Finish implementing char validation

This commit is contained in:
Adolfo Ochagavía 2018-11-06 17:05:06 +01:00
parent d1b242262a
commit c56db92d1f
3 changed files with 93 additions and 8 deletions

View File

@ -219,7 +219,7 @@ mod tests {
#[test]
fn test_unicode_escapes() {
let unicode_escapes = &[r"{DEAD}", "{BEEF}", "{FF}", ""];
let unicode_escapes = &[r"{DEAD}", "{BEEF}", "{FF}", "{}", ""];
for escape in unicode_escapes {
let escape_sequence = format!(r"'\u{}'", escape);
let component = closed_char_component(&escape_sequence);

View File

@ -1,3 +1,5 @@
use std::u32;
use crate::{
algo::visit::{visitor_ctx, VisitorCtx},
ast::{self, AstNode},
@ -42,15 +44,82 @@ fn validate_char(node: ast::Char, errors: &mut Vec<SyntaxError>) {
}
}
AsciiCodeEscape => {
// TODO:
// * First digit is octal
// * Second digit is hex
// An AsciiCodeEscape has 4 chars, example: `\xDD`
if text.len() < 4 {
errors.push(SyntaxError::new(TooShortAsciiCodeEscape, range));
} else {
assert!(text.chars().count() == 4, "AsciiCodeEscape cannot be longer than 4 chars");
match u8::from_str_radix(&text[2..], 16) {
Ok(code) if code < 128 => { /* Escape code is valid */ },
Ok(_) => errors.push(SyntaxError::new(AsciiCodeEscapeOutOfRange, range)),
Err(_) => errors.push(SyntaxError::new(MalformedAsciiCodeEscape, range)),
}
}
}
UnicodeEscape => {
// TODO:
// * Only hex digits or underscores allowed
// * Max 6 chars
// * Within allowed range (must be at most 10FFFF)
assert!(&text[..2] == "\\u", "UnicodeEscape always starts with \\u");
if text.len() == 2 {
// No starting `{`
errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
return;
}
if text.len() == 3 {
// Only starting `{`
errors.push(SyntaxError::new(UnclosedUnicodeEscape, range));
return;
}
let mut code = String::new();
let mut closed = false;
for c in text[3..].chars() {
assert!(!closed, "no characters after escape is closed");
if c.is_digit(16) {
code.push(c);
} else if c == '_' {
// Reject leading _
if code.len() == 0 {
errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
return;
}
} else if c == '}' {
closed = true;
} else {
errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
return;
}
}
if !closed {
errors.push(SyntaxError::new(UnclosedUnicodeEscape, range))
}
if code.len() == 0 {
errors.push(SyntaxError::new(EmptyUnicodeEcape, range));
return;
}
if code.len() > 6 {
errors.push(SyntaxError::new(OverlongUnicodeEscape, range));
}
match u32::from_str_radix(&code, 16) {
Ok(code_u32) if code_u32 > 0x10FFFF => {
errors.push(SyntaxError::new(UnicodeEscapeOutOfRange, range));
}
Ok(_) => {
// Valid escape code
}
Err(_) => {
errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
}
}
// FIXME: we really need tests for this
}
// Code points are always valid
CodePoint => (),

View File

@ -69,6 +69,14 @@ pub enum SyntaxErrorKind {
LongChar,
EmptyAsciiEscape,
InvalidAsciiEscape,
TooShortAsciiCodeEscape,
AsciiCodeEscapeOutOfRange,
MalformedAsciiCodeEscape,
UnclosedUnicodeEscape,
MalformedUnicodeEscape,
EmptyUnicodeEcape,
OverlongUnicodeEscape,
UnicodeEscapeOutOfRange,
}
#[derive(Debug, Clone, PartialEq, Eq, Hash)]
@ -83,6 +91,14 @@ impl fmt::Display for SyntaxErrorKind {
EmptyChar => write!(f, "Empty char literal"),
UnclosedChar => write!(f, "Unclosed char literal"),
LongChar => write!(f, "Char literal should be one character long"),
TooShortAsciiCodeEscape => write!(f, "Escape sequence should have two digits"),
AsciiCodeEscapeOutOfRange => write!(f, "Escape sequence should be between \\x00 and \\x7F"),
MalformedAsciiCodeEscape => write!(f, "Escape sequence should be a hexadecimal number"),
UnclosedUnicodeEscape => write!(f, "Missing `}}`"),
MalformedUnicodeEscape => write!(f, "Malformed unicode escape sequence"),
EmptyUnicodeEcape => write!(f, "Empty unicode escape sequence"),
OverlongUnicodeEscape => write!(f, "Unicode escape sequence should have at most 6 digits"),
UnicodeEscapeOutOfRange => write!(f, "Unicode escape code should be at most 0x10FFFF"),
ParseError(msg) => write!(f, "{}", msg.0),
}
}