github
/
boa
mirror of https://github.com/boa-dev/boa.git
1
0
Fork 0
Code Issues Releases Wiki Activity
Browse Source

Fix `GetSubstitution` (#2933)

pull/2939/head
Haled Odat 2 years ago committed by GitHub
parent
eb37614d70
commit
190eeb388b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 153 additions and 134 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Unified View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 185
      boa_engine/src/builtins/regexp/mod.rs
  2. 102
      boa_engine/src/builtins/string/mod.rs

185
boa_engine/src/builtins/regexp/mod.rs
Unescape View File

@ -1267,8 +1267,14 @@ impl RegExp {
args: &[JsValue], args: &[JsValue],
context: &mut Context<'_>, context: &mut Context<'_>,
) -> JsResult<JsValue> { ) -> JsResult<JsValue> {
// Helper enum.
enum CallableOrString<'a> {
FunctionalReplace(&'a JsObject),
ReplaceValue(JsString),
}
// 1. Let rx be the this value. // 1. Let rx be the this value.
// 2. If Type(rx) is not Object, throw a TypeError exception. // 2. If rx is not an Object, throw a TypeError exception.
let rx = this.as_object().ok_or_else(|| { let rx = this.as_object().ok_or_else(|| {
JsNativeError::typ().with_message( JsNativeError::typ().with_message(
"RegExp.prototype[Symbol.replace] method called on incompatible value", "RegExp.prototype[Symbol.replace] method called on incompatible value",
@ -1276,84 +1282,101 @@ impl RegExp {
})?; })?;
// 3. Let S be ? ToString(string). // 3. Let S be ? ToString(string).
let arg_str = args.get_or_undefined(0).to_string(context)?; let s = args.get_or_undefined(0).to_string(context)?;
// 4. Let lengthS be the number of code unit elements in S. // 4. Let lengthS be the length of S.
let length_arg_str = arg_str.len(); let length_s = s.len();
let replace_value = args.get_or_undefined(1);
// 5. Let functionalReplace be IsCallable(replaceValue). // 5. Let functionalReplace be IsCallable(replaceValue).
let replace_value = args.get_or_undefined(1).clone(); let functional_replace = replace_value.as_callable();
let replace = if let Some(f) = replace_value.as_callable() {
Ok(f) // 6. If functionalReplace is false, then
let replace_value = if let Some(callable) = functional_replace {
CallableOrString::FunctionalReplace(callable)
} else { } else {
// 6. If functionalReplace is false, then
// a. Set replaceValue to ? ToString(replaceValue). // a. Set replaceValue to ? ToString(replaceValue).
Err(replace_value.to_string(context)?) CallableOrString::ReplaceValue(replace_value.to_string(context)?)
}; };
// 7. Let global be ! ToBoolean(? Get(rx, "global")). // 7. Let flags be ? ToString(? Get(rx, "flags")).
let global = rx.get(utf16!("global"), context)?.to_boolean(); let flags = rx.get(utf16!("flags"), context)?.to_string(context)?;
// 8. If global is true, then // 8. If flags contains "g", let global be true. Otherwise, let global be false.
let mut unicode = false; let global = flags.as_slice().contains(&u16::from(b'g'));
if global {
// a. Let fullUnicode be ! ToBoolean(? Get(rx, "unicode")). // 9. If global is true, then
unicode = rx.get(utf16!("unicode"), context)?.to_boolean(); let full_unicode = if global {
// a. If flags contains "u", let fullUnicode be true. Otherwise, let fullUnicode be false.
let full_unicode = flags.contains(&u16::from(b'u'));
// b. Perform ? Set(rx, "lastIndex", +0𝔽, true). // b. Perform ? Set(rx, "lastIndex", +0𝔽, true).
rx.set(utf16!("lastIndex"), 0, true, context)?; rx.set(utf16!("lastIndex"), 0, true, context)?;
}
// 9. Let results be a new empty List. full_unicode
} else {
false
};
// 10. Let results be a new empty List.
let mut results = Vec::new(); let mut results = Vec::new();
// 10. Let done be false. // SKIPPED: 11. Let done be false.
// 11. Repeat, while done is false, //
// NOTE(HalidOdat): We don't keep track of `done`, we just break when done is true.
