Fix number parsing.

This commit is contained in:
Stephen Chung 2020-03-03 21:39:25 +08:00
parent 9f80bf03c4
commit 71a3c79915
4 changed files with 133 additions and 69 deletions

View File

@ -683,12 +683,29 @@ a.x = 500;
a.update();
```
## Numbers
```rust
let x = 123; // i64
let x = 123.4; // f64
let x = 123_456_789; // separators can be put anywhere inside the number
let x = 0x12abcd; // i64 in hex
let x = 0o777; // i64 in oct
let x = 0b1010_1111; // i64 in binary
```
Conversion functions:
* `to_int` - converts an `f32` or `f64` to `i64`
* `to_float` - converts an integer type to `f64`
## Strings and Chars
```rust
let name = "Bob";
let middle_initial = 'C';
let last = 'Davis';
let last = "Davis";
let full_name = name + " " + middle_initial + ". " + last;
full_name == "Bob C. Davis";
@ -706,9 +723,13 @@ let c = "foo"[0]; // a syntax error for now - cannot index into literals
let c = ts.s[0]; // a syntax error for now - cannot index into properties
let c = record[0]; // this works
// Escape sequences in strings
record += " \u2764\n"; // escape sequence of '❤' in Unicode
record == "Bob C. Davis: age 42 ❤\n"; // '\n' = new-line
// Unlike Rust, Rhai strings can be modified
record[4] = 'Z';
record == "Bob Z. Davis: age 42";
record[4] = '\x58'; // 0x58 = 'X'
record == "Bob X. Davis: age 42 ❤\n";
```
The following standard functions operate on strings:
@ -727,6 +748,7 @@ full_name.len() == 14;
full_name.trim();
full_name.len() == 12;
full_name == "Bob C. Davis";
full_name.pad(15, '$');
full_name.len() == 15;

View File

@ -34,9 +34,24 @@
// needs to be here, because order matters for macros
macro_rules! debug_println {
() => (#[cfg(feature = "debug_msgs")] {print!("\n")});
($fmt:expr) => (#[cfg(feature = "debug_msgs")] {print!(concat!($fmt, "\n"))});
($fmt:expr, $($arg:tt)*) => (#[cfg(feature = "debug_msgs")] {print!(concat!($fmt, "\n"), $($arg)*)});
() => (
#[cfg(feature = "debug_msgs")]
{
print!("\n");
}
);
($fmt:expr) => (
#[cfg(feature = "debug_msgs")]
{
print!(concat!($fmt, "\n"));
}
);
($fmt:expr, $($arg:tt)*) => (
#[cfg(feature = "debug_msgs")]
{
print!(concat!($fmt, "\n"), $($arg)*);
}
);
}
mod any;
@ -50,6 +65,6 @@ mod scope;
pub use any::Dynamic;
pub use engine::{Array, Engine, EvalAltResult};
pub use scope::Scope;
pub use fn_register::{RegisterDynamicFn, RegisterFn};
pub use parser::{ParseError, ParseErrorType, AST};
pub use scope::Scope;

View File

@ -9,9 +9,9 @@ use std::str::Chars;
pub enum LexError {
UnexpectedChar(char),
UnterminatedString,
MalformedEscapeSequence,
MalformedNumber,
MalformedChar,
MalformedEscapeSequence(String),
MalformedNumber(String),
MalformedChar(String),
InputError(String),
}
@ -22,9 +22,9 @@ impl Error for LexError {
match *self {
LERR::UnexpectedChar(_) => "Unexpected character",
LERR::UnterminatedString => "Open string is not terminated",
LERR::MalformedEscapeSequence => "Unexpected values in escape sequence",
LERR::MalformedNumber => "Unexpected characters in number",
LERR::MalformedChar => "Char constant not a single character",
LERR::MalformedEscapeSequence(_) => "Unexpected values in escape sequence",
LERR::MalformedNumber(_) => "Unexpected characters in number",
LERR::MalformedChar(_) => "Char constant not a single character",
LERR::InputError(_) => "Input error",
}
}
@ -34,6 +34,9 @@ impl fmt::Display for LexError {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
match self {
LERR::UnexpectedChar(c) => write!(f, "Unexpected '{}'", c),
LERR::MalformedEscapeSequence(s) => write!(f, "Invalid escape sequence: '{}'", s),
LERR::MalformedNumber(s) => write!(f, "Invalid number: '{}'", s),
LERR::MalformedChar(s) => write!(f, "Invalid character: '{}'", s),
LERR::InputError(s) => write!(f, "{}", s),
_ => write!(f, "{}", self.description()),
}
