Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(902)

Unified Diff: runtime/vm/regexp_parser.cc

Issue 754383002: Revert "Integrate the Irregexp Regular Expression Engine." (Closed) Base URL: https://dart.googlecode.com/svn/branches/bleeding_edge/dart
Patch Set: Created 6 years, 1 month ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « runtime/vm/regexp_parser.h ('k') | runtime/vm/symbols.h » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: runtime/vm/regexp_parser.cc
diff --git a/runtime/vm/regexp_parser.cc b/runtime/vm/regexp_parser.cc
index 79e802b2d092ff37f0fb0aa6a6fe3988f099ee2b..72e8d7ffcc992905cf0e5fe112624c7ba45ac535 100644
--- a/runtime/vm/regexp_parser.cc
+++ b/runtime/vm/regexp_parser.cc
@@ -2,23 +2,17 @@
// for details. All rights reserved. Use of this source code is governed by a
// BSD-style license that can be found in the LICENSE file.
-#include "vm/longjump.h"
-#include "vm/object_store.h"
#include "vm/regexp_parser.h"
-namespace dart {
-
-#define I isolate()
+// SNIP
-// Enables possessive quantifier syntax for testing.
-static const bool FLAG_regexp_possessive_quantifier = false;
+namespace dart {
-RegExpBuilder::RegExpBuilder()
- : isolate_(Isolate::Current()),
+RegExpBuilder::RegExpBuilder(Zone* zone)
+ : zone_(zone),
pending_empty_(false),
characters_(NULL),
terms_(),
- text_(),
alternatives_()
#ifdef DEBUG
, last_added_(ADD_NONE)
@@ -29,9 +23,9 @@ RegExpBuilder::RegExpBuilder()
void RegExpBuilder::FlushCharacters() {
pending_empty_ = false;
if (characters_ != NULL) {
- RegExpTree* atom = new(I) RegExpAtom(characters_);
+ RegExpTree* atom = new(zone()) RegExpAtom(characters_->ToConstVector());
characters_ = NULL;
- text_.Add(atom);
+ text_.Add(atom, zone());
LAST(ADD_ATOM);
}
}
@@ -39,27 +33,27 @@ void RegExpBuilder::FlushCharacters() {
void RegExpBuilder::FlushText() {
FlushCharacters();
- intptr_t num_text = text_.length();
+ int num_text = text_.length();
if (num_text == 0) {
return;
} else if (num_text == 1) {
- terms_.Add(text_.Last());
+ terms_.Add(text_.last(), zone());
} else {
- RegExpText* text = new(I) RegExpText();
- for (intptr_t i = 0; i < num_text; i++)
- text_[i]->AppendToText(text);
- terms_.Add(text);
+ RegExpText* text = new(zone()) RegExpText(zone());
+ for (int i = 0; i < num_text; i++)
+ text_.Get(i)->AppendToText(text, zone());
+ terms_.Add(text, zone());
}
text_.Clear();
}
-void RegExpBuilder::AddCharacter(uint16_t c) {
+void RegExpBuilder::AddCharacter(uc16 c) {
pending_empty_ = false;
if (characters_ == NULL) {
- characters_ = new(I) ZoneGrowableArray<uint16_t>(4);
+ characters_ = new(zone()) ZoneList<uc16>(4, zone());
}
- characters_->Add(c);
+ characters_->Add(c, zone());
LAST(ADD_CHAR);
}
@@ -76,10 +70,10 @@ void RegExpBuilder::AddAtom(RegExpTree* term) {
}
if (term->IsTextElement()) {
FlushCharacters();
- text_.Add(term);
+ text_.Add(term, zone());
} else {
FlushText();
- terms_.Add(term);
+ terms_.Add(term, zone());
}
LAST(ADD_ATOM);
}
@@ -87,7 +81,7 @@ void RegExpBuilder::AddAtom(RegExpTree* term) {
void RegExpBuilder::AddAssertion(RegExpTree* assert) {
FlushText();
- terms_.Add(assert);
+ terms_.Add(assert, zone());
LAST(ADD_ASSERT);
}