// 12. Repeat, while done is false,
loop { loop {
// a. Let result be ? RegExpExec(rx, S). // a. Let result be ? RegExpExec(rx, S).
let result = Self::abstract_exec(rx, arg_str.clone(), context)?; let result = Self::abstract_exec(rx, s.clone(), context)?;
// b. If result is null, set done to true. // b. If result is null, set done to true.
let Some(result) = result else {
// SKIPPED: 1. Set done to true.
break;
};
// c. Else, // c. Else,
if let Some(result) = result { // i. Append result to results.
// i. Append result to the end of results. results.push(result.clone());
results.push(result.clone());
// ii. If global is false, set done to true. // ii. If global is false, then
if !global {
// SKIPPED: 1. Set done to true.
break;
}
if !global { // iii. Else,
break; // 1. Let matchStr be ? ToString(? Get(result, "0")).
} let match_str = result.get(0, context)?.to_string(context)?;
// iii. Else,
// 1. Let matchStr be ? ToString(? Get(result, "0")).
let match_str = result.get(0, context)?.to_string(context)?;
// 2. If matchStr is the empty String, then // 2. If matchStr is the empty String, then
if match_str.is_empty() { if match_str.is_empty() {
// a. Let thisIndex be ℝ(? ToLength(? Get(rx, "lastIndex"))). // a. Let thisIndex be ℝ(? ToLength(? Get(rx, "lastIndex"))).
let this_index = rx.get(utf16!("lastIndex"), context)?.to_length(context)?; let this_index = rx.get(utf16!("lastIndex"), context)?.to_length(context)?;
// b. Let nextIndex be AdvanceStringIndex(S, thisIndex, fullUnicode). // b. Let nextIndex be AdvanceStringIndex(S, thisIndex, fullUnicode).
let next_index = advance_string_index(&arg_str, this_index, unicode); let next_index = advance_string_index(&s, this_index, full_unicode);
// c. Perform ? Set(rx, "lastIndex", 𝔽(nextIndex), true). // c. Perform ? Set(rx, "lastIndex", 𝔽(nextIndex), true).
rx.set(utf16!("lastIndex"), JsValue::new(next_index), true, context)?; rx.set(utf16!("lastIndex"), JsValue::new(next_index), true, context)?;
}
} else {
break;
} }
} }
// 12. Let accumulatedResult be the empty String. // 16. If nextSourcePosition ≥ lengthS, return accumulatedResult.
// 17. Return the string-concatenation of accumulatedResult and the substring of S from nextSourcePosition.
// 13. Let accumulatedResult be the empty String.
let mut accumulated_result = vec![]; let mut accumulated_result = vec![];
// 13. Let nextSourcePosition be 0. // 14. Let nextSourcePosition be 0.
let mut next_source_position = 0; let mut next_source_position = 0;
// 14. For each element result of results, do // 15. For each element result of results, do
for result in results { for result in results {
// a. Let resultLength be ? LengthOfArrayLike(result). // a. Let resultLength be ? LengthOfArrayLike(result).
let result_length = result.length_of_array_like(context)? as isize; let result_length = result.length_of_array_like(context)? as i64;
// b. Let nCaptures be max(resultLength - 1, 0). // b. Let nCaptures be max(resultLength - 1, 0).
let n_captures = std::cmp::max(result_length - 1, 0); let n_captures = std::cmp::max(result_length - 1, 0);
@ -1361,7 +1384,7 @@ impl RegExp {
// c. Let matched be ? ToString(? Get(result, "0")). // c. Let matched be ? ToString(? Get(result, "0")).
let matched = result.get(0, context)?.to_string(context)?; let matched = result.get(0, context)?.to_string(context)?;
// d. Let matchLength be the number of code units in matched. // d. Let matchLength be the length of matched.
let match_length = matched.len(); let match_length = matched.len();
// e. Let position be ? ToIntegerOrInfinity(? Get(result, "index")). // e. Let position be ? ToIntegerOrInfinity(? Get(result, "index")).
@ -1370,13 +1393,12 @@ impl RegExp {
.to_integer_or_infinity(context)?; .to_integer_or_infinity(context)?;
// f. Set position to the result of clamping position between 0 and lengthS. // f. Set position to the result of clamping position between 0 and lengthS.
