// option. This file may not be copied, modified, or distributed
// except according to those terms.
-use ast;
-use syntax_pos::{self, BytePos, CharPos, Pos, Span};
-use codemap::CodeMap;
-use errors::{FatalError, Handler, DiagnosticBuilder};
-use ext::tt::transcribe::tt_next_token;
-use parse::token::{self, keywords, str_to_ident};
+use ast::{self, Ident};
+use syntax_pos::{self, BytePos, CharPos, Pos, Span, NO_EXPANSION};
+use codemap::{CodeMap, FilePathMapping};
+use errors::{FatalError, DiagnosticBuilder};
+use parse::{token, ParseSess};
use str::char_at;
-use rustc_unicode::property::Pattern_White_Space;
+use symbol::{Symbol, keywords};
+use core::unicode::property::Pattern_White_Space;
use std::borrow::Cow;
use std::char;
use std::mem::replace;
-use std::rc::Rc;
-
-pub use ext::tt::transcribe::{TtReader, new_tt_reader, new_tt_reader_with_doc_flag};
+use rustc_data_structures::sync::Lrc;
pub mod comments;
+mod tokentrees;
mod unicode_chars;
-pub trait Reader {
- fn is_eof(&self) -> bool;
- fn try_next_token(&mut self) -> Result<TokenAndSpan, ()>;
+#[derive(Clone, PartialEq, Eq, Debug)]
+pub struct TokenAndSpan {
+ pub tok: token::Token,
+ pub sp: Span,
+}
+
+impl Default for TokenAndSpan {
+ fn default() -> Self {
+ TokenAndSpan { tok: token::Whitespace, sp: syntax_pos::DUMMY_SP }
+ }
+}
+
+pub struct StringReader<'a> {
+ pub sess: &'a ParseSess,
+ /// The absolute offset within the codemap of the next character to read
+ pub next_pos: BytePos,
+ /// The absolute offset within the codemap of the current character
+ pub pos: BytePos,
+ /// The column of the next character to read
+ pub col: CharPos,
+ /// The current character (which has been read from self.pos)
+ pub ch: Option<char>,
+ pub filemap: Lrc<syntax_pos::FileMap>,
+ /// If Some, stop reading the source at this position (inclusive).
+ pub terminator: Option<BytePos>,
+ /// Whether to record new-lines and multibyte chars in filemap.
+ /// This is only necessary the first time a filemap is lexed.
+ /// If part of a filemap is being re-lexed, this should be set to false.
+ pub save_new_lines_and_multibyte: bool,
+ // cached:
+ pub peek_tok: token::Token,
+ pub peek_span: Span,
+ pub fatal_errs: Vec<DiagnosticBuilder<'a>>,
+ // cache a direct reference to the source text, so that we don't have to
+ // retrieve it via `self.filemap.src.as_ref().unwrap()` all the time.
+ source_text: Lrc<String>,
+ /// Stack of open delimiters and their spans. Used for error message.
+ token: token::Token,
+ span: Span,
+ open_braces: Vec<(token::DelimToken, Span)>,
+ pub override_span: Option<Span>,
+}
+
+impl<'a> StringReader<'a> {
+ fn mk_sp(&self, lo: BytePos, hi: BytePos) -> Span {
+ unwrap_or!(self.override_span, Span::new(lo, hi, NO_EXPANSION))
+ }
+ fn mk_ident(&self, string: &str) -> Ident {
+ let mut ident = Ident::from_str(string);
+ if let Some(span) = self.override_span {
+ ident.span = span;
+ }
+ ident
+ }
+
fn next_token(&mut self) -> TokenAndSpan where Self: Sized {
let res = self.try_next_token();
self.unwrap_or_abort(res)
}
- /// Report a fatal error with the current span.
- fn fatal(&self, &str) -> FatalError;
- /// Report a non-fatal error with the current span.
- fn err(&self, &str);
- fn emit_fatal_errors(&mut self);
fn unwrap_or_abort(&mut self, res: Result<TokenAndSpan, ()>) -> TokenAndSpan {
match res {
Ok(tok) => tok,
Err(_) => {
self.emit_fatal_errors();
- panic!(FatalError);
+ FatalError.raise();
}
}
}
- fn peek(&self) -> TokenAndSpan;
- /// Get a token the parser cares about.
fn try_real_token(&mut self) -> Result<TokenAndSpan, ()> {
let mut t = self.try_next_token()?;
loop {
_ => break,
}
}
+ self.token = t.tok.clone();
+ self.span = t.sp;
Ok(t)
}
- fn real_token(&mut self) -> TokenAndSpan {
+ pub fn real_token(&mut self) -> TokenAndSpan {
let res = self.try_real_token();
self.unwrap_or_abort(res)
}
-}
-
-#[derive(Clone, PartialEq, Eq, Debug)]
-pub struct TokenAndSpan {
- pub tok: token::Token,
- pub sp: Span,
-}
-
-pub struct StringReader<'a> {
- pub span_diagnostic: &'a Handler,
- /// The absolute offset within the codemap of the next character to read
- pub pos: BytePos,
- /// The absolute offset within the codemap of the last character read(curr)
- pub last_pos: BytePos,
- /// The column of the next character to read
- pub col: CharPos,
- /// The last character to be read
- pub curr: Option<char>,
- pub filemap: Rc<syntax_pos::FileMap>,
- /// If Some, stop reading the source at this position (inclusive).
- pub terminator: Option<BytePos>,
- /// Whether to record new-lines in filemap. This is only necessary the first
- /// time a filemap is lexed. If part of a filemap is being re-lexed, this
- /// should be set to false.
- pub save_new_lines: bool,
- // cached:
- pub peek_tok: token::Token,
- pub peek_span: Span,
- pub fatal_errs: Vec<DiagnosticBuilder<'a>>,
- // cache a direct reference to the source text, so that we don't have to
- // retrieve it via `self.filemap.src.as_ref().unwrap()` all the time.
- source_text: Rc<String>,
-}
-
-impl<'a> Reader for StringReader<'a> {
fn is_eof(&self) -> bool {
- if self.curr.is_none() {
+ if self.ch.is_none() {
return true;
}
match self.terminator {
- Some(t) => self.pos > t,
+ Some(t) => self.next_pos > t,
None => false,
}
}
/// Return the next token. EFFECT: advances the string_reader.
- fn try_next_token(&mut self) -> Result<TokenAndSpan, ()> {
+ pub fn try_next_token(&mut self) -> Result<TokenAndSpan, ()> {
assert!(self.fatal_errs.is_empty());
let ret_val = TokenAndSpan {
- tok: replace(&mut self.peek_tok, token::Underscore),
+ tok: replace(&mut self.peek_tok, token::Whitespace),
sp: self.peek_span,
};
self.advance_token()?;
Ok(ret_val)
}
+
+ fn fail_unterminated_raw_string(&self, pos: BytePos, hash_count: u16) {
+ let mut err = self.struct_span_fatal(pos, pos, "unterminated raw string");
+ err.span_label(self.mk_sp(pos, pos), "unterminated raw string");
+ if hash_count > 0 {
+ err.note(&format!("this raw string should be terminated with `\"{}`",
+ "#".repeat(hash_count as usize)));
+ }
+ err.emit();
+ FatalError.raise();
+ }
+
fn fatal(&self, m: &str) -> FatalError {
self.fatal_span(self.peek_span, m)
}
- fn err(&self, m: &str) {
- self.err_span(self.peek_span, m)
- }
- fn emit_fatal_errors(&mut self) {
+ pub fn emit_fatal_errors(&mut self) {
for err in &mut self.fatal_errs {
err.emit();
}
self.fatal_errs.clear();
}
- fn peek(&self) -> TokenAndSpan {
+ pub fn peek(&self) -> TokenAndSpan {
// FIXME(pcwalton): Bad copy!