@@ -99,21 +93,16 @@ void RegExpBuilder::NewAlternative() {
void RegExpBuilder::FlushTerms() {
FlushText();
- intptr_t num_terms = terms_.length();
+ int num_terms = terms_.length();
RegExpTree* alternative;
if (num_terms == 0) {
alternative = RegExpEmpty::GetInstance();
} else if (num_terms == 1) {
- alternative = terms_.Last();
+ alternative = terms_.last();
} else {
- ZoneGrowableArray<RegExpTree*>* terms =
- new(I) ZoneGrowableArray<RegExpTree*>();
- for (intptr_t i = 0; i < terms_.length(); i++) {
- terms->Add(terms_[i]);
- }
- alternative = new(I) RegExpAlternative(terms);
+ alternative = new(zone()) RegExpAlternative(terms_.GetList(zone()));
}
- alternatives_.Add(alternative);
+ alternatives_.Add(alternative, zone());
terms_.Clear();
LAST(ADD_NONE);
}
@@ -121,59 +110,43 @@ void RegExpBuilder::FlushTerms() {
RegExpTree* RegExpBuilder::ToRegExp() {
FlushTerms();
- intptr_t num_alternatives = alternatives_.length();
+ int num_alternatives = alternatives_.length();
if (num_alternatives == 0) {
return RegExpEmpty::GetInstance();
}
if (num_alternatives == 1) {
- return alternatives_.Last();
- }
- ZoneGrowableArray<RegExpTree*>* alternatives =
- new(I) ZoneGrowableArray<RegExpTree*>();
- for (intptr_t i = 0; i < alternatives_.length(); i++) {
- alternatives->Add(alternatives_[i]);
+ return alternatives_.last();
}
- return new(I) RegExpDisjunction(alternatives);
+ return new(zone()) RegExpDisjunction(alternatives_.GetList(zone()));
}
void RegExpBuilder::AddQuantifierToAtom(
- intptr_t min,
- intptr_t max,
- RegExpQuantifier::QuantifierType quantifier_type) {
+ int min, int max, RegExpQuantifier::QuantifierType quantifier_type) {
if (pending_empty_) {
pending_empty_ = false;
return;
}
RegExpTree* atom;
if (characters_ != NULL) {
- DEBUG_ASSERT(last_added_ == ADD_CHAR);
+ DCHECK(last_added_ == ADD_CHAR);
// Last atom was character.
-
- ZoneGrowableArray<uint16_t> *char_vector =
- new(I) ZoneGrowableArray<uint16_t>();
- char_vector->AddArray(*characters_);
- intptr_t num_chars = char_vector->length();
+ Vector<const uc16> char_vector = characters_->ToConstVector();
+ int num_chars = char_vector.length();
if (num_chars > 1) {
- ZoneGrowableArray<uint16_t> *prefix =
- new(I) ZoneGrowableArray<uint16_t>();
- for (intptr_t i = 0; i < num_chars - 1; i++) {
- prefix->Add(char_vector->At(i));
- }
- text_.Add(new(I) RegExpAtom(prefix));
- ZoneGrowableArray<uint16_t> *tail = new(I) ZoneGrowableArray<uint16_t>();
- tail->Add(char_vector->At(num_chars - 1));
- char_vector = tail;
+ Vector<const uc16> prefix = char_vector.SubVector(0, num_chars - 1);
+ text_.Add(new(zone()) RegExpAtom(prefix), zone());
+ char_vector = char_vector.SubVector(num_chars - 1, num_chars);
}
characters_ = NULL;
- atom = new(I) RegExpAtom(char_vector);
+ atom = new(zone()) RegExpAtom(char_vector);
FlushText();
} else if (text_.length() > 0) {
- DEBUG_ASSERT(last_added_ == ADD_ATOM);
+ DCHECK(last_added_ == ADD_ATOM);
atom = text_.RemoveLast();
FlushText();
} else if (terms_.length() > 0) {
- DEBUG_ASSERT(last_added_ == ADD_ATOM);
+ DCHECK(last_added_ == ADD_ATOM);
atom = terms_.RemoveLast();
if (atom->max_match() == 0) {
// Guaranteed to only match an empty string.