//position = position. let position = position.clamp_finite(0, length_s as i64) as usize;
let position = position.clamp_finite(0, length_arg_str as i64) as usize;
// h. Let captures be a new empty List. // g. Let captures be a new empty List.
let mut captures = Vec::new(); let mut captures = Vec::new();
// g. Let n be 1. // h. Let n be 1.
// i. Repeat, while n ≤ nCaptures, // i. Repeat, while n ≤ nCaptures,
for n in 1..=n_captures { for n in 1..=n_captures {
// i. Let capN be ? Get(result, ! ToString(𝔽(n))). // i. Let capN be ? Get(result, ! ToString(𝔽(n))).
@ -1388,42 +1410,43 @@ impl RegExp {
cap_n = cap_n.to_string(context)?.into(); cap_n = cap_n.to_string(context)?.into();
} }
// iii. Append capN as the last element of captures. // iii. Append capN to captures.
captures.push(cap_n); captures.push(cap_n);
// iv. Set n to n + 1. // iv. NOTE: When n = 1, the preceding step puts the first element into captures (at index 0).
// More generally, the nth capture (the characters captured by the nth set of capturing parentheses)
// is at captures[n - 1].
//
// v. Set n to n + 1.
} }
// j. Let namedCaptures be ? Get(result, "groups"). // j. Let namedCaptures be ? Get(result, "groups").
let mut named_captures = result.get(utf16!("groups"), context)?; let mut named_captures = result.get(utf16!("groups"), context)?;
// k. If functionalReplace is true, then let replacement = match replace_value {
let replacement = match replace { // k. If functionalReplace is true, then
Ok(replace_fn) => { CallableOrString::FunctionalReplace(replace_value) => {
// i. Let replacerArgs be « matched ». // i. Let replacerArgs be the list-concatenation of « matched », captures, and « 𝔽(position), S ».
let mut replacer_args = vec![JsValue::new(matched)]; let mut replacer_args = vec![JsValue::new(matched)];
// ii. Append in List order the elements of captures to the end of the List replacerArgs.
replacer_args.extend(captures); replacer_args.extend(captures);
// iii. Append 𝔽(position) and S to replacerArgs.
replacer_args.push(position.into()); replacer_args.push(position.into());
replacer_args.push(arg_str.clone().into()); replacer_args.push(s.clone().into());
// iv. If namedCaptures is not undefined, then // ii. If namedCaptures is not undefined, then
if !named_captures.is_undefined() { if !named_captures.is_undefined() {
// 1. Append namedCaptures as the last element of replacerArgs. // 1. Append namedCaptures to replacerArgs.
replacer_args.push(named_captures); replacer_args.push(named_captures);
} }
// v. Let replValue be ? Call(replaceValue, undefined, replacerArgs). // iii. Let replValue be ? Call(replaceValue, undefined, replacerArgs).
// vi. Let replacement be ? ToString(replValue). let repl_value =
replace_fn replace_value.call(&JsValue::undefined(), &replacer_args, context)?;
.call(&JsValue::undefined(), &replacer_args, context)?
.to_string(context)? // iv. Let replacement be ? ToString(replValue).
repl_value.to_string(context)?
} }
// l. Else, // l. Else,
Err(ref replace_str) => { CallableOrString::ReplaceValue(ref replace_value) => {
// i. If namedCaptures is not undefined, then // i. If namedCaptures is not undefined, then
if !named_captures.is_undefined() { if !named_captures.is_undefined() {
// 1. Set namedCaptures to ? ToObject(namedCaptures). // 1. Set namedCaptures to ? ToObject(namedCaptures).
@ -1433,11 +1456,11 @@ impl RegExp {
// ii. Let replacement be ? GetSubstitution(matched, S, position, captures, namedCaptures, replaceValue). // ii. Let replacement be ? GetSubstitution(matched, S, position, captures, namedCaptures, replaceValue).
string::get_substitution( string::get_substitution(
&matched, &matched,
&arg_str, &s,
position, position,
&captures, &captures,
&named_captures, &named_captures,
replace_str, replace_value,
context, context,
)? )?
