), - - ('❵', "Medium Right Curly Bracket Ornament", '}'), - ('}', "Fullwidth Right Curly Bracket", '}'), - - ('⁎', "Low Asterisk", '*'), - ('٭', "Arabic Five Pointed Star", '*'), - ('∗', "Asterisk Operator", '*'), - ('𐌟', "Old Italic Letter Ess", '*'), - ('*', "Fullwidth Asterisk", '*'), - - ('᜵', "Philippine Single Punctuation", '/'), - ('⁁', "Caret Insertion Point", '/'), - ('∕', "Division Slash", '/'), - ('⁄', "Fraction Slash", '/'), - ('╱', "Box Drawings Light Diagonal Upper Right To Lower Left", '/'), - ('⟋', "Mathematical Rising Diagonal", '/'), - ('⧸', "Big Solidus", '/'), - ('𝈺', "Greek Instrumental Notation Symbol-47", '/'), - ('㇓', "CJK Stroke Sp", '/'), - ('〳', "Vertical Kana Repeat Mark Upper Half", '/'), - ('Ⳇ', "Coptic Capital Letter Old Coptic Esh", '/'), - ('ノ', "Katakana Letter No", '/'), - ('丿', "CJK Unified Ideograph-4E3F", '/'), - ('⼃', "Kangxi Radical Slash", '/'), - ('/', "Fullwidth Solidus", '/'), - - ('\', "Fullwidth Reverse Solidus", '\\'), - ('﹨', "Small Reverse Solidus", '\\'), - ('∖', "Set Minus", '\\'), - ('⟍', "Mathematical Falling Diagonal", '\\'), - ('⧵', "Reverse Solidus Operator", '\\'), - ('⧹', "Big Reverse Solidus", '\\'), - ('⧹', "Greek Vocal Notation Symbol-16", '\\'), - ('⧹', "Greek Instrumental Symbol-48", '\\'), - ('㇔', "CJK Stroke D", '\\'), - ('丶', "CJK Unified Ideograph-4E36", '\\'), - ('⼂', "Kangxi Radical Dot", '\\'), - ('、', "Ideographic Comma", '\\'), - ('ヽ', "Katakana Iteration Mark", '\\'), - - ('ꝸ', "Latin Small Letter Um", '&'), - ('&', "Fullwidth Ampersand", '&'), - - ('᛭', "Runic Cross Punctuation", '+'), - ('➕', "Heavy Plus Sign", '+'), - ('𐊛', "Lycian Letter H", '+'), - ('﬩', "Hebrew Letter Alternative Plus Sign", '+'), - ('+', "Fullwidth Plus Sign", '+'), - - ('‹', "Single Left-Pointing Angle Quotation Mark", '<'), - ('❮', "Heavy Left-Pointing Angle Quotation Mark Ornament", '<'), - ('˂', "Modifier Letter Left Arrowhead", '<'), - ('𝈶', "Greek Instrumental Symbol-40", '<'), - ('ᐸ', "Canadian Syllabics Pa", '<'), - ('ᚲ', "Runic Letter Kauna", '<'), - ('❬', "Medium Left-Pointing Angle Bracket Ornament", '<'), - ('⟨', "Mathematical Left Angle Bracket", '<'), - ('〈', "Left-Pointing Angle Bracket", '<'), - ('〈', "Left Angle Bracket", '<'), - ('㇛', "CJK Stroke Pd", '<'), - ('く', "Hiragana Letter Ku", '<'), - ('𡿨', "CJK Unified Ideograph-21FE8", '<'), - ('《', "Left Double Angle Bracket", '<'), - ('<', "Fullwidth Less-Than Sign", '<'), - - ('᐀', "Canadian Syllabics Hyphen", '='), - ('⹀', "Double Hyphen", '='), - ('゠', "Katakana-Hiragana Double Hyphen", '='), - ('꓿', "Lisu Punctuation Full Stop", '='), - ('=', "Fullwidth Equals Sign", '='), - - ('›', "Single Right-Pointing Angle Quotation Mark", '>'), - ('❯', "Heavy Right-Pointing Angle Quotation Mark Ornament", '>'), - ('˃', "Modifier Letter Right Arrowhead", '>'), - ('𝈷', "Greek Instrumental Symbol-42", '>'), - ('ᐳ', "Canadian Syllabics Po", '>'), - ('𖼿', "Miao Letter Archaic Zza", '>'), - ('❭', "Medium Right-Pointing Angle Bracket Ornament", '>'), - ('⟩', "Mathematical Right Angle Bracket", '>'), - ('〉', "Right-Pointing Angle Bracket", '>'), - ('〉', "Right Angle Bracket", '>'), - ('》', "Right Double Angle Bracket", '>'), - ('>', "Fullwidth Greater-Than Sign", '>'), -]; - -// FIXME: the lexer could be used to turn the ASCII version of unicode homoglyphs, instead of -// keeping the substitution token in this table. Ideally, this should be inside `rustc_lexer`. -// However, we should first remove compound tokens like `<<` from `rustc_lexer`, and then add -// fancier error recovery to it, as there will be less overall work to do this way. -const ASCII_ARRAY: &[(char, &str, Option<token::TokenKind>)] = &[ - (' ', "Space", Some(token::Whitespace)), - ('_', "Underscore", Some(token::Ident(kw::Underscore, false))), - ('-', "Minus/Hyphen", Some(token::BinOp(token::Minus))), - (',', "Comma", Some(token::Comma)), - (';', "Semicolon", Some(token::Semi)), - (':', "Colon", Some(token::Colon)), - ('!', "Exclamation Mark", Some(token::Not)), - ('?', "Question Mark", Some(token::Question)), - ('.', "Period", Some(token::Dot)), - ('(', "Left Parenthesis", Some(token::OpenDelim(token::Paren))), - (')', "Right Parenthesis", Some(token::CloseDelim(token::Paren))), - ('[', "Left Square Bracket", Some(token::OpenDelim(token::Bracket))), - (']', "Right Square Bracket", Some(token::CloseDelim(token::Bracket))), - ('{', "Left Curly Brace", Some(token::OpenDelim(token::Brace))), - ('}', "Right Curly Brace", Some(token::CloseDelim(token::Brace))), - ('*', "Asterisk", Some(token::BinOp(token::Star))), - ('/', "Slash", Some(token::BinOp(token::Slash))), - ('\\', "Backslash", None), - ('&', "Ampersand", Some(token::BinOp(token::And))), - ('+', "Plus Sign", Some(token::BinOp(token::Plus))), - ('<', "Less-Than Sign", Some(token::Lt)), - ('=', "Equals Sign", Some(token::Eq)), - ('>', "Greater-Than Sign", Some(token::Gt)), - // FIXME: Literals are already lexed by this point, so we can't recover gracefully just by - // spitting the correct token out. - ('\'', "Single Quote", None), - ('"', "Quotation Mark", None), -]; - -crate fn check_for_substitution<'a>( - reader: &StringReader<'a>, - pos: BytePos, - ch: char, - err: &mut DiagnosticBuilder<'a>, -) -> Option<token::TokenKind> { - let (u_name, ascii_char) = match UNICODE_ARRAY.iter().find(|&&(c, _, _)| c == ch) { - Some(&(_u_char, u_name, ascii_char)) => (u_name, ascii_char), - None => return None, - }; - - let span = Span::with_root_ctxt(pos, pos + Pos::from_usize(ch.len_utf8())); - - let (ascii_name, token) = match ASCII_ARRAY.iter().find(|&&(c, _, _)| c == ascii_char) { - Some((_ascii_char, ascii_name, token)) => (ascii_name, token), - None => { - let msg = format!("substitution character not found for '{}'", ch); - reader.sess.span_diagnostic.span_bug_no_panic(span, &msg); - return None; - } - }; - - // special help suggestion for "directed" double quotes - if let Some(s) = peek_delimited(&reader.src[reader.src_index(pos)..], '“', '”') { - let msg = format!( - "Unicode characters '“' (Left Double Quotation Mark) and \ - '”' (Right Double Quotation Mark) look like '{}' ({}), but are not", - ascii_char, ascii_name - ); - err.span_suggestion( - Span::with_root_ctxt( - pos, - pos + Pos::from_usize('“'.len_utf8() + s.len() + '”'.len_utf8()), - ), - &msg, - format!("\"{}\"", s), - Applicability::MaybeIncorrect, - ); - } else { - let msg = format!( - "Unicode character '{}' ({}) looks like '{}' ({}), but it is not", - ch, u_name, ascii_char, ascii_name - ); - err.span_suggestion(span, &msg, ascii_char.to_string(), Applicability::MaybeIncorrect); - } - token.clone() -} - -/// Extract string if found at current position with given delimiters -fn peek_delimited(text: &str, from_ch: char, to_ch: char) -> Option<&str> { - let mut chars = text.chars(); - let first_char = chars.next()?; - if first_char != from_ch { - return None; - } - let last_char_idx = chars.as_str().find(to_ch)?; - Some(&chars.as_str()[..last_char_idx]) -} |