@@ -181,7 +154,7 @@ void RegExpBuilder::AddQuantifierToAtom(
if (min == 0) {
return;
}
- terms_.Add(atom);
+ terms_.Add(atom, zone());
return;
}
} else {
@@ -189,17 +162,22 @@ void RegExpBuilder::AddQuantifierToAtom(
UNREACHABLE();
return;
}
- terms_.Add(new(I) RegExpQuantifier(min, max, quantifier_type, atom));
+ terms_.Add(
+ new(zone()) RegExpQuantifier(min, max, quantifier_type, atom), zone());
LAST(ADD_TERM);
}
-// ----------------------------------------------------------------------------
-// Implementation of Parser
+// SNIP
-RegExpParser::RegExpParser(const String& in,
- String* error,
- bool multiline)
- : isolate_(Isolate::Current()),
+// ----------------------------------------------------------------------------
+// Regular expressions
+
+RegExpParser::RegExpParser(FlatStringReader* in,
+ Handle<String>* error,
+ bool multiline,
+ Zone* zone)
+ : isolate_(zone->isolate()),
+ zone_(zone),
error_(error),
captures_(NULL),
in_(in),
@@ -216,35 +194,9 @@ RegExpParser::RegExpParser(const String& in,
}
-bool RegExpParser::ParseFunction(ParsedFunction *parsed_function) {
- Isolate* isolate = parsed_function->isolate();
- JSRegExp& regexp = JSRegExp::Handle(parsed_function->function().regexp());
-
- const String& pattern = String::Handle(regexp.pattern());
- const bool multiline = regexp.is_multi_line();
-
- RegExpCompileData* compile_data = new(isolate) RegExpCompileData();
- if (!RegExpParser::ParseRegExp(pattern, multiline, compile_data)) {
- // Parsing failures are handled in the JSRegExp factory constructor.
- UNREACHABLE();
- }
-
- regexp.set_num_bracket_expressions(compile_data->capture_count);
- if (compile_data->simple) {
- regexp.set_is_simple();
- } else {
- regexp.set_is_complex();
- }
-
- parsed_function->SetRegExpCompileData(compile_data);
-
- return true;
-}
-
-
-uint32_t RegExpParser::Next() {
+uc32 RegExpParser::Next() {
if (has_next()) {
- return in().CharAt(next_pos_);
+ return in()->Get(next_pos_);
} else {
return kEndMarker;
}
@@ -252,9 +204,16 @@ uint32_t RegExpParser::Next() {
void RegExpParser::Advance() {
- if (next_pos_ < in().Length()) {
- current_ = in().CharAt(next_pos_);
- next_pos_++;
+ if (next_pos_ < in()->length()) {
+ StackLimitCheck check(isolate());
+ if (check.HasOverflowed()) {
+ ReportError(CStrVector(Isolate::kStackOverflowMessage));
+ } else if (zone()->excess_allocation()) {
+ ReportError(CStrVector("Regular expression too large"));
+ } else {
+ current_ = in()->Get(next_pos_);
+ next_pos_++;
+ }
} else {
current_ = kEndMarker;
has_more_ = false;
@@ -262,14 +221,14 @@ void RegExpParser::Advance() {
}
-void RegExpParser::Reset(intptr_t pos) {
+void RegExpParser::Reset(int pos) {
next_pos_ = pos;
- has_more_ = (pos < in().Length());
+ has_more_ = (pos < in()->length());
Advance();
}
-void RegExpParser::Advance(intptr_t dist) {
+void RegExpParser::Advance(int dist) {
next_pos_ += dist - 1;
Advance();
}
@@ -280,27 +239,24 @@ bool RegExpParser::simple() {
}
-void RegExpParser::ReportError(const char* message) {
+RegExpTree* RegExpParser::ReportError(Vector<const char> message) {
failed_ = true;
- *error_ = String::New(message);
+ *error_ = isolate()->factory()->NewStringFromAscii(message).ToHandleChecked();
// Zip to the end to make sure the no more input is read.
current_ = kEndMarker;
- next_pos_ = in().Length();
-
- const Error& error = Error::Handle(LanguageError::New(*error_));
- Report::LongJump(error);
- UNREACHABLE();
+ next_pos_ = in()->length();
+ return NULL;
}
// Pattern ::
// Disjunction
RegExpTree* RegExpParser::ParsePattern() {
- RegExpTree* result = ParseDisjunction();
- ASSERT(!has_more());
+ RegExpTree* result = ParseDisjunction(CHECK_FAILED);
+ DCHECK(!has_more());
// If the result of parsing is a literal string atom, and it has the
// same length as the input, then the atom is identical to the input.