} }
@ -1446,12 +1469,12 @@ impl RegExp {
// m. If position ≥ nextSourcePosition, then // m. If position ≥ nextSourcePosition, then
if position >= next_source_position { if position >= next_source_position {
// i. NOTE: position should not normally move backwards. // i. NOTE: position should not normally move backwards.
// If it does, it is an indication of an ill-behaving RegExp subclass // If it does, it is an indication of an ill-behaving RegExp subclass or use of
// or use of an access triggered side-effect to change the global flag or other characteristics of rx. // an access triggered side-effect to change the global flag or other characteristics of rx.
// In such cases, the corresponding substitution is ignored. // In such cases, the corresponding substitution is ignored.
// ii. Set accumulatedResult to the string-concatenation of accumulatedResult,
// the substring of S from nextSourcePosition to position, and replacement. // ii. Set accumulatedResult to the string-concatenation of accumulatedResult, the substring of S from nextSourcePosition to position, and replacement.
accumulated_result.extend_from_slice(&arg_str[next_source_position..position]); accumulated_result.extend_from_slice(&s[next_source_position..position]);
accumulated_result.extend_from_slice(&replacement); accumulated_result.extend_from_slice(&replacement);
// iii. Set nextSourcePosition to position + matchLength. // iii. Set nextSourcePosition to position + matchLength.
@ -1459,13 +1482,13 @@ impl RegExp {
} }
} }
// 15. If nextSourcePosition ≥ lengthS, return accumulatedResult. // 16. If nextSourcePosition ≥ lengthS, return accumulatedResult.
if next_source_position >= length_arg_str { if next_source_position >= length_s {
return Ok(js_string!(accumulated_result).into()); return Ok(js_string!(accumulated_result).into());
} }
// 16. Return the string-concatenation of accumulatedResult and the substring of S from nextSourcePosition. // 17. Return the string-concatenation of accumulatedResult and the substring of S from nextSourcePosition.
Ok(js_string!(&accumulated_result[..], &arg_str[next_source_position..]).into()) Ok(js_string!(&accumulated_result[..], &s[next_source_position..]).into())
} }
/// `RegExp.prototype[ @@search ]( string )` /// `RegExp.prototype[ @@search ]( string )`

102
boa_engine/src/builtins/string/mod.rs
Unescape View File

@ -2599,9 +2599,6 @@ pub(crate) fn get_substitution(
// 8. Let tailPos be position + matchLength. // 8. Let tailPos be position + matchLength.
let tail_pos = position + match_length; let tail_pos = position + match_length;
// 9. Let m be the number of elements in captures.
let m = captures.len();
// 10. Let result be the String value derived from replacement by copying code unit elements // 10. Let result be the String value derived from replacement by copying code unit elements
// from replacement to result while performing replacements as specified in Table 58. // from replacement to result while performing replacements as specified in Table 58.
// These $ replacements are done left-to-right, and, once such a replacement is performed, // These $ replacements are done left-to-right, and, once such a replacement is performed,
@ -2616,35 +2613,25 @@ pub(crate) fn get_substitution(
.and_then(CodePoint::as_char) .and_then(CodePoint::as_char)
.as_ref() .as_ref()
.map_or(false, char::is_ascii_digit); .map_or(false, char::is_ascii_digit);
// we use peek so that it is still in the iterator if not used
let third = if second_is_digit {
chars.peek().copied()
} else {
None
};
let third_is_digit = third
.and_then(CodePoint::as_char)
.as_ref()
.map_or(false, char::is_ascii_digit);
match (second, third) { match second {
// $$ // $$
(Some(CodePoint::Unicode('$')), _) => { Some(CodePoint::Unicode('$')) => {
// $ // $
result.push('$' as u16); result.push('$' as u16);
} }
// $& // $&
(Some(CodePoint::Unicode('&')), _) => { Some(CodePoint::Unicode('&')) => {
// matched // matched
result.extend_from_slice(matched); result.extend_from_slice(matched);
} }
// $` // $`
(Some(CodePoint::Unicode('`')), _) => { Some(CodePoint::Unicode('`')) => {
// The replacement is the substring of str from 0 to position. // The replacement is the substring of str from 0 to position.
result.extend_from_slice(&str[..position]); result.extend_from_slice(&str[..position]);
} }
// $' // $'
(Some(CodePoint::Unicode('\'')), _) => { Some(CodePoint::Unicode('\'')) => {
// If tailPos ≥ stringLength, the replacement is the empty String. // If tailPos ≥ stringLength, the replacement is the empty String.