TokenAndSpan {
tok: self.peek_tok.clone(),
}
}
-impl<'a> Reader for TtReader<'a> {
- fn is_eof(&self) -> bool {
- self.cur_tok == token::Eof
- }
- fn try_next_token(&mut self) -> Result<TokenAndSpan, ()> {
- assert!(self.fatal_errs.is_empty());
- let r = tt_next_token(self);
- debug!("TtReader: r={:?}", r);
- Ok(r)
- }
- fn fatal(&self, m: &str) -> FatalError {
- self.sp_diag.span_fatal(self.cur_span, m)
- }
- fn err(&self, m: &str) {
- self.sp_diag.span_err(self.cur_span, m);
- }
- fn emit_fatal_errors(&mut self) {
- for err in &mut self.fatal_errs {
- err.emit();
- }
- self.fatal_errs.clear();
- }
- fn peek(&self) -> TokenAndSpan {
- TokenAndSpan {
- tok: self.cur_tok.clone(),
- sp: self.cur_span,
- }
- }
-}
-
impl<'a> StringReader<'a> {
- /// For comments.rs, which hackily pokes into pos and curr
- pub fn new_raw<'b>(span_diagnostic: &'b Handler,
- filemap: Rc<syntax_pos::FileMap>)
- -> StringReader<'b> {
- let mut sr = StringReader::new_raw_internal(span_diagnostic, filemap);
+ /// For comments.rs, which hackily pokes into next_pos and ch
+ pub fn new_raw(sess: &'a ParseSess, filemap: Lrc<syntax_pos::FileMap>) -> Self {
+ let mut sr = StringReader::new_raw_internal(sess, filemap);
sr.bump();
sr
}
- fn new_raw_internal<'b>(span_diagnostic: &'b Handler,
- filemap: Rc<syntax_pos::FileMap>)
- -> StringReader<'b> {
+ fn new_raw_internal(sess: &'a ParseSess, filemap: Lrc<syntax_pos::FileMap>) -> Self {
if filemap.src.is_none() {
- span_diagnostic.bug(&format!("Cannot lex filemap \
- without source: {}",
- filemap.name)[..]);
+ sess.span_diagnostic.bug(&format!("Cannot lex filemap without source: {}",
+ filemap.name));
}
let source_text = (*filemap.src.as_ref().unwrap()).clone();
StringReader {
- span_diagnostic: span_diagnostic,
+ sess,
+ next_pos: filemap.start_pos,
pos: filemap.start_pos,
- last_pos: filemap.start_pos,
col: CharPos(0),
- curr: Some('\n'),
- filemap: filemap,
+ ch: Some('\n'),
+ filemap,
terminator: None,
- save_new_lines: true,
+ save_new_lines_and_multibyte: true,
// dummy values; not read
peek_tok: token::Eof,
peek_span: syntax_pos::DUMMY_SP,
- source_text: source_text,
+ source_text,
fatal_errs: Vec::new(),
+ token: token::Eof,
+ span: syntax_pos::DUMMY_SP,
+ open_braces: Vec::new(),
+ override_span: None,
}
}
- pub fn new<'b>(span_diagnostic: &'b Handler,
- filemap: Rc<syntax_pos::FileMap>)
- -> StringReader<'b> {
- let mut sr = StringReader::new_raw(span_diagnostic, filemap);
- if let Err(_) = sr.advance_token() {
+ pub fn new(sess: &'a ParseSess, filemap: Lrc<syntax_pos::FileMap>) -> Self {
+ let mut sr = StringReader::new_raw(sess, filemap);
+ if sr.advance_token().is_err() {
sr.emit_fatal_errors();
- panic!(FatalError);
+ FatalError.raise();
}
sr
}
- pub fn curr_is(&self, c: char) -> bool {
- self.curr == Some(c)
+ pub fn retokenize(sess: &'a ParseSess, mut span: Span) -> Self {
+ let begin = sess.codemap().lookup_byte_offset(span.lo());
+ let end = sess.codemap().lookup_byte_offset(span.hi());
+
+ // Make the range zero-length if the span is invalid.
+ if span.lo() > span.hi() || begin.fm.start_pos != end.fm.start_pos {
+ span = span.shrink_to_lo();
+ }
+
+ let mut sr = StringReader::new_raw_internal(sess, begin.fm);
+
+ // Seek the lexer to the right byte range.
+ sr.save_new_lines_and_multibyte = false;
+ sr.next_pos = span.lo();
+ sr.terminator = Some(span.hi());
+
+ sr.bump();
+
+ if sr.advance_token().is_err() {
+ sr.emit_fatal_errors();
+ FatalError.raise();
+ }
+ sr
+ }
+
+ pub fn ch_is(&self, c: char) -> bool {
+ self.ch == Some(c)
}
/// Report a fatal lexical error with a given span.
pub fn fatal_span(&self, sp: Span, m: &str) -> FatalError {
- self.span_diagnostic.span_fatal(sp, m)
+ self.sess.span_diagnostic.span_fatal(sp, m)
}
/// Report a lexical error with a given span.
pub fn err_span(&self, sp: Span, m: &str) {
- self.span_diagnostic.span_err(sp, m)
+ self.sess.span_diagnostic.span_err(sp, m)
}
/// 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(syntax_pos::mk_sp(from_pos, to_pos), m)
+ 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(syntax_pos::mk_sp(from_pos, to_pos), m)
+ self.err_span(self.mk_sp(from_pos, to_pos), m)
+ }
+
+ /// Pushes a character to a message string for error reporting
+ fn push_escaped_char_for_msg(m: &mut String, c: char) {
+ match c {
+ '\u{20}'...'\u{7e}' => {
+ // Don't escape \, ' or " for user-facing messages
+ m.push(c);
+ }
+ _ => {
+ for c in c.escape_default() {
+ m.push(c);
+ }
+ }
+ }
}
/// Report a lexical error spanning [`from_pos`, `to_pos`), appending an
fn fatal_span_char(&self, from_pos: BytePos, to_pos: BytePos, m: &str, c: char) -> FatalError {
let mut m = m.to_string();
m.push_str(": ");
- for c in c.escape_default() {
- m.push(c)
- }
+ Self::push_escaped_char_for_msg(&mut m, c);
self.fatal_span_(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,
-> DiagnosticBuilder<'a> {
let mut m = m.to_string();
m.push_str(": ");
- for c in c.escape_default() {
- m.push(c)
- }
- self.span_diagnostic.struct_span_fatal(syntax_pos::mk_sp(from_pos, to_pos), &m[..])
+ Self::push_escaped_char_for_msg(&mut m, c);
+ self.sess.span_diagnostic.struct_span_fatal(self.mk_sp(from_pos, to_pos), &m[..])
}
/// Report a lexical error spanning [`from_pos`, `to_pos`), appending an
fn err_span_char(&self, from_pos: BytePos, to_pos: BytePos, m: &str, c: char) {
let mut m = m.to_string();
m.push_str(": ");
- for c in c.escape_default() {
- m.push(c)
- }
+ Self::push_escaped_char_for_msg(&mut m, c);
self.err_span_(from_pos, to_pos, &m[..]);
}
fn struct_err_span_char(&self,
-> DiagnosticBuilder<'a> {
let mut m = m.to_string();
m.push_str(": ");
- for c in c.escape_default() {
- m.push(c)
- }
- self.span_diagnostic.struct_span_err(syntax_pos::mk_sp(from_pos, to_pos), &m[..])
+ Self::push_escaped_char_for_msg(&mut m, c);
+ self.sess.span_diagnostic.struct_span_err(self.mk_sp(from_pos, to_pos), &m[..])
}
/// Report a lexical error spanning [`from_pos`, `to_pos`), appending the
None => {
if self.is_eof() {
self.peek_tok = token::Eof;
- self.peek_span = syntax_pos::mk_sp(self.filemap.end_pos, self.filemap.end_pos);
+ self.peek_span = self.mk_sp(self.filemap.end_pos, self.filemap.end_pos);
} else {
- let start_bytepos = self.last_pos;
+ let start_bytepos = self.pos;
self.peek_tok = self.next_token_inner()?;
- self.peek_span = syntax_pos::mk_sp(start_bytepos, self.last_pos);
+ self.peek_span = self.mk_sp(start_bytepos, self.pos);
};
}
}
}
/// Calls `f` with a string slice of the source text spanning from `start`
- /// up to but excluding `self.last_pos`, meaning the slice does not include
- /// the character `self.curr`.
+ /// up to but excluding `self.pos`, meaning the slice does not include
+ /// the character `self.ch`.
pub fn with_str_from<T, F>(&self, start: BytePos, f: F) -> T
where F: FnOnce(&str) -> T
{
- self.with_str_from_to(start, self.last_pos, f)
+ self.with_str_from_to(start, self.pos, f)
}
/// Create a Name from a given offset to the current offset, each
/// adjusted 1 towards each other (assumes that on either side there is a
/// single-byte delimiter).
pub fn name_from(&self, start: BytePos) -> ast::Name {
- debug!("taking an ident from {:?} to {:?}", start, self.last_pos);
- self.with_str_from(start, token::intern)
+ debug!("taking an ident from {:?} to {:?}", start, self.pos);
+ self.with_str_from(start, Symbol::intern)
}
/// As name_from, with an explicit endpoint.
pub fn name_from_to(&self, start: BytePos, end: BytePos) -> ast::Name {
debug!("taking an ident from {:?} to {:?}", start, end);
- self.with_str_from_to(start, end, token::intern)
+ self.with_str_from_to(start, end, Symbol::intern)
}
/// Calls `f` with a string slice of the source text spanning from `start`
/// Advance the StringReader by one character. If a newline is
/// discovered, add it to the FileMap's list of line start offsets.