- if (result->IsAtom() && result->AsAtom()->length() == in().Length()) {
+ if (result->IsAtom() && result->AsAtom()->length() == in()->length()) {
simple_ = true;
}
return result;
@@ -319,7 +275,7 @@ RegExpTree* RegExpParser::ParsePattern() {
// Atom Quantifier
RegExpTree* RegExpParser::ParseDisjunction() {
// Used to store current state while parsing subexpressions.
- RegExpParserState initial_state(NULL, INITIAL, 0, I);
+ RegExpParserState initial_state(NULL, INITIAL, 0, zone());
RegExpParserState* stored_state = &initial_state;
// Cache the builder in a local variable for quick access.
RegExpBuilder* builder = initial_state.builder();
@@ -328,27 +284,25 @@ RegExpTree* RegExpParser::ParseDisjunction() {
case kEndMarker:
if (stored_state->IsSubexpression()) {
// Inside a parenthesized group when hitting end of input.
- ReportError("Unterminated group");
- UNREACHABLE();
+ ReportError(CStrVector("Unterminated group") CHECK_FAILED);
}
- ASSERT(INITIAL == stored_state->group_type());
+ DCHECK_EQ(INITIAL, stored_state->group_type());
// Parsing completed successfully.
return builder->ToRegExp();
case ')': {
if (!stored_state->IsSubexpression()) {
- ReportError("Unmatched ')'");
- UNREACHABLE();
+ ReportError(CStrVector("Unmatched ')'") CHECK_FAILED);
}
- ASSERT(INITIAL != stored_state->group_type());
+ DCHECK_NE(INITIAL, stored_state->group_type());
Advance();
// End disjunction parsing and convert builder content to new single
// regexp atom.
RegExpTree* body = builder->ToRegExp();
- intptr_t end_capture_index = captures_started();
+ int end_capture_index = captures_started();
- intptr_t capture_index = stored_state->capture_index();
+ int capture_index = stored_state->capture_index();
SubexpressionType group_type = stored_state->group_type();
// Restore previous state.
@@ -357,20 +311,20 @@ RegExpTree* RegExpParser::ParseDisjunction() {
// Build result of subexpression.
if (group_type == CAPTURE) {
- RegExpCapture* capture = new(I) RegExpCapture(body, capture_index);
- (*captures_)[capture_index - 1] = capture;
+ RegExpCapture* capture = new(zone()) RegExpCapture(body, capture_index);
+ captures_->at(capture_index - 1) = capture;
body = capture;
} else if (group_type != GROUPING) {
- ASSERT(group_type == POSITIVE_LOOKAHEAD ||
+ DCHECK(group_type == POSITIVE_LOOKAHEAD ||
group_type == NEGATIVE_LOOKAHEAD);
bool is_positive = (group_type == POSITIVE_LOOKAHEAD);
- body = new(I) RegExpLookahead(body,
- is_positive,
- end_capture_index - capture_index,
- capture_index);
+ body = new(zone()) RegExpLookahead(body,
+ is_positive,
+ end_capture_index - capture_index,
+ capture_index);
}
builder->AddAtom(body);
- // For compatibility with JSC and ES3, we allow quantifiers after
+ // For compatability with JSC and ES3, we allow quantifiers after
// lookaheads, and break in all cases.