// Otherwise the replacement is the substring of str from tailPos. // Otherwise the replacement is the substring of str from tailPos.
if tail_pos < str_length { if tail_pos < str_length {
@ -2652,52 +2639,61 @@ pub(crate) fn get_substitution(
} }
} }
// $nn // $nn
(Some(CodePoint::Unicode(second)), Some(CodePoint::Unicode(third))) // f. Else if templateRemainder starts with "$" followed by 1 or more decimal digits, then
if second_is_digit && third_is_digit => Some(CodePoint::Unicode(second)) if second_is_digit => {
{ // i. If templateRemainder starts with "$" followed by 2 or more decimal digits, let digitCount be 2. Otherwise, let digitCount be 1.
// The nnth element of captures, where nn is a two-digit decimal number in the range 01 to 99. // ii. Let ref be the substring of templateRemainder from 0 to 1 + digitCount.
let tens = second // iii. Let digits be the substring of templateRemainder from 1 to 1 + digitCount.
.to_digit(10) // iv. Let index be ℝ(StringToNumber(digits)).
.expect("could not convert character to digit after checking it") let mut index = second
as usize;
let units = third
.to_digit(10) .to_digit(10)
.expect("could not convert character to digit after checking it") .expect("could not convert character to digit after checking it")
as usize; as usize;
let nn = 10 * tens + units;
// vi. Let captureLen be the number of elements in captures.
// If nn ≤ m and the nnth element of captures is undefined, use the empty String instead. let capture_len = captures.len();
// If nn is 00 or nn > m, no replacement is done.
if nn == 0 || nn > m { // NOTE(HalidOdat): We deviate from the spec, because of a bug in GetSubstitutions
result.extend_from_slice(&['$' as u16, second as u16, third as u16]); //
} else if let Some(capture) = captures.get(nn - 1) { // See: https://github.com/tc39/ecma262/issues/1426
if let Some(s) = capture.as_string() { if let Some(digit) = chars
result.extend_from_slice(s); .peek()
.copied()
.and_then(CodePoint::as_char)
.and_then(|n| n.to_digit(10))
{
// If there is two digits, and it's not in range fallback to one digit.
let two_digit_index = index * 10 + digit as usize;
if (1..=capture_len).contains(&two_digit_index) {
index = two_digit_index;
chars.next();
} }
} }
chars.next(); // v. Assert: 0 ≤ index ≤ 99.
} debug_assert!((0..=99).contains(&index));
// $n
(Some(CodePoint::Unicode(second)), _) if second_is_digit => { // vii. If 1 ≤ index ≤ captureLen, then
// The nth element of captures, where n is a single digit in the range 1 to 9. if (1..=capture_len).contains(&index) {
let n = second // 1. Let capture be captures[index - 1].
.to_digit(10) // 2. If capture is undefined, then
.expect("could not convert character to digit after checking it") // a. Let refReplacement be the empty String.
as usize; // 3. Else,
// a. Let refReplacement be capture.
if let Some(capture) = captures.get(index - 1) {
if let Some(s) = capture.as_string() {
result.extend_from_slice(s);
}
}
// If n ≤ m and the nth element of captures is undefined, use the empty String instead. // viii. Else,
// If n > m, no replacement is done. } else {
if n == 0 || n > m { // 1. Let refReplacement be ref.
result.extend_from_slice(&['$' as u16, second as u16]); result.extend_from_slice(&['$' as u16, second as u16]);
} else if let Some(capture) = captures.get(n - 1) {
if let Some(s) = capture.as_string() {
result.extend_from_slice(s);
}
} }
} }
// $< // $<
(Some(CodePoint::Unicode('<')), _) => { Some(CodePoint::Unicode('<')) => {
// 1. If namedCaptures is undefined, the replacement text is the String "$<". // 1. If namedCaptures is undefined, the replacement text is the String "$<".
// 2. Else, // 2. Else,
if named_captures.is_undefined() { if named_captures.is_undefined() {

Write Preview
Loading…
Cancel
Save