pub fn bump(&mut self) {
- self.last_pos = self.pos;
- let current_byte_offset = self.byte_offset(self.pos).to_usize();
- if current_byte_offset < self.source_text.len() {
- assert!(self.curr.is_some());
- let last_char = self.curr.unwrap();
- let ch = char_at(&self.source_text, current_byte_offset);
- let next = current_byte_offset + ch.len_utf8();
- let byte_offset_diff = next - current_byte_offset;
- self.pos = self.pos + Pos::from_usize(byte_offset_diff);
- self.curr = Some(ch);
- self.col = self.col + CharPos(1);
- if last_char == '\n' {
- if self.save_new_lines {
- self.filemap.next_line(self.last_pos);
+ let new_pos = self.next_pos;
+ let new_byte_offset = self.byte_offset(new_pos).to_usize();
+ let end = self.terminator.map_or(self.source_text.len(), |t| {
+ self.byte_offset(t).to_usize()
+ });
+ if new_byte_offset < end {
+ let old_ch_is_newline = self.ch.unwrap() == '\n';
+ let new_ch = char_at(&self.source_text, new_byte_offset);
+ let new_ch_len = new_ch.len_utf8();
+
+ self.ch = Some(new_ch);
+ self.pos = new_pos;
+ self.next_pos = new_pos + Pos::from_usize(new_ch_len);
+ if old_ch_is_newline {
+ if self.save_new_lines_and_multibyte {
+ self.filemap.next_line(self.pos);
}
self.col = CharPos(0);
+ } else {
+ self.col = self.col + CharPos(1);
}
-
- if byte_offset_diff > 1 {
- self.filemap.record_multibyte_char(self.last_pos, byte_offset_diff);
+ if new_ch_len > 1 {
+ if self.save_new_lines_and_multibyte {
+ self.filemap.record_multibyte_char(self.pos, new_ch_len);
+ }
}
+ self.filemap.record_width(self.pos, new_ch);
} else {
- self.curr = None;
+ self.ch = None;
+ self.pos = new_pos;
}
}
pub fn nextch(&self) -> Option<char> {
- let offset = self.byte_offset(self.pos).to_usize();
+ let offset = self.byte_offset(self.next_pos).to_usize();
if offset < self.source_text.len() {
Some(char_at(&self.source_text, offset))
} else {
}
pub fn nextnextch(&self) -> Option<char> {
- let offset = self.byte_offset(self.pos).to_usize();
+ let offset = self.byte_offset(self.next_pos).to_usize();
let s = &self.source_text[..];
if offset >= s.len() {
return None;
/// Eats <XID_start><XID_continue>*, if possible.
fn scan_optional_raw_name(&mut self) -> Option<ast::Name> {
- if !ident_start(self.curr) {
+ if !ident_start(self.ch) {
return None;
}
- let start = self.last_pos;
- while ident_continue(self.curr) {
+ let start = self.pos;
+ while ident_continue(self.ch) {
self.bump();
}
self.with_str_from(start, |string| {
if string == "_" {
+ self.sess.span_diagnostic
+ .struct_span_warn(self.mk_sp(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(token::intern(string))
+ Some(Symbol::intern(string))
}
})
}
- /// PRECONDITION: self.curr is not whitespace
+ /// PRECONDITION: self.ch is not whitespace
/// Eats any kind of comment.
fn scan_comment(&mut self) -> Option<TokenAndSpan> {
- if let Some(c) = self.curr {
+ if let Some(c) = self.ch {
if c.is_whitespace() {
- self.span_diagnostic.span_err(syntax_pos::mk_sp(self.last_pos, self.last_pos),
- "called consume_any_line_comment, but there \
- was whitespace");
+ let msg = "called consume_any_line_comment, but there was whitespace";
+ self.sess.span_diagnostic.span_err(self.mk_sp(self.pos, self.pos), msg);
}
}
- if self.curr_is('/') {
+ if self.ch_is('/') {
match self.nextch() {
Some('/') => {
self.bump();
self.bump();
// line comments starting with "///" or "//!" are doc-comments
- let doc_comment = self.curr_is('/') || self.curr_is('!');
- let start_bpos = if doc_comment {
- self.pos - BytePos(3)
- } else {
- self.last_pos - BytePos(2)
- };
+ let doc_comment = (self.ch_is('/') && !self.nextch_is('/')) || self.ch_is('!');
+ let start_bpos = self.pos - BytePos(2);
while !self.is_eof() {
- match self.curr.unwrap() {
+ match self.ch.unwrap() {
'\n' => break,
'\r' => {
if self.nextch_is('\n') {
// CRLF
break;
} else if doc_comment {
- self.err_span_(self.last_pos,
- self.pos,
+ self.err_span_(self.pos,
+ self.next_pos,
"bare CR not allowed in doc-comment");
}
}
self.bump();
}
- return if doc_comment {
+ if doc_comment {
self.with_str_from(start_bpos, |string| {
// comments with only more "/"s are not doc comments
let tok = if is_doc_comment(string) {
- token::DocComment(token::intern(string))
+ token::DocComment(Symbol::intern(string))
} else {
token::Comment
};
Some(TokenAndSpan {
- tok: tok,
- sp: syntax_pos::mk_sp(start_bpos, self.last_pos),
+ tok,
+ sp: self.mk_sp(start_bpos, self.pos),
})
})
} else {
Some(TokenAndSpan {
tok: token::Comment,
- sp: syntax_pos::mk_sp(start_bpos, self.last_pos),
+ sp: self.mk_sp(start_bpos, self.pos),
})
- };
+ }
}
Some('*') => {
self.bump();
}
_ => None,
}
- } else if self.curr_is('#') {
+ } else if self.ch_is('#') {
if self.nextch_is('!') {
// Parse an inner attribute.
// I guess this is the only way to figure out if
// we're at the beginning of the file...
- let cmap = CodeMap::new();
- cmap.files.borrow_mut().push(self.filemap.clone());
- let loc = cmap.lookup_char_pos_adj(self.last_pos);
+ let cmap = CodeMap::new(FilePathMapping::empty());
+ cmap.files.borrow_mut().file_maps.push(self.filemap.clone());
+ let loc = cmap.lookup_char_pos_adj(self.pos);
debug!("Skipping a shebang");
if loc.line == 1 && loc.col == CharPos(0) {
// FIXME: Add shebang "token", return it
- let start = self.last_pos;
- while !self.curr_is('\n') && !self.is_eof() {
+ let start = self.pos;
+ while !self.ch_is('\n') && !self.is_eof() {
self.bump();
}
return Some(TokenAndSpan {
tok: token::Shebang(self.name_from(start)),
- sp: syntax_pos::mk_sp(start, self.last_pos),
+ sp: self.mk_sp(start, self.pos),
});
}
}
/// If there is whitespace, shebang, or a comment, scan it. Otherwise,
/// return None.
fn scan_whitespace_or_comment(&mut self) -> Option<TokenAndSpan> {
- match self.curr.unwrap_or('\0') {
+ match self.ch.unwrap_or('\0') {
// # to handle shebang at start of file -- this is the entry point
// for skipping over all "junk"
'/' | '#' => {
c
},
c if is_pattern_whitespace(Some(c)) => {
- let start_bpos = self.last_pos;
- while is_pattern_whitespace(self.curr) {
+ let start_bpos = self.pos;
+ while is_pattern_whitespace(self.ch) {
self.bump();
}
let c = Some(TokenAndSpan {
tok: token::Whitespace,
- sp: syntax_pos::mk_sp(start_bpos, self.last_pos),
+ sp: self.mk_sp(start_bpos, self.pos),
});
debug!("scanning whitespace: {:?}", c);
c
/// Might return a sugared-doc-attr
fn scan_block_comment(&mut self) -> Option<TokenAndSpan> {
// block comments starting with "/**" or "/*!" are doc-comments
- let is_doc_comment = self.curr_is('*') || self.curr_is('!');
- let start_bpos = self.last_pos - BytePos(2);
+ let is_doc_comment = self.ch_is('*') || self.ch_is('!');
+ let start_bpos = self.pos - BytePos(2);
let mut level: isize = 1;
let mut has_cr = false;
} else {
"unterminated block comment"
};
- let last_bpos = self.last_pos;
- panic!(self.fatal_span_(start_bpos, last_bpos, msg));
+ let last_bpos = self.pos;
+ self.fatal_span_(start_bpos, last_bpos, msg).raise();
}
- let n = self.curr.unwrap();
+ let n = self.ch.unwrap();
match n {
'/' if self.nextch_is('*') => {
level += 1;
} else {
string.into()
};
- token::DocComment(token::intern(&string[..]))
+ token::DocComment(Symbol::intern(&string[..]))
} else {
token::Comment
};
Some(TokenAndSpan {
- tok: tok,
- sp: syntax_pos::mk_sp(start_bpos, self.last_pos),
+ tok,
+ sp: self.mk_sp(start_bpos, self.pos),
})
})
}
assert!(real_radix <= scan_radix);
let mut len = 0;
loop {
- let c = self.curr;
+ let c = self.ch;
if c == Some('_') {
debug!("skipping a _");
self.bump();
// check that the hypothetical digit is actually
// in range for the true radix
if c.unwrap().to_digit(real_radix).is_none() {
- self.err_span_(self.last_pos,
- self.pos,
+ self.err_span_(self.pos,
+ self.next_pos,
&format!("invalid digit for a base {} literal", real_radix));
}
len += 1;
fn scan_number(&mut self, c: char) -> token::Lit {
let num_digits;
let mut base = 10;
- let start_bpos = self.last_pos;
+ let start_bpos = self.pos;
self.bump();
if c == '0' {
- match self.curr.unwrap_or('\0') {
+ match self.ch.unwrap_or('\0') {
'b' => {
self.bump();
base = 2;
base = 16;
num_digits = self.scan_digits(16, 16);
}
- '0'...'9' | '_' | '.' => {
+ '0'...'9' | '_' | '.' | 'e' | 'E' => {
num_digits = self.scan_digits(10, 10) + 1;
}
_ => {
if num_digits == 0 {
self.err_span_(start_bpos,
- self.last_pos,
+ self.pos,
"no valid digits found for number");
- return token::Integer(token::intern("0"));
+ return token::Integer(Symbol::intern("0"));
}
// might be a float, but don't be greedy if this is actually an
// integer literal followed by field/method access or a range pattern
// (`0..2` and `12.foo()`)
- if self.curr_is('.') && !self.nextch_is('.') &&
- !self.nextch()
- .unwrap_or('\0')
- .is_xid_start() {
+ if self.ch_is('.') && !self.nextch_is('.') &&
+ !ident_start(self.nextch()) {
// might have stuff after the ., and if it does, it needs to start
// with a number
self.bump();
- if self.curr.unwrap_or('\0').is_digit(10) {
+ if self.ch.unwrap_or('\0').is_digit(10) {
self.scan_digits(10, 10);
self.scan_float_exponent();
}
- let last_pos = self.last_pos;
- self.check_float_base(start_bpos, last_pos, base);
- return token::Float(self.name_from(start_bpos));
+ let pos = self.pos;
+ self.check_float_base(start_bpos, pos, base);
+ token::Float(self.name_from(start_bpos))
} else {
// it might be a float if it has an exponent
- if self.curr_is('e') || self.curr_is('E') {
+ if self.ch_is('e') || self.ch_is('E') {
self.scan_float_exponent();
- let last_pos = self.last_pos;
- self.check_float_base(start_bpos, last_pos, base);
+ let pos = self.pos;
+ self.check_float_base(start_bpos, pos, base);
return token::Float(self.name_from(start_bpos));
}
// but we certainly have an integer!