break;
}
@@ -382,16 +336,15 @@ RegExpTree* RegExpParser::ParseDisjunction() {
case '*':
case '+':
case '?':
- ReportError("Nothing to repeat");
- UNREACHABLE();
+ return ReportError(CStrVector("Nothing to repeat"));
case '^': {
Advance();
if (multiline_) {
builder->AddAssertion(
- new(I) RegExpAssertion(RegExpAssertion::START_OF_LINE));
+ new(zone()) RegExpAssertion(RegExpAssertion::START_OF_LINE));
} else {
builder->AddAssertion(
- new(I) RegExpAssertion(RegExpAssertion::START_OF_INPUT));
+ new(zone()) RegExpAssertion(RegExpAssertion::START_OF_INPUT));
set_contains_anchor();
}
continue;
@@ -401,16 +354,16 @@ RegExpTree* RegExpParser::ParseDisjunction() {
RegExpAssertion::AssertionType assertion_type =
multiline_ ? RegExpAssertion::END_OF_LINE :
RegExpAssertion::END_OF_INPUT;
- builder->AddAssertion(new RegExpAssertion(assertion_type));
+ builder->AddAssertion(new(zone()) RegExpAssertion(assertion_type));
continue;
}
case '.': {
Advance();
// everything except \x0a, \x0d, \u2028 and \u2029
- ZoneGrowableArray<CharacterRange>* ranges =
- new ZoneGrowableArray<CharacterRange>(2);
- CharacterRange::AddClassEscape('.', ranges);
- RegExpTree* atom = new RegExpCharacterClass(ranges, false);
+ ZoneList<CharacterRange>* ranges =
+ new(zone()) ZoneList<CharacterRange>(2, zone());
+ CharacterRange::AddClassEscape('.', ranges, zone());
+ RegExpTree* atom = new(zone()) RegExpCharacterClass(ranges, false);
builder->AddAtom(atom);
break;
}
@@ -429,28 +382,27 @@ RegExpTree* RegExpParser::ParseDisjunction() {
subexpr_type = NEGATIVE_LOOKAHEAD;
break;
default:
- ReportError("Invalid group");
- UNREACHABLE();
+ ReportError(CStrVector("Invalid group") CHECK_FAILED);
+ break;
}
Advance(2);
} else {
if (captures_ == NULL) {
- captures_ = new ZoneGrowableArray<RegExpCapture*>(2);
+ captures_ = new(zone()) ZoneList<RegExpCapture*>(2, zone());
}
if (captures_started() >= kMaxCaptures) {
- ReportError("Too many captures");
- UNREACHABLE();
+ ReportError(CStrVector("Too many captures") CHECK_FAILED);
}
- captures_->Add(NULL);
+ captures_->Add(NULL, zone());
}
// Store current state and begin new disjunction parsing.
- stored_state = new RegExpParserState(stored_state, subexpr_type,
- captures_started(), I);
+ stored_state = new(zone()) RegExpParserState(stored_state, subexpr_type,
+ captures_started(), zone());
builder = stored_state->builder();
continue;
}
case '[': {
- RegExpTree* atom = ParseCharacterClass();
+ RegExpTree* atom = ParseCharacterClass(CHECK_FAILED);
builder->AddAtom(atom);
break;
}
@@ -459,17 +411,16 @@ RegExpTree* RegExpParser::ParseDisjunction() {
case '\\':
switch (Next()) {
case kEndMarker:
- ReportError("\\ at end of pattern");
- UNREACHABLE();
+ return ReportError(CStrVector("\\ at end of pattern"));
case 'b':
Advance(2);
builder->AddAssertion(
- new RegExpAssertion(RegExpAssertion::BOUNDARY));
+ new(zone()) RegExpAssertion(RegExpAssertion::BOUNDARY));
continue;
case 'B':
Advance(2);
builder->AddAssertion(
- new RegExpAssertion(RegExpAssertion::NON_BOUNDARY));
+ new(zone()) RegExpAssertion(RegExpAssertion::NON_BOUNDARY));
continue;
// AtomEscape ::
// CharacterClassEscape
@@ -477,32 +428,32 @@ RegExpTree* RegExpParser::ParseDisjunction() {
// CharacterClassEscape :: one of
// d D s S w W
case 'd': case 'D': case 's': case 'S': case 'w': case 'W': {
- uint32_t c = Next();
+ uc32 c = Next();
Advance(2);
- ZoneGrowableArray<CharacterRange>* ranges =
- new ZoneGrowableArray<CharacterRange>(2);
- CharacterRange::AddClassEscape(c, ranges);
- RegExpTree* atom = new RegExpCharacterClass(ranges, false);
+ ZoneList<CharacterRange>* ranges =
+ new(zone()) ZoneList<CharacterRange>(2, zone());
+ CharacterRange::AddClassEscape(c, ranges, zone());
+ RegExpTree* atom = new(zone()) RegExpCharacterClass(ranges, false);
builder->AddAtom(atom);
break;
}
case '1': case '2': case '3': case '4': case '5': case '6':
case '7': case '8': case '9': {
- intptr_t index = 0;
+ int index = 0;
if (ParseBackReferenceIndex(&index)) {
RegExpCapture* capture = NULL;
if (captures_ != NULL && index <= captures_->length()) {
- capture = captures_->At(index - 1);
+ capture = captures_->at(index - 1);
}
if (capture == NULL) {
builder->AddEmpty();
break;
}
- RegExpTree* atom = new RegExpBackReference(capture);
+ RegExpTree* atom = new(zone()) RegExpBackReference(capture);
builder->AddAtom(atom);
break;
}
- uint32_t first_digit = Next();
+ uc32 first_digit = Next();
if (first_digit == '8' || first_digit == '9') {
// Treat as identity escape
builder->AddCharacter(first_digit);
@@ -513,7 +464,7 @@ RegExpTree* RegExpParser::ParseDisjunction() {
// FALLTHROUGH
case '0': {
Advance();
- uint32_t octal = ParseOctalLiteral();
+ uc32 octal = ParseOctalLiteral();
builder->AddCharacter(octal);
break;
}
@@ -541,10 +492,10 @@ RegExpTree* RegExpParser::ParseDisjunction() {
break;
case 'c': {
Advance();
- uint32_t controlLetter = Next();
+ uc32 controlLetter = Next();
// Special case if it is an ASCII letter.