@ -535,7 +538,7 @@ impl<'a> TokenIterator<'a> {
enclosing_char: char,
) -> Result<String, (LexError, Position)> {
let mut result = Vec::new();
let mut escape = false;
let mut escape = String::with_capacity(12);
loop {
let next_char = self.char_stream.next();
@ -547,107 +550,123 @@ impl<'a> TokenIterator<'a> {
self.advance()?;
match next_char.unwrap() {
'\\' if !escape => escape = true,
'\\' if escape => {
escape = false;
'\\' if escape.is_empty() => {
escape.push('\\');
}
'\\' if !escape.is_empty() => {
escape.clear();
result.push('\\');
}
't' if escape => {
escape = false;
't' if !escape.is_empty() => {
escape.clear();
result.push('\t');
}
'n' if escape => {
escape = false;
'n' if !escape.is_empty() => {
escape.clear();
result.push('\n');
}
'r' if escape => {
escape = false;
'r' if !escape.is_empty() => {
escape.clear();
result.push('\r');
}
'x' if escape => {
escape = false;
'x' if !escape.is_empty() => {
let mut seq = escape.clone();
seq.push('x');
escape.clear();
let mut out_val: u32 = 0;
for _ in 0..2 {
if let Some(c) = self.char_stream.next() {
seq.push(c);
self.advance()?;
if let Some(d1) = c.to_digit(16) {
out_val *= 16;
out_val += d1;
} else {
return Err((LERR::MalformedEscapeSequence, self.pos));
return Err((LERR::MalformedEscapeSequence(seq), self.pos));
}
} else {
return Err((LERR::MalformedEscapeSequence, self.pos));
return Err((LERR::MalformedEscapeSequence(seq), self.pos));
}
self.advance()?;
}
if let Some(r) = char::from_u32(out_val) {
result.push(r);
} else {
return Err((LERR::MalformedEscapeSequence, self.pos));
return Err((LERR::MalformedEscapeSequence(seq), self.pos));
}
}
'u' if escape => {
escape = false;
'u' if !escape.is_empty() => {
let mut seq = escape.clone();
seq.push('u');
escape.clear();
let mut out_val: u32 = 0;
for _ in 0..4 {
if let Some(c) = self.char_stream.next() {
seq.push(c);
self.advance()?;
if let Some(d1) = c.to_digit(16) {
out_val *= 16;
out_val += d1;
} else {
return Err((LERR::MalformedEscapeSequence, self.pos));
return Err((LERR::MalformedEscapeSequence(seq), self.pos));
}
} else {
return Err((LERR::MalformedEscapeSequence, self.pos));
return Err((LERR::MalformedEscapeSequence(seq), self.pos));
}
self.advance()?;
}
if let Some(r) = char::from_u32(out_val) {
result.push(r);
} else {
return Err((LERR::MalformedEscapeSequence, self.pos));
return Err((LERR::MalformedEscapeSequence(seq), self.pos));
}
}
'U' if escape => {
escape = false;
'U' if !escape.is_empty() => {
let mut seq = escape.clone();
seq.push('U');
escape.clear();
let mut out_val: u32 = 0;
for _ in 0..8 {
if let Some(c) = self.char_stream.next() {
seq.push(c);
self.advance()?;
if let Some(d1) = c.to_digit(16) {
out_val *= 16;
out_val += d1;
} else {
return Err((LERR::MalformedEscapeSequence, self.pos));
return Err((LERR::MalformedEscapeSequence(seq), self.pos));
}
} else {
return Err((LERR::MalformedEscapeSequence, self.pos));
return Err((LERR::MalformedEscapeSequence(seq), self.pos));
}
self.advance()?;
}
if let Some(r) = char::from_u32(out_val) {
result.push(r);
} else {
return Err((LERR::MalformedEscapeSequence, self.pos));
return Err((LERR::MalformedEscapeSequence(seq), self.pos));
}
}
x if enclosing_char == x && escape => result.push(x),
x if enclosing_char == x && !escape => break,
_ if escape => return Err((LERR::MalformedEscapeSequence, self.pos)),
x if enclosing_char == x && !escape.is_empty() => result.push(x),
x if enclosing_char == x && escape.is_empty() => break,
_ if !escape.is_empty() => {
return Err((LERR::MalformedEscapeSequence(escape), self.pos))
}
'\n' => {
self.rewind()?;
return Err((LERR::UnterminatedString, self.pos));
}
x => {
escape = false;
escape.clear();
result.push(x);
}
}
}
let out: String = result.iter().cloned().collect();
let out: String = result.iter().collect();
Ok(out)
}