- return token::Integer(self.name_from(start_bpos));
+ token::Integer(self.name_from(start_bpos))
}
}
/// error if too many or too few digits are encountered.
fn scan_hex_digits(&mut self, n_digits: usize, delim: char, below_0x7f_only: bool) -> bool {
debug!("scanning {} digits until {:?}", n_digits, delim);
- let start_bpos = self.last_pos;
+ let start_bpos = self.pos;
let mut accum_int = 0;
let mut valid = true;
for _ in 0..n_digits {
if self.is_eof() {
- let last_bpos = self.last_pos;
- panic!(self.fatal_span_(start_bpos,
- last_bpos,
- "unterminated numeric character escape"));
+ let last_bpos = self.pos;
+ self.fatal_span_(start_bpos,
+ last_bpos,
+ "unterminated numeric character escape").raise();
}
- if self.curr_is(delim) {
- let last_bpos = self.last_pos;
+ if self.ch_is(delim) {
+ let last_bpos = self.pos;
self.err_span_(start_bpos,
last_bpos,
"numeric character escape is too short");
valid = false;
break;
}
- let c = self.curr.unwrap_or('\x00');
+ let c = self.ch.unwrap_or('\x00');
accum_int *= 16;
accum_int += c.to_digit(16).unwrap_or_else(|| {
- self.err_span_char(self.last_pos,
- self.pos,
+ self.err_span_char(self.pos,
+ self.next_pos,
"invalid character in numeric character escape",
c);
if below_0x7f_only && accum_int >= 0x80 {
self.err_span_(start_bpos,
- self.last_pos,
+ self.pos,
"this form of character escape may only be used with characters in \
the range [\\x00-\\x7f]");
valid = false;
match char::from_u32(accum_int) {
Some(_) => valid,
None => {
- let last_bpos = self.last_pos;
+ let last_bpos = self.pos;
self.err_span_(start_bpos, last_bpos, "invalid numeric character escape");
false
}
match first_source_char {
'\\' => {
// '\X' for some X must be a character constant:
- let escaped = self.curr;
- let escaped_pos = self.last_pos;
+ let escaped = self.ch;
+ let escaped_pos = self.pos;
self.bump();
match escaped {
None => {} // EOF here is an error that will be checked later.
'n' | 'r' | 't' | '\\' | '\'' | '"' | '0' => true,
'x' => self.scan_byte_escape(delim, !ascii_only),
'u' => {
- let valid = if self.curr_is('{') {
+ let valid = if self.ch_is('{') {
self.scan_unicode_escape(delim) && !ascii_only
} else {
- let span = syntax_pos::mk_sp(start, self.last_pos);
- self.span_diagnostic
+ let span = self.mk_sp(start, self.pos);
+ self.sess.span_diagnostic
.struct_span_err(span, "incorrect unicode escape sequence")
.span_help(span,
"format of unicode escape sequences is \
};
if ascii_only {
self.err_span_(start,
- self.last_pos,
+ self.pos,
"unicode escape sequences cannot be used as a \
byte or in a byte string");
}
self.consume_whitespace();
true
}
- '\r' if delim == '"' && self.curr_is('\n') => {
+ '\r' if delim == '"' && self.ch_is('\n') => {
self.consume_whitespace();
true
}
c => {
- let last_pos = self.last_pos;
+ let pos = self.pos;
let mut err = self.struct_err_span_char(escaped_pos,
- last_pos,
+ pos,
if ascii_only {
"unknown byte escape"
} else {
},
c);
if e == '\r' {
- err.span_help(syntax_pos::mk_sp(escaped_pos, last_pos),
+ err.span_help(self.mk_sp(escaped_pos, pos),
"this is an isolated carriage return; consider \
checking your editor and version control \
settings");
}
if (e == '{' || e == '}') && !ascii_only {
- err.span_help(syntax_pos::mk_sp(escaped_pos, last_pos),
+ err.span_help(self.mk_sp(escaped_pos, pos),
"if used in a formatting string, curly braces \
are escaped with `{{` and `}}`");
}
}
}
'\t' | '\n' | '\r' | '\'' if delim == '\'' => {
- let last_pos = self.last_pos;
+ let pos = self.pos;
self.err_span_char(start,
- last_pos,
+ pos,
if ascii_only {
"byte constant must be escaped"
} else {
return false;
}
'\r' => {
- if self.curr_is('\n') {
+ if self.ch_is('\n') {
self.bump();
return true;
} else {
self.err_span_(start,
- self.last_pos,
+ self.pos,
"bare CR not allowed in string, use \\r instead");
return false;
}
}
_ => {
if ascii_only && first_source_char > '\x7F' {
- let last_pos = self.last_pos;
+ let pos = self.pos;
self.err_span_(start,
- last_pos,
+ pos,
"byte constant must be ASCII. Use a \\xHH escape for a \
non-ASCII byte");
return false;
true
}
- /// Scan over a \u{...} escape
+ /// Scan over a `\u{...}` escape
///
- /// At this point, we have already seen the \ and the u, the { is the current character. We
- /// will read at least one digit, and up to 6, and pass over the }.
+ /// At this point, we have already seen the `\` and the `u`, the `{` is the current character.
+ /// We will read a hex number (with `_` separators), with 1 to 6 actual digits,
+ /// and pass over the `}`.
fn scan_unicode_escape(&mut self, delim: char) -> bool {
self.bump(); // past the {
- let start_bpos = self.last_pos;
- let mut count = 0;
- let mut accum_int = 0;
+ let start_bpos = self.pos;
let mut valid = true;
- while !self.curr_is('}') && count <= 6 {
- let c = match self.curr {
- Some(c) => c,
- None => {
- panic!(self.fatal_span_(start_bpos,
- self.last_pos,
- "unterminated unicode escape (found EOF)"));
- }
- };
- accum_int *= 16;
- accum_int += c.to_digit(16).unwrap_or_else(|| {
- if c == delim {
- panic!(self.fatal_span_(self.last_pos,
- self.pos,
- "unterminated unicode escape (needed a `}`)"));
- } else {
- self.err_span_char(self.last_pos,
- self.pos,
- "invalid character in unicode escape",
- c);
- }
- valid = false;
- 0
- });
- self.bump();
- count += 1;
+ if let Some('_') = self.ch {
+ // disallow leading `_`
+ self.err_span_(self.pos,
+ self.next_pos,
+ "invalid start of unicode escape");
+ valid = false;
}
+ let count = self.scan_digits(16, 16);
+
if count > 6 {
self.err_span_(start_bpos,
- self.last_pos,
- "overlong unicode escape (can have at most 6 hex digits)");
+ self.pos,
+ "overlong unicode escape (must have at most 6 hex digits)");
valid = false;
}
-
- if valid && (char::from_u32(accum_int).is_none() || count == 0) {
- self.err_span_(start_bpos,
- self.last_pos,
- "invalid unicode character escape");
- valid = false;
+ loop {
+ match self.ch {
+ Some('}') => {
+ if valid && count == 0 {
+ self.err_span_(start_bpos,
+ self.pos,
+ "empty unicode escape (must have at least 1 hex digit)");
+ valid = false;
+ }
+ self.bump(); // past the ending `}`
+ break;
+ },
+ Some(c) => {
+ if c == delim {
+ self.err_span_(self.pos,
+ self.pos,
+ "unterminated unicode escape (needed a `}`)");
+ valid = false;
+ break;
+ } else if valid {
+ self.err_span_char(start_bpos,
+ self.pos,
+ "invalid character in unicode escape",
+ c);
+ valid = false;
+ }
+ },
+ None => {
+ self.fatal_span_(start_bpos,
+ self.pos,
+ "unterminated unicode escape (found EOF)").raise();
+ }
+ }
+ self.bump();
}
-
- self.bump(); // past the ending }
valid
}
/// Scan over a float exponent.
fn scan_float_exponent(&mut self) {
- if self.curr_is('e') || self.curr_is('E') {
+ if self.ch_is('e') || self.ch_is('E') {
self.bump();
- if self.curr_is('-') || self.curr_is('+') {
+ if self.ch_is('-') || self.ch_is('+') {
self.bump();
}
if self.scan_digits(10, 10) == 0 {
- self.err_span_(self.last_pos,
- self.pos,
+ self.err_span_(self.pos,
+ self.next_pos,
"expected at least one digit in exponent")
}
}
fn binop(&mut self, op: token::BinOpToken) -> token::Token {
self.bump();
- if self.curr_is('=') {
+ if self.ch_is('=') {
self.bump();
- return token::BinOpEq(op);
+ token::BinOpEq(op)
} else {
- return token::BinOp(op);
+ token::BinOp(op)
}
}
/// Return the next token from the string, advances the input past that
/// token, and updates the interner
fn next_token_inner(&mut self) -> Result<token::Token, ()> {
- let c = self.curr;
- if ident_start(c) &&
- match (c.unwrap(), self.nextch(), self.nextnextch()) {
- // Note: r as in r" or r#" is part of a raw string literal,
- // b as in b' is part of a byte literal.
- // They are not identifiers, and are handled further down.
- ('r', Some('"'), _) |
- ('r', Some('#'), _) |
- ('b', Some('"'), _) |
- ('b', Some('\''), _) |
- ('b', Some('r'), Some('"')) |
- ('b', Some('r'), Some('#')) => false,
- _ => true,
- } {
- let start = self.last_pos;
- while ident_continue(self.curr) {
- self.bump();
- }
+ let c = self.ch;
+
+ if ident_start(c) {
+ let (is_ident_start, is_raw_ident) =
+ match (c.unwrap(), self.nextch(), self.nextnextch()) {
+ // r# followed by an identifier starter is a raw identifier.
+ // This is an exception to the r# case below.
+ ('r', Some('#'), x) if ident_start(x) => (true, true),
+ // r as in r" or r#" is part of a raw string literal.
+ // b as in b' is part of a byte literal.
+ // They are not identifiers, and are handled further down.
+ ('r', Some('"'), _) |
+ ('r', Some('#'), _) |
+ ('b', Some('"'), _) |
+ ('b', Some('\''), _) |
+ ('b', Some('r'), Some('"')) |
+ ('b', Some('r'), Some('#')) => (false, false),
+ _ => (true, false),
+ };
+ if is_ident_start {
+ let raw_start = self.pos;
+ if is_raw_ident {
+ // Consume the 'r#' characters.
+ self.bump();
+ self.bump();
+ }
- return Ok(self.with_str_from(start, |string| {
- if string == "_" {
- token::Underscore
- } else {
- // FIXME: perform NFKC normalization here. (Issue #2253)
- token::Ident(str_to_ident(string))
+ let start = self.pos;
+ while ident_continue(self.ch) {
+ self.bump();
}
- }));
+
+ return Ok(self.with_str_from(start, |string| {
+ // FIXME: perform NFKC normalization here. (Issue #2253)
+ let ident = self.mk_ident(string);
+ if is_raw_ident && (token::is_path_segment_keyword(ident) ||
+ ident.name == keywords::Underscore.name()) {
+ self.fatal_span_(raw_start, self.pos,
+ &format!("`r#{}` is not currently supported.", ident.name)
+ ).raise();
+ }
+ if is_raw_ident {
+ let span = self.mk_sp(raw_start, self.pos);
+ self.sess.raw_identifier_spans.borrow_mut().push(span);
+ }
+ token::Ident(ident, is_raw_ident)
+ }));
+ }
}
if is_dec_digit(c) {
// One-byte tokens.
';' => {
self.bump();
- return Ok(token::Semi);
+ Ok(token::Semi)
}
',' => {
self.bump();
- return Ok(token::Comma);
+ Ok(token::Comma)
}
'.' => {
self.bump();
- return if self.curr_is('.') {
+ if self.ch_is('.') {
self.bump();
- if self.curr_is('.') {
+ if self.ch_is('.') {
self.bump();
Ok(token::DotDotDot)
+ } else if self.ch_is('=') {
+ self.bump();
+ Ok(token::DotDotEq)
} else {
Ok(token::DotDot)
}
} else {
Ok(token::Dot)
- };
+ }
}
'(' => {
self.bump();
- return Ok(token::OpenDelim(token::Paren));
+ Ok(token::OpenDelim(token::Paren))
}
')' => {
self.bump();
- return Ok(token::CloseDelim(token::Paren));
+ Ok(token::CloseDelim(token::Paren))
}
'{' => {
self.bump();
- return Ok(token::OpenDelim(token::Brace));
+ Ok(token::OpenDelim(token::Brace))
}
'}' => {
self.bump();
- return Ok(token::CloseDelim(token::Brace));
+ Ok(token::CloseDelim(token::Brace))
}
'[' => {
self.bump();
- return Ok(token::OpenDelim(token::Bracket));
+ Ok(token::OpenDelim(token::Bracket))
}
']' => {
self.bump();
- return Ok(token::CloseDelim(token::Bracket));
+ Ok(token::CloseDelim(token::Bracket))
}
'@' => {
self.bump();
- return Ok(token::At);
+ Ok(token::At)
}
'#' => {
self.bump();
- return Ok(token::Pound);
+ Ok(token::Pound)
}
'~' => {
self.bump();
- return Ok(token::Tilde);
+ Ok(token::Tilde)
}
'?' => {
self.bump();
- return Ok(token::Question);
+ Ok(token::Question)
}
':' => {
self.bump();
- if self.curr_is(':') {
+ if self.ch_is(':') {
self.bump();
- return Ok(token::ModSep);
+ Ok(token::ModSep)
} else {
- return Ok(token::Colon);
+ Ok(token::Colon)
}
}
'$' => {
self.bump();
- return Ok(token::Dollar);
+ Ok(token::Dollar)
}
// Multi-byte tokens.
'=' => {
self.bump();
- if self.curr_is('=') {
+ if self.ch_is('=') {
self.bump();
- return Ok(token::EqEq);
- } else if self.curr_is('>') {
+ Ok(token::EqEq)
+ } else if self.ch_is('>') {
self.bump();
- return Ok(token::FatArrow);
+ Ok(token::FatArrow)
} else {
- return Ok(token::Eq);
+ Ok(token::Eq)
}
}
'!' => {
self.bump();
- if self.curr_is('=') {
+ if self.ch_is('=') {
self.bump();
- return Ok(token::Ne);
+ Ok(token::Ne)
} else {
- return Ok(token::Not);
+ Ok(token::Not)
}
}
'<' => {
self.bump();
- match self.curr.unwrap_or('\x00') {
+ match self.ch.unwrap_or('\x00') {
'=' => {
self.bump();
- return Ok(token::Le);
+ Ok(token::Le)
}
'<' => {
- return Ok(self.binop(token::Shl));
+ Ok(self.binop(token::Shl))
}
'-' => {
self.bump();
- match self.curr.unwrap_or('\x00') {
+ match self.ch.unwrap_or('\x00') {
_ => {
- return Ok(token::LArrow);
+ Ok(token::LArrow)
}
}
}
_ => {
- return Ok(token::Lt);
+ Ok(token::Lt)
}
}
}
'>' => {
self.bump();
- match self.curr.unwrap_or('\x00') {
+ match self.ch.unwrap_or('\x00') {
'=' => {
self.bump();
- return Ok(token::Ge);
+ Ok(token::Ge)
}
'>' => {
- return Ok(self.binop(token::Shr));
+ Ok(self.binop(token::Shr))
}
_ => {
- return Ok(token::Gt);
+ Ok(token::Gt)
}
}
}
'\'' => {
// Either a character constant 'a' OR a lifetime name 'abc
- let start_with_quote = self.last_pos;
+ let start_with_quote = self.pos;
self.bump();
- let start = self.last_pos;
+ let start = self.pos;
// the eof will be picked up by the final `'` check below
- let c2 = self.curr.unwrap_or('\x00');
+ let c2 = self.ch.unwrap_or('\x00');
self.bump();
// If the character is an ident start not followed by another single
// quote, then this is a lifetime name:
- if ident_start(Some(c2)) && !self.curr_is('\'') {
- while ident_continue(self.curr) {
+ if ident_start(Some(c2)) && !self.ch_is('\'') {
+ while ident_continue(self.ch) {
self.bump();
}
// lifetimes shouldn't end with a single quote
// if we find one, then this is an invalid character literal
- if self.curr_is('\'') {
- panic!(self.fatal_span_verbose(
- start_with_quote, self.pos,
- String::from("character literal may only contain one codepoint")));
+ if self.ch_is('\'') {
+ self.fatal_span_verbose(start_with_quote, self.next_pos,
+ String::from("character literal may only contain one codepoint"))
+ .raise();
}
// expansion purposes. See #12512 for the gory details of why
// this is necessary.
let ident = self.with_str_from(start, |lifetime_name| {
- str_to_ident(&format!("'{}", lifetime_name))
+ self.mk_ident(&format!("'{}", lifetime_name))
});
- // Conjure up a "keyword checking ident" to make sure that
- // the lifetime name is not a keyword.