// Convert lower case letters to uppercase.
- uint32_t letter = controlLetter & ~('a' ^ 'A');
+ uc32 letter = controlLetter & ~('a' ^ 'A');
if (letter < 'A' || 'Z' < letter) {
// controlLetter is not in range 'A'-'Z' or 'a'-'z'.
// This is outside the specification. We match JSC in
@@ -559,7 +510,7 @@ RegExpTree* RegExpParser::ParseDisjunction() {
}
case 'x': {
Advance(2);
- uint32_t value;
+ uc32 value;
if (ParseHexEscape(2, &value)) {
builder->AddCharacter(value);
} else {
@@ -569,7 +520,7 @@ RegExpTree* RegExpParser::ParseDisjunction() {
}
case 'u': {
Advance(2);
- uint32_t value;
+ uc32 value;
if (ParseHexEscape(4, &value)) {
builder->AddCharacter(value);
} else {
@@ -585,10 +536,9 @@ RegExpTree* RegExpParser::ParseDisjunction() {
}
break;
case '{': {
- intptr_t dummy;
+ int dummy;
if (ParseIntervalQuantifier(&dummy, &dummy)) {
- ReportError("Nothing to repeat");
- UNREACHABLE();
+ ReportError(CStrVector("Nothing to repeat") CHECK_FAILED);
}
// fallthrough
}
@@ -598,8 +548,8 @@ RegExpTree* RegExpParser::ParseDisjunction() {
break;
} // end switch(current())
- intptr_t min;
- intptr_t max;
+ int min;
+ int max;
switch (current()) {
// QuantifierPrefix ::
// *
@@ -624,8 +574,8 @@ RegExpTree* RegExpParser::ParseDisjunction() {
case '{':
if (ParseIntervalQuantifier(&min, &max)) {
if (max < min) {
- ReportError("numbers out of order in {} quantifier.");
- UNREACHABLE();
+ ReportError(CStrVector("numbers out of order in {} quantifier.")
+ CHECK_FAILED);
}
break;
} else {
@@ -649,8 +599,8 @@ RegExpTree* RegExpParser::ParseDisjunction() {
#ifdef DEBUG
-// Currently only used in an ASSERT.
-static bool IsSpecialClassEscape(uint32_t c) {
+// Currently only used in an DCHECK.
+static bool IsSpecialClassEscape(uc32 c) {
switch (c) {
case 'd': case 'D':
case 's': case 'S':
@@ -671,9 +621,9 @@ static bool IsSpecialClassEscape(uint32_t c) {
// characters.
void RegExpParser::ScanForCaptures() {
// Start with captures started previous to current position
- intptr_t capture_count = captures_started();
+ int capture_count = captures_started();
// Add count of captures after this position.
- intptr_t n;
+ int n;
while ((n = current()) != kEndMarker) {
Advance();
switch (n) {
@@ -681,7 +631,7 @@ void RegExpParser::ScanForCaptures() {
Advance();
break;
case '[': {
- intptr_t c;
+ int c;
while ((c = current()) != kEndMarker) {
Advance();
if (c == '\\') {
@@ -702,21 +652,16 @@ void RegExpParser::ScanForCaptures() {
}
-static inline bool IsDecimalDigit(int32_t c) {
- return '0' <= c && c <= '9';
-}
-
-
-bool RegExpParser::ParseBackReferenceIndex(intptr_t* index_out) {
- ASSERT('\\' == current());
- ASSERT('1' <= Next() && Next() <= '9');
+bool RegExpParser::ParseBackReferenceIndex(int* index_out) {
+ DCHECK_EQ('\\', current());
+ DCHECK('1' <= Next() && Next() <= '9');
// Try to parse a decimal literal that is no greater than the total number
// of left capturing parentheses in the input.