@ -672,7 +691,7 @@ impl<'a> TokenIterator<'a> {
while let Some(&next_char) = self.char_stream.peek() {
match next_char {
'0'..='9' => {
'0'..='9' | '_' => {
result.push(next_char);
self.char_stream.next();
if let Err(err) = self.advance_token() {
@ -687,7 +706,7 @@ impl<'a> TokenIterator<'a> {
}
while let Some(&next_char_in_float) = self.char_stream.peek() {
match next_char_in_float {
'0'..='9' => {
'0'..='9' | '_' => {
result.push(next_char_in_float);
self.char_stream.next();
if let Err(err) = self.advance_token() {
@ -698,7 +717,7 @@ impl<'a> TokenIterator<'a> {
}
}
}
'x' | 'X' => {
'x' | 'X' if c == '0' => {
result.push(next_char);
self.char_stream.next();
if let Err(err) = self.advance_token() {
@ -706,7 +725,7 @@ impl<'a> TokenIterator<'a> {
}
while let Some(&next_char_in_hex) = self.char_stream.peek() {
match next_char_in_hex {
'0'..='9' | 'a'..='f' | 'A'..='F' => {
'0'..='9' | 'a'..='f' | 'A'..='F' | '_' => {
result.push(next_char_in_hex);
self.char_stream.next();
if let Err(err) = self.advance_token() {
@ -718,7 +737,7 @@ impl<'a> TokenIterator<'a> {
}
radix_base = Some(16);
}
'o' | 'O' => {
'o' | 'O' if c == '0' => {
result.push(next_char);
self.char_stream.next();
if let Err(err) = self.advance_token() {
@ -726,7 +745,7 @@ impl<'a> TokenIterator<'a> {
}
while let Some(&next_char_in_oct) = self.char_stream.peek() {
match next_char_in_oct {
'0'..='8' => {
'0'..='8' | '_' => {
result.push(next_char_in_oct);
self.char_stream.next();
if let Err(err) = self.advance_token() {
@ -738,7 +757,7 @@ impl<'a> TokenIterator<'a> {
}
radix_base = Some(8);
}
'b' | 'B' => {
'b' | 'B' if c == '0' => {
result.push(next_char);
self.char_stream.next();
if let Err(err) = self.advance_token() {
@ -763,19 +782,19 @@ impl<'a> TokenIterator<'a> {
}
if let Some(radix) = radix_base {
let out: String = result
.iter()
.cloned()
.skip(2)
.filter(|c| c != &'_')
.collect();
let out: String = result.iter().skip(2).filter(|&&c| c != '_').collect();
if let Ok(val) = i64::from_str_radix(&out, radix) {
return Some((Token::IntegerConstant(val), pos));
}
return Some((
if let Ok(val) = i64::from_str_radix(&out, radix) {
Token::IntegerConstant(val)
} else {
Token::LexErr(LERR::MalformedNumber(result.iter().collect()))
},
pos,
));
}
let out: String = result.iter().cloned().collect();
let out: String = result.iter().filter(|&&c| c != '_').collect();
return Some((
if let Ok(val) = out.parse::<i64>() {
@ -783,7 +802,7 @@ impl<'a> TokenIterator<'a> {
} else if let Ok(val) = out.parse::<f64>() {
Token::FloatConstant(val)
} else {
Token::LexErr(LERR::MalformedNumber)
Token::LexErr(LERR::MalformedNumber(result.iter().collect()))
},
pos,
));
@ -805,7 +824,7 @@ impl<'a> TokenIterator<'a> {
}
}
let out: String = result.iter().cloned().collect();
let out: String = result.iter().collect();
return Some((
match out.as_str() {
@ -840,12 +859,12 @@ impl<'a> TokenIterator<'a> {
return Some((
if let Some(first_char) = chars.next() {
if chars.count() != 0 {
Token::LexErr(LERR::MalformedChar)
Token::LexErr(LERR::MalformedChar(format!("'{}'", result)))
} else {
Token::CharConstant(first_char)
}
} else {
Token::LexErr(LERR::MalformedChar)
Token::LexErr(LERR::MalformedChar(format!("'{}'", result)))
},
pos,
));

View File

@ -5,13 +5,21 @@ fn test_string() -> Result<(), EvalAltResult> {
let mut engine = Engine::new();
assert_eq!(
engine.eval::<String>("\"Test string: \\u2764\"")?,
engine.eval::<String>(r#""Test string: \u2764""#)?,
"Test string: ❤".to_string()
);
assert_eq!(
engine.eval::<String>("\"foo\" + \"bar\"")?,
engine.eval::<String>(r#""Test string: \x58""#)?,
"Test string: X".to_string()
);
assert_eq!(
engine.eval::<String>(r#""foo" + "bar""#)?,
"foobar".to_string()
);
assert_eq!(
engine.eval::<String>(r#""foo" + 123.4556"#)?,
"foo123.4556".to_string()
);
Ok(())
}