Fix exponent numbers causing record ids to fail (#4049)
This commit is contained in:
parent
47054b2891
commit
8077c15f41
2 changed files with 52 additions and 29 deletions
|
@ -50,7 +50,7 @@ impl Lexer<'_> {
|
||||||
match self.reader.peek() {
|
match self.reader.peek() {
|
||||||
Some(b'd' | b'f') => {
|
Some(b'd' | b'f') => {
|
||||||
// not an integer but parse anyway for error reporting.
|
// not an integer but parse anyway for error reporting.
|
||||||
return self.lex_suffix(false, false);
|
return self.lex_suffix(false, false, false);
|
||||||
}
|
}
|
||||||
Some(x) if x.is_ascii_alphabetic() => return self.invalid_suffix_token(),
|
Some(x) if x.is_ascii_alphabetic() => return self.invalid_suffix_token(),
|
||||||
_ => {}
|
_ => {}
|
||||||
|
@ -96,7 +96,7 @@ impl Lexer<'_> {
|
||||||
return self.finish_number_token(NumberKind::Integer);
|
return self.finish_number_token(NumberKind::Integer);
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
b'f' | b'd' => return self.lex_suffix(false, false),
|
b'f' | b'd' => return self.lex_suffix(false, false, false),
|
||||||
// Oxc2 is the start byte of 'µ'
|
// Oxc2 is the start byte of 'µ'
|
||||||
0xc2 | b'n' | b'u' | b'm' | b'h' | b'w' | b'y' | b's' => {
|
0xc2 | b'n' | b'u' | b'm' | b'h' | b'w' | b'y' | b's' => {
|
||||||
// duration suffix, switch to lexing duration.
|
// duration suffix, switch to lexing duration.
|
||||||
|
@ -132,13 +132,13 @@ impl Lexer<'_> {
|
||||||
}
|
}
|
||||||
|
|
||||||
/// Lex a number suffix, either 'f' or 'dec'.
|
/// Lex a number suffix, either 'f' or 'dec'.
|
||||||
fn lex_suffix(&mut self, had_mantissa: bool, had_exponent: bool) -> Token {
|
fn lex_suffix(&mut self, had_mantissa: bool, had_exponent: bool, had_operator: bool) -> Token {
|
||||||
match self.reader.peek() {
|
match self.reader.peek() {
|
||||||
Some(b'f') => {
|
Some(b'f') => {
|
||||||
// float suffix
|
// float suffix
|
||||||
self.reader.next();
|
self.reader.next();
|
||||||
if let Some(true) = self.reader.peek().map(|x| x.is_identifier_continue()) {
|
if let Some(true) = self.reader.peek().map(|x| x.is_identifier_continue()) {
|
||||||
if self.flexible_ident && !had_mantissa {
|
if self.flexible_ident && !had_mantissa && !had_operator {
|
||||||
self.scratch.push('f');
|
self.scratch.push('f');
|
||||||
self.lex_ident()
|
self.lex_ident()
|
||||||
} else {
|
} else {
|
||||||
|
@ -158,7 +158,7 @@ impl Lexer<'_> {
|
||||||
self.reader.next();
|
self.reader.next();
|
||||||
let checkpoint = self.reader.offset();
|
let checkpoint = self.reader.offset();
|
||||||
if !self.eat(b'e') {
|
if !self.eat(b'e') {
|
||||||
if !had_mantissa && !had_exponent {
|
if !had_mantissa && !had_exponent && !had_operator {
|
||||||
self.reader.backup(checkpoint - 1);
|
self.reader.backup(checkpoint - 1);
|
||||||
return self.lex_duration();
|
return self.lex_duration();
|
||||||
} else if !had_mantissa && self.flexible_ident {
|
} else if !had_mantissa && self.flexible_ident {
|
||||||
|
@ -215,7 +215,7 @@ impl Lexer<'_> {
|
||||||
self.scratch.push('e');
|
self.scratch.push('e');
|
||||||
return self.lex_exponent(true);
|
return self.lex_exponent(true);
|
||||||
}
|
}
|
||||||
b'f' | b'd' => return self.lex_suffix(true, false),
|
b'f' | b'd' => return self.lex_suffix(true, false, false),
|
||||||
b'a'..=b'z' | b'A'..=b'Z' => {
|
b'a'..=b'z' | b'A'..=b'Z' => {
|
||||||
// invalid token, random identifier characters immediately after number.
|
// invalid token, random identifier characters immediately after number.