- intptr_t start = position();
- intptr_t value = Next() - '0';
+ int start = position();
+ int value = Next() - '0';
Advance(2);
while (true) {
- uint32_t c = current();
+ uc32 c = current();
if (IsDecimalDigit(c)) {
value = 10 * value + (c - '0');
if (value > kMaxCaptures) {
@@ -730,7 +675,7 @@ bool RegExpParser::ParseBackReferenceIndex(intptr_t* index_out) {
}
if (value > captures_started()) {
if (!is_scanned_for_captures_) {
- intptr_t saved_position = position();
+ int saved_position = position();
ScanForCaptures();
Reset(saved_position);
}
@@ -751,18 +696,17 @@ bool RegExpParser::ParseBackReferenceIndex(intptr_t* index_out) {
//
// Returns true if parsing succeeds, and set the min_out and max_out
// values. Values are truncated to RegExpTree::kInfinity if they overflow.
-bool RegExpParser::ParseIntervalQuantifier(intptr_t* min_out,
- intptr_t* max_out) {
- ASSERT(current() == '{');
- intptr_t start = position();
+bool RegExpParser::ParseIntervalQuantifier(int* min_out, int* max_out) {
+ DCHECK_EQ(current(), '{');
+ int start = position();
Advance();
- intptr_t min = 0;
+ int min = 0;
if (!IsDecimalDigit(current())) {
Reset(start);
return false;
}
while (IsDecimalDigit(current())) {
- intptr_t next = current() - '0';
+ int next = current() - '0';
if (min > (RegExpTree::kInfinity - next) / 10) {
// Overflow. Skip past remaining decimal digits and return -1.
do {
@@ -774,7 +718,7 @@ bool RegExpParser::ParseIntervalQuantifier(intptr_t* min_out,
min = 10 * min + next;
Advance();
}
- intptr_t max = 0;
+ int max = 0;
if (current() == '}') {
max = min;
Advance();
@@ -785,7 +729,7 @@ bool RegExpParser::ParseIntervalQuantifier(intptr_t* min_out,
Advance();
} else {
while (IsDecimalDigit(current())) {
- intptr_t next = current() - '0';
+ int next = current() - '0';
if (max > (RegExpTree::kInfinity - next) / 10) {
do {
Advance();
@@ -812,11 +756,11 @@ bool RegExpParser::ParseIntervalQuantifier(intptr_t* min_out,
}
-uint32_t RegExpParser::ParseOctalLiteral() {
- ASSERT(('0' <= current() && current() <= '7') || current() == kEndMarker);
+uc32 RegExpParser::ParseOctalLiteral() {
+ DCHECK(('0' <= current() && current() <= '7') || current() == kEndMarker);
// For compatibility with some other browsers (not all), we parse
// up to three octal digits with a value below 256.
- uint32_t value = current() - '0';
+ uc32 value = current() - '0';
Advance();
if ('0' <= current() && current() <= '7') {
value = value * 8 + current() - '0';
@@ -830,24 +774,13 @@ uint32_t RegExpParser::ParseOctalLiteral() {
}
-// Returns the value (0 .. 15) of a hexadecimal character c.
-// If c is not a legal hexadecimal character, returns a value < 0.
-static inline intptr_t HexValue(uint32_t c) {
- c -= '0';
- if (static_cast<unsigned>(c) <= 9) return c;
- c = (c | 0x20) - ('a' - '0'); // detect 0x11..0x16 and 0x31..0x36.
- if (static_cast<unsigned>(c) <= 5) return c + 10;
- return -1;
-}
-
-
-bool RegExpParser::ParseHexEscape(intptr_t length, uint32_t *value) {
- intptr_t start = position();
- uint32_t val = 0;
+bool RegExpParser::ParseHexEscape(int length, uc32 *value) {
+ int start = position();
+ uc32 val = 0;
bool done = false;