- let keyword_checking_ident = self.with_str_from(start, |lifetime_name| {
- str_to_ident(lifetime_name)
- });
- let keyword_checking_token = &token::Ident(keyword_checking_ident);
- let last_bpos = self.last_pos;
- if keyword_checking_token.is_any_keyword() &&
- !keyword_checking_token.is_keyword(keywords::Static) {
- self.err_span_(start, last_bpos, "lifetimes cannot use keyword names");
- }
-
return Ok(token::Lifetime(ident));
}
false,
'\'');
- if !self.curr_is('\'') {
- panic!(self.fatal_span_verbose(
- start_with_quote, self.last_pos,
- String::from("character literal may only contain one codepoint")));
+ if !self.ch_is('\'') {
+ let pos = self.pos;
+ loop {
+ self.bump();
+ if self.ch_is('\'') {
+ let start = self.byte_offset(start).to_usize();
+ let end = self.byte_offset(self.pos).to_usize();
+ self.bump();
+ let span = self.mk_sp(start_with_quote, self.pos);
+ self.sess.span_diagnostic
+ .struct_span_err(span,
+ "character literal may only contain one codepoint")
+ .span_suggestion(span,
+ "if you meant to write a `str` literal, \
+ use double quotes",
+ format!("\"{}\"",
+ &self.source_text[start..end]))
+ .emit();
+ return Ok(token::Literal(token::Str_(Symbol::intern("??")), None))
+ }
+ if self.ch_is('\n') || self.is_eof() || self.ch_is('/') {
+ // Only attempt to infer single line string literals. If we encounter
+ // a slash, bail out in order to avoid nonsensical suggestion when
+ // involving comments.
+ break;
+ }
+ }
+ self.fatal_span_verbose(start_with_quote, pos,
+ String::from("character literal may only contain one codepoint")).raise();
}
let id = if valid {
self.name_from(start)
} else {
- token::intern("0")
+ Symbol::intern("0")
};
- self.bump(); // advance curr past token
+ self.bump(); // advance ch past token
let suffix = self.scan_optional_raw_name();
- return Ok(token::Literal(token::Char(id), suffix));
+ Ok(token::Literal(token::Char(id), suffix))
}
'b' => {
self.bump();
- let lit = match self.curr {
+ let lit = match self.ch {
Some('\'') => self.scan_byte(),
Some('"') => self.scan_byte_string(),
Some('r') => self.scan_raw_byte_string(),
_ => unreachable!(), // Should have been a token::Ident above.
};
let suffix = self.scan_optional_raw_name();
- return Ok(token::Literal(lit, suffix));
+ Ok(token::Literal(lit, suffix))
}
'"' => {
- let start_bpos = self.last_pos;
+ let start_bpos = self.pos;
let mut valid = true;
self.bump();
- while !self.curr_is('"') {
+ while !self.ch_is('"') {
if self.is_eof() {
- let last_bpos = self.last_pos;
- panic!(self.fatal_span_(start_bpos,
- last_bpos,
- "unterminated double quote string"));
+ let last_bpos = self.pos;
+ self.fatal_span_(start_bpos,
+ last_bpos,
+ "unterminated double quote string").raise();
}
- let ch_start = self.last_pos;
- let ch = self.curr.unwrap();
+ let ch_start = self.pos;
+ let ch = self.ch.unwrap();
self.bump();
valid &= self.scan_char_or_byte(ch_start,
ch,
let id = if valid {
self.name_from(start_bpos + BytePos(1))
} else {
- token::intern("??")
+ Symbol::intern("??")
};
self.bump();
let suffix = self.scan_optional_raw_name();
- return Ok(token::Literal(token::Str_(id), suffix));
+ Ok(token::Literal(token::Str_(id), suffix))
}
'r' => {
- let start_bpos = self.last_pos;
+ let start_bpos = self.pos;
self.bump();
- let mut hash_count = 0;
- while self.curr_is('#') {
+ let mut hash_count: u16 = 0;
+ while self.ch_is('#') {
self.bump();
hash_count += 1;
}
if self.is_eof() {
- let last_bpos = self.last_pos;
- panic!(self.fatal_span_(start_bpos, last_bpos, "unterminated raw string"));
- } else if !self.curr_is('"') {
- let last_bpos = self.last_pos;
- let curr_char = self.curr.unwrap();
- panic!(self.fatal_span_char(start_bpos,
- last_bpos,
- "found invalid character; only `#` is allowed \
- in raw string delimitation",
- curr_char));
+ self.fail_unterminated_raw_string(start_bpos, hash_count);
+ } else if !self.ch_is('"') {
+ let last_bpos = self.pos;
+ let curr_char = self.ch.unwrap();
+ self.fatal_span_char(start_bpos,
+ last_bpos,
+ "found invalid character; only `#` is allowed \
+ in raw string delimitation",
+ curr_char).raise();
}
self.bump();
- let content_start_bpos = self.last_pos;
+ let content_start_bpos = self.pos;
let mut content_end_bpos;
let mut valid = true;
'outer: loop {
if self.is_eof() {
- let last_bpos = self.last_pos;
- panic!(self.fatal_span_(start_bpos, last_bpos, "unterminated raw string"));
+ self.fail_unterminated_raw_string(start_bpos, hash_count);
}
- // if self.curr_is('"') {
- // content_end_bpos = self.last_pos;
+ // if self.ch_is('"') {
+ // content_end_bpos = self.pos;
// for _ in 0..hash_count {
// self.bump();
- // if !self.curr_is('#') {
+ // if !self.ch_is('#') {
// continue 'outer;
- let c = self.curr.unwrap();
+ let c = self.ch.unwrap();
match c {
'"' => {
- content_end_bpos = self.last_pos;
+ content_end_bpos = self.pos;
for _ in 0..hash_count {
self.bump();
- if !self.curr_is('#') {
+ if !self.ch_is('#') {
continue 'outer;
}
}
}
'\r' => {
if !self.nextch_is('\n') {
- let last_bpos = self.last_pos;
+ let last_bpos = self.pos;
self.err_span_(start_bpos,
last_bpos,
"bare CR not allowed in raw string, use \\r \
let id = if valid {
self.name_from_to(content_start_bpos, content_end_bpos)
} else {
- token::intern("??")
+ Symbol::intern("??")
};
let suffix = self.scan_optional_raw_name();
- return Ok(token::Literal(token::StrRaw(id, hash_count), suffix));
+ Ok(token::Literal(token::StrRaw(id, hash_count), suffix))
}
'-' => {
if self.nextch_is('>') {
self.bump();
self.bump();
- return Ok(token::RArrow);
+ Ok(token::RArrow)
} else {
- return Ok(self.binop(token::Minus));
+ Ok(self.binop(token::Minus))
}
}
'&' => {
if self.nextch_is('&') {
self.bump();
self.bump();
- return Ok(token::AndAnd);
+ Ok(token::AndAnd)
} else {
- return Ok(self.binop(token::And));
+ Ok(self.binop(token::And))
}
}
'|' => {
Some('|') => {
self.bump();
self.bump();
- return Ok(token::OrOr);
+ Ok(token::OrOr)
}
_ => {
- return Ok(self.binop(token::Or));
+ Ok(self.binop(token::Or))
}
}
}
'+' => {
- return Ok(self.binop(token::Plus));
+ Ok(self.binop(token::Plus))
}
'*' => {
- return Ok(self.binop(token::Star));
+ Ok(self.binop(token::Star))
}
'/' => {
- return Ok(self.binop(token::Slash));
+ Ok(self.binop(token::Slash))
}
'^' => {
- return Ok(self.binop(token::Caret));
+ Ok(self.binop(token::Caret))
}
'%' => {
- return Ok(self.binop(token::Percent));
+ Ok(self.binop(token::Percent))
}
c => {
- let last_bpos = self.last_pos;
- let bpos = self.pos;
+ let last_bpos = self.pos;
+ let bpos = self.next_pos;
let mut err = self.struct_fatal_span_char(last_bpos,
bpos,
"unknown start of token",
c);
- unicode_chars::check_for_substitution(&self, c, &mut err);
+ unicode_chars::check_for_substitution(self, c, &mut err);
self.fatal_errs.push(err);
Err(())
}
}
fn consume_whitespace(&mut self) {
- while is_pattern_whitespace(self.curr) && !self.is_eof() {
+ while is_pattern_whitespace(self.ch) && !self.is_eof() {
self.bump();
}
}
fn read_to_eol(&mut self) -> String {
let mut val = String::new();
- while !self.curr_is('\n') && !self.is_eof() {
- val.push(self.curr.unwrap());
+ while !self.ch_is('\n') && !self.is_eof() {
+ val.push(self.ch.unwrap());
self.bump();
}
- if self.curr_is('\n') {
+ if self.ch_is('\n') {
self.bump();
}
- return val;
+ val
}
fn read_one_line_comment(&mut self) -> String {
let val = self.read_to_eol();
assert!((val.as_bytes()[0] == b'/' && val.as_bytes()[1] == b'/') ||
(val.as_bytes()[0] == b'#' && val.as_bytes()[1] == b'!'));
- return val;
+ val
}
fn consume_non_eol_whitespace(&mut self) {
- while is_pattern_whitespace(self.curr) && !self.curr_is('\n') && !self.is_eof() {
+ while is_pattern_whitespace(self.ch) && !self.ch_is('\n') && !self.is_eof() {
self.bump();
}
}
fn peeking_at_comment(&self) -> bool {
- (self.curr_is('/') && self.nextch_is('/')) || (self.curr_is('/') && self.nextch_is('*')) ||
+ (self.ch_is('/') && self.nextch_is('/')) || (self.ch_is('/') && self.nextch_is('*')) ||
// consider shebangs comments, but not inner attributes
- (self.curr_is('#') && self.nextch_is('!') && !self.nextnextch_is('['))
+ (self.ch_is('#') && self.nextch_is('!') && !self.nextnextch_is('['))
}
fn scan_byte(&mut self) -> token::Lit {
self.bump();
- let start = self.last_pos;
+ let start = self.pos;
// the eof will be picked up by the final `'` check below
- let c2 = self.curr.unwrap_or('\x00');
+ let c2 = self.ch.unwrap_or('\x00');
self.bump();
let valid = self.scan_char_or_byte(start,
// ascii_only =
true,
'\'');
- if !self.curr_is('\'') {
+ if !self.ch_is('\'') {
// Byte offsetting here is okay because the
// character before position `start` are an
// ascii single quote and ascii 'b'.
- let last_pos = self.last_pos;
- panic!(self.fatal_span_verbose(start - BytePos(2),
- last_pos,
- "unterminated byte constant".to_string()));
+ let pos = self.pos;
+ self.fatal_span_verbose(start - BytePos(2),
+ pos,
+ "unterminated byte constant".to_string()).raise();
}
let id = if valid {
self.name_from(start)
} else {
- token::intern("?")
+ Symbol::intern("?")
};
- self.bump(); // advance curr past token
- return token::Byte(id);
+ self.bump(); // advance ch past token
+ token::Byte(id)
}
fn scan_byte_escape(&mut self, delim: char, below_0x7f_only: bool) -> bool {
fn scan_byte_string(&mut self) -> token::Lit {
self.bump();
- let start = self.last_pos;
+ let start = self.pos;
let mut valid = true;
- while !self.curr_is('"') {
+ while !self.ch_is('"') {
if self.is_eof() {
- let last_pos = self.last_pos;
- panic!(self.fatal_span_(start, last_pos, "unterminated double quote byte string"));
+ let pos = self.pos;
+ self.fatal_span_(start, pos, "unterminated double quote byte string").raise();
}
- let ch_start = self.last_pos;
- let ch = self.curr.unwrap();
+ let ch_start = self.pos;
+ let ch = self.ch.unwrap();
self.bump();
valid &= self.scan_char_or_byte(ch_start,
ch,
let id = if valid {
self.name_from(start)
} else {
- token::intern("??")
+ Symbol::intern("??")
};
self.bump();
- return token::ByteStr(id);
+ token::ByteStr(id)
}
fn scan_raw_byte_string(&mut self) -> token::Lit {
- let start_bpos = self.last_pos;
+ let start_bpos = self.pos;
self.bump();
let mut hash_count = 0;
- while self.curr_is('#') {
+ while self.ch_is('#') {
self.bump();
hash_count += 1;
}
if self.is_eof() {
- let last_pos = self.last_pos;
- panic!(self.fatal_span_(start_bpos, last_pos, "unterminated raw string"));
- } else if !self.curr_is('"') {
- let last_pos = self.last_pos;
- let ch = self.curr.unwrap();
- panic!(self.fatal_span_char(start_bpos,
- last_pos,
+ self.fail_unterminated_raw_string(start_bpos, hash_count);
+ } else if !self.ch_is('"') {
+ let pos = self.pos;
+ let ch = self.ch.unwrap();
+ self.fatal_span_char(start_bpos,
+ pos,
"found invalid character; only `#` is allowed in raw \
string delimitation",
- ch));
+ ch).raise();
}
self.bump();
- let content_start_bpos = self.last_pos;
+ let content_start_bpos = self.pos;
let mut content_end_bpos;
'outer: loop {
- match self.curr {
+ match self.ch {
None => {
- let last_pos = self.last_pos;
- panic!(self.fatal_span_(start_bpos, last_pos, "unterminated raw string"))
+ self.fail_unterminated_raw_string(start_bpos, hash_count);
}
Some('"') => {
- content_end_bpos = self.last_pos;
+ content_end_bpos = self.pos;
for _ in 0..hash_count {
self.bump();
- if !self.curr_is('#') {
+ if !self.ch_is('#') {
continue 'outer;
}
}
}
Some(c) => {
if c > '\x7F' {
- let last_pos = self.last_pos;
- self.err_span_char(last_pos, last_pos, "raw byte string must be ASCII", c);
+ let pos = self.pos;
+ self.err_span_char(pos, pos, "raw byte string must be ASCII", c);
}
}
}
self.bump();
}
self.bump();
- return token::ByteStrRaw(self.name_from_to(content_start_bpos, content_end_bpos),
- hash_count);
+ token::ByteStrRaw(self.name_from_to(content_start_bpos, content_end_bpos),
+ hash_count)
}
}
}
fn is_dec_digit(c: Option<char>) -> bool {
- return in_range(c, '0', '9');
+ in_range(c, '0', '9')
}
pub fn is_doc_comment(s: &str) -> bool {
mod tests {
use super::*;
+ use ast::{Ident, CrateConfig};
+ use symbol::Symbol;
use syntax_pos::{BytePos, Span, NO_EXPANSION};
use codemap::CodeMap;
use errors;
+ use feature_gate::UnstableFeatures;
use parse::token;
- use parse::token::str_to_ident;
+ use std::collections::HashSet;
use std::io;
- use std::rc::Rc;
-
- fn mk_sh(cm: Rc<CodeMap>) -> errors::Handler {
- // FIXME (#22405): Replace `Box::new` with `box` here when/if possible.
+ use std::path::PathBuf;
+ use diagnostics::plugin::ErrorMap;
+ use rustc_data_structures::sync::Lock;
+ use with_globals;
+ fn mk_sess(cm: Lrc<CodeMap>) -> ParseSess {
let emitter = errors::emitter::EmitterWriter::new(Box::new(io::sink()),
- Some(cm));
- errors::Handler::with_emitter(true, false, Box::new(emitter))
+ Some(cm.clone()),
+ false,
+ false);
+ ParseSess {
+ span_diagnostic: errors::Handler::with_emitter(true, false, Box::new(emitter)),
+ unstable_features: UnstableFeatures::from_environment(),
+ config: CrateConfig::new(),
+ included_mod_stack: Lock::new(Vec::new()),
+ code_map: cm,
+ missing_fragment_specifiers: Lock::new(HashSet::new()),
+ raw_identifier_spans: Lock::new(Vec::new()),
+ registered_diagnostics: Lock::new(ErrorMap::new()),
+ non_modrs_mods: Lock::new(vec![]),
+ }
}
// open a string reader for the given string
fn setup<'a>(cm: &CodeMap,
- span_handler: &'a errors::Handler,
+ sess: &'a ParseSess,
teststr: String)
-> StringReader<'a> {
- let fm = cm.new_filemap("zebra.rs".to_string(), None, teststr);
- StringReader::new(span_handler, fm)
+ let fm = cm.new_filemap(PathBuf::from("zebra.rs").into(), teststr);
+ StringReader::new(sess, fm)
}
#[test]
fn t1() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- let mut string_reader = setup(&cm,
- &sh,
- "/* my source file */ fn main() { println!(\"zebra\"); }\n"
- .to_string());
- let id = str_to_ident("fn");
- assert_eq!(string_reader.next_token().tok, token::Comment);
- assert_eq!(string_reader.next_token().tok, token::Whitespace);
- let tok1 = string_reader.next_token();
- let tok2 = TokenAndSpan {
- tok: token::Ident(id),
- sp: Span {
- lo: BytePos(21),
- hi: BytePos(23),
- expn_id: NO_EXPANSION,
- },
- };
- assert_eq!(tok1, tok2);
- assert_eq!(string_reader.next_token().tok, token::Whitespace);
- // the 'main' id is already read:
- assert_eq!(string_reader.last_pos.clone(), BytePos(28));
- // read another token:
- let tok3 = string_reader.next_token();
- let tok4 = TokenAndSpan {
- tok: token::Ident(str_to_ident("main")),
- sp: Span {
- lo: BytePos(24),
- hi: BytePos(28),
- expn_id: NO_EXPANSION,
- },
- };
- assert_eq!(tok3, tok4);
- // the lparen is already read:
- assert_eq!(string_reader.last_pos.clone(), BytePos(29))
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ let mut string_reader = setup(&cm,
+ &sh,
+ "/* my source file */ fn main() { println!(\"zebra\"); }\n"
+ .to_string());
+ let id = Ident::from_str("fn");
+ assert_eq!(string_reader.next_token().tok, token::Comment);
+ assert_eq!(string_reader.next_token().tok, token::Whitespace);
+ let tok1 = string_reader.next_token();
+ let tok2 = TokenAndSpan {
+ tok: token::Ident(id, false),
+ sp: Span::new(BytePos(21), BytePos(23), NO_EXPANSION),
+ };
+ assert_eq!(tok1, tok2);
+ assert_eq!(string_reader.next_token().tok, token::Whitespace);
+ // the 'main' id is already read:
+ assert_eq!(string_reader.pos.clone(), BytePos(28));
+ // read another token:
+ let tok3 = string_reader.next_token();
+ let tok4 = TokenAndSpan {
+ tok: mk_ident("main"),
+ sp: Span::new(BytePos(24), BytePos(28), NO_EXPANSION),
+ };
+ assert_eq!(tok3, tok4);
+ // the lparen is already read:
+ assert_eq!(string_reader.pos.clone(), BytePos(29))
+ })
}
// check that the given reader produces the desired stream
// make the identifier by looking up the string in the interner
fn mk_ident(id: &str) -> token::Token {
- token::Ident(str_to_ident(id))
+ token::Token::from_ast_ident(Ident::from_str(id))
}
#[test]
fn doublecolonparsing() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- check_tokenization(setup(&cm, &sh, "a b".to_string()),
- vec![mk_ident("a"), token::Whitespace, mk_ident("b")]);
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ check_tokenization(setup(&cm, &sh, "a b".to_string()),
+ vec![mk_ident("a"), token::Whitespace, mk_ident("b")]);
+ })
}
#[test]
fn dcparsing_2() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- check_tokenization(setup(&cm, &sh, "a::b".to_string()),
- vec![mk_ident("a"), token::ModSep, mk_ident("b")]);
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ check_tokenization(setup(&cm, &sh, "a::b".to_string()),
+ vec![mk_ident("a"), token::ModSep, mk_ident("b")]);
+ })
}
#[test]
fn dcparsing_3() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- check_tokenization(setup(&cm, &sh, "a ::b".to_string()),
- vec![mk_ident("a"), token::Whitespace, token::ModSep, mk_ident("b")]);
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ check_tokenization(setup(&cm, &sh, "a ::b".to_string()),
+ vec![mk_ident("a"), token::Whitespace, token::ModSep, mk_ident("b")]);
+ })
}
#[test]
fn dcparsing_4() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- check_tokenization(setup(&cm, &sh, "a:: b".to_string()),
- vec![mk_ident("a"), token::ModSep, token::Whitespace, mk_ident("b")]);
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ check_tokenization(setup(&cm, &sh, "a:: b".to_string()),
+ vec![mk_ident("a"), token::ModSep, token::Whitespace, mk_ident("b")]);
+ })
}
#[test]
fn character_a() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- assert_eq!(setup(&cm, &sh, "'a'".to_string()).next_token().tok,
- token::Literal(token::Char(token::intern("a")), None));
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ assert_eq!(setup(&cm, &sh, "'a'".to_string()).next_token().tok,
+ token::Literal(token::Char(Symbol::intern("a")), None));
+ })
}
#[test]
fn character_space() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- assert_eq!(setup(&cm, &sh, "' '".to_string()).next_token().tok,
- token::Literal(token::Char(token::intern(" ")), None));
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ assert_eq!(setup(&cm, &sh, "' '".to_string()).next_token().tok,
+ token::Literal(token::Char(Symbol::intern(" ")), None));
+ })
}
#[test]
fn character_escaped() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- assert_eq!(setup(&cm, &sh, "'\\n'".to_string()).next_token().tok,
- token::Literal(token::Char(token::intern("\\n")), None));
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ assert_eq!(setup(&cm, &sh, "'\\n'".to_string()).next_token().tok,
+ token::Literal(token::Char(Symbol::intern("\\n")), None));
+ })
}
#[test]
fn lifetime_name() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- assert_eq!(setup(&cm, &sh, "'abc".to_string()).next_token().tok,
- token::Lifetime(token::str_to_ident("'abc")));
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ assert_eq!(setup(&cm, &sh, "'abc".to_string()).next_token().tok,
+ token::Lifetime(Ident::from_str("'abc")));
+ })
}
#[test]
fn raw_string() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- assert_eq!(setup(&cm, &sh, "r###\"\"#a\\b\x00c\"\"###".to_string())
- .next_token()
- .tok,
- token::Literal(token::StrRaw(token::intern("\"#a\\b\x00c\""), 3), None));
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ assert_eq!(setup(&cm, &sh, "r###\"\"#a\\b\x00c\"\"###".to_string())
+ .next_token()
+ .tok,
+ token::Literal(token::StrRaw(Symbol::intern("\"#a\\b\x00c\""), 3), None));
+ })
}
#[test]
fn literal_suffixes() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- macro_rules! test {
- ($input: expr, $tok_type: ident, $tok_contents: expr) => {{
- assert_eq!(setup(&cm, &sh, format!("{}suffix", $input)).next_token().tok,
- token::Literal(token::$tok_type(token::intern($tok_contents)),
- Some(token::intern("suffix"))));
- // with a whitespace separator:
- assert_eq!(setup(&cm, &sh, format!("{} suffix", $input)).next_token().tok,
- token::Literal(token::$tok_type(token::intern($tok_contents)),
- None));
- }}
- }
-
- test!("'a'", Char, "a");
- test!("b'a'", Byte, "a");
- test!("\"a\"", Str_, "a");
- test!("b\"a\"", ByteStr, "a");
- test!("1234", Integer, "1234");
- test!("0b101", Integer, "0b101");
- test!("0xABC", Integer, "0xABC");
- test!("1.0", Float, "1.0");
- test!("1.0e10", Float, "1.0e10");
-
- assert_eq!(setup(&cm, &sh, "2us".to_string()).next_token().tok,
- token::Literal(token::Integer(token::intern("2")),
- Some(token::intern("us"))));
- assert_eq!(setup(&cm, &sh, "r###\"raw\"###suffix".to_string()).next_token().tok,
- token::Literal(token::StrRaw(token::intern("raw"), 3),
- Some(token::intern("suffix"))));
- assert_eq!(setup(&cm, &sh, "br###\"raw\"###suffix".to_string()).next_token().tok,
- token::Literal(token::ByteStrRaw(token::intern("raw"), 3),
- Some(token::intern("suffix"))));
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ macro_rules! test {
+ ($input: expr, $tok_type: ident, $tok_contents: expr) => {{
+ assert_eq!(setup(&cm, &sh, format!("{}suffix", $input)).next_token().tok,
+ token::Literal(token::$tok_type(Symbol::intern($tok_contents)),
+ Some(Symbol::intern("suffix"))));
+ // with a whitespace separator:
+ assert_eq!(setup(&cm, &sh, format!("{} suffix", $input)).next_token().tok,
+ token::Literal(token::$tok_type(Symbol::intern($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(&cm, &sh, "2us".to_string()).next_token().tok,
+ token::Literal(token::Integer(Symbol::intern("2")),
+ Some(Symbol::intern("us"))));
+ assert_eq!(setup(&cm, &sh, "r###\"raw\"###suffix".to_string()).next_token().tok,
+ token::Literal(token::StrRaw(Symbol::intern("raw"), 3),
+ Some(Symbol::intern("suffix"))));
+ assert_eq!(setup(&cm, &sh, "br###\"raw\"###suffix".to_string()).next_token().tok,
+ token::Literal(token::ByteStrRaw(Symbol::intern("raw"), 3),
+ Some(Symbol::intern("suffix"))));
+ })
}
#[test]
#[test]
fn nested_block_comments() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- let mut lexer = setup(&cm, &sh, "/* /* */ */'a'".to_string());
- match lexer.next_token().tok {
- token::Comment => {}
- _ => panic!("expected a comment!"),
- }
- assert_eq!(lexer.next_token().tok,
- token::Literal(token::Char(token::intern("a")), None));
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ let mut lexer = setup(&cm, &sh, "/* /* */ */'a'".to_string());
+ match lexer.next_token().tok {
+ token::Comment => {}
+ _ => panic!("expected a comment!"),
+ }
+ assert_eq!(lexer.next_token().tok,
+ token::Literal(token::Char(Symbol::intern("a")), None));
+ })
}
#[test]
fn crlf_comments() {
- let cm = Rc::new(CodeMap::new());
- let sh = mk_sh(cm.clone());
- let mut lexer = setup(&cm, &sh, "// test\r\n/// test\r\n".to_string());
- let comment = lexer.next_token();
- assert_eq!(comment.tok, token::Comment);
- assert_eq!(comment.sp, ::syntax_pos::mk_sp(BytePos(0), BytePos(7)));
- assert_eq!(lexer.next_token().tok, token::Whitespace);
- assert_eq!(lexer.next_token().tok,
- token::DocComment(token::intern("/// test")));
+ with_globals(|| {
+ let cm = Lrc::new(CodeMap::new(FilePathMapping::empty()));
+ let sh = mk_sess(cm.clone());
+ let mut lexer = setup(&cm, &sh, "// test\r\n/// test\r\n".to_string());
+ let comment = lexer.next_token();
+ assert_eq!(comment.tok, token::Comment);
+ assert_eq!((comment.sp.lo(), comment.sp.hi()), (BytePos(0), BytePos(7)));
+ assert_eq!(lexer.next_token().tok, token::Whitespace);
+ assert_eq!(lexer.next_token().tok,
+ token::DocComment(Symbol::intern("/// test")));
+ })
}
}
projects / rustc.git / blobdiff