|
||||||
self.scratch.clear();
|
self.scratch.clear();
|
||||||
|
@ -230,33 +230,39 @@ impl Lexer<'_> {
|
||||||
|
|
||||||
/// Lexes the exponent of a number, i.e. `e10` in `1.1e10`;
|
/// Lexes the exponent of a number, i.e. `e10` in `1.1e10`;
|
||||||
fn lex_exponent(&mut self, had_mantissa: bool) -> Token {
|
fn lex_exponent(&mut self, had_mantissa: bool) -> Token {
|
||||||
loop {
|
let mut had_operator = false;
|
||||||
match self.reader.peek() {
|
let mut peek = self.reader.peek();
|
||||||
Some(x @ b'-' | x @ b'+') => {
|
|
||||||
|
if let Some(x @ b'-' | x @ b'+') = peek {
|
||||||
|
had_operator = true;
|
||||||
self.reader.next();
|
self.reader.next();
|
||||||
self.scratch.push(x as char);
|
self.scratch.push(x as char);
|
||||||
|
peek = self.reader.peek();
|
||||||
}
|
}
|
||||||
Some(x @ b'0'..=b'9') => {
|
|
||||||
|
if let Some(x @ b'0'..=b'9') = peek {
|
||||||
|
self.reader.next();
|
||||||
self.scratch.push(x as char);
|
self.scratch.push(x as char);
|
||||||
break;
|
} else {
|
||||||
}
|
if self.flexible_ident && !had_mantissa && !had_operator {
|
||||||
_ => {
|
|
||||||
if self.flexible_ident && !had_mantissa {
|
|
||||||
return self.lex_ident();
|
return self.lex_ident();
|
||||||
}
|
}
|
||||||
// random other character, expected atleast one digit.
|
|
||||||
return self.invalid_token(LexError::Number(Error::DigitExpectedExponent));
|
return self.invalid_token(LexError::Number(Error::DigitExpectedExponent));
|
||||||
}
|
}
|
||||||
}
|
|
||||||
}
|
|
||||||
self.reader.next();
|
|
||||||
loop {
|
loop {
|
||||||
match self.reader.peek() {
|
match self.reader.peek() {
|
||||||
Some(x @ (b'0'..=b'9' | b'_')) => {
|
Some(x @ (b'0'..=b'9' | b'_')) => {
|
||||||
self.reader.next();
|
self.reader.next();
|
||||||
self.scratch.push(x as char);
|
self.scratch.push(x as char);
|
||||||
}
|
}
|
||||||
Some(b'f' | b'd') => return self.lex_suffix(had_mantissa, true),
|
Some(b'f' | b'd') => return self.lex_suffix(had_mantissa, true, had_operator),
|
||||||
|
Some(x) if x.is_identifier_continue() => {
|
||||||
|
if self.flexible_ident && !had_operator && !had_mantissa {
|
||||||
|
return self.lex_ident();
|
||||||
|
}
|
||||||
|
return self.invalid_token(LexError::Number(Error::InvalidSuffix));
|
||||||
|
}
|
||||||
_ => {
|
_ => {
|
||||||
let kind = if had_mantissa {
|
let kind = if had_mantissa {
|
||||||
NumberKind::MantissaExponent
|
NumberKind::MantissaExponent
|
||||||
|
|
|
@ -277,15 +277,30 @@ impl Parser<'_> {
|
||||||
Ok(Id::String(text))
|
Ok(Id::String(text))
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
TokenKind::Number(NumberKind::Decimal | NumberKind::DecimalExponent)
|
TokenKind::Number(NumberKind::Exponent) if self.flexible_record_id => {
|
||||||
if self.flexible_record_id =>
|
let text = self.lexer.string.take().unwrap();
|
||||||
{
|
if text.bytes().any(|x| !x.is_ascii_alphanumeric()) {
|
||||||
|
unexpected!(self, token.kind, "a identifier");
|
||||||
|
}
|
||||||
|
Ok(Id::String(text))
|
||||||
|
}
|
||||||
|
TokenKind::Number(NumberKind::Decimal) if self.flexible_record_id => {
|
||||||
let mut text = self.lexer.string.take().unwrap();
|
let mut text = self.lexer.string.take().unwrap();
|
||||||
text.push('d');
|
text.push('d');
|
||||||
text.push('e');
|
text.push('e');
|
||||||
text.push('c');
|
text.push('c');
|
||||||
Ok(Id::String(text))
|
Ok(Id::String(text))
|
||||||
}
|
}
|
||||||
|
TokenKind::Number(NumberKind::DecimalExponent) if self.flexible_record_id => {
|
||||||
|
let mut text = self.lexer.string.take().unwrap();
|
||||||
|
if text.bytes().any(|x| !x.is_ascii_alphanumeric()) {
|
||||||
|
unexpected!(self, token.kind, "a identifier");
|
||||||
|
}
|
||||||
|
text.push('d');
|
||||||
|
text.push('e');
|
||||||
|
text.push('c');
|
||||||
|
Ok(Id::String(text))
|
||||||
|
}
|
||||||
TokenKind::Number(NumberKind::Float) if self.flexible_record_id => {
|
TokenKind::Number(NumberKind::Float) if self.flexible_record_id => {
|
||||||
let mut text = self.lexer.string.take().unwrap();
|
let mut text = self.lexer.string.take().unwrap();
|
||||||
text.push('f');
|
text.push('f');
|
||||||
|
@ -565,5 +580,7 @@ mod tests {
|
||||||
assert_ident_parses_correctly("1ns");
|
assert_ident_parses_correctly("1ns");
|
||||||
assert_ident_parses_correctly("1ns1");
|
assert_ident_parses_correctly("1ns1");
|
||||||
assert_ident_parses_correctly("1ns1h");
|
assert_ident_parses_correctly("1ns1h");
|
||||||
|
assert_ident_parses_correctly("000e8");
|
||||||
|
assert_ident_parses_correctly("000e8bla");
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
Loading…
Reference in a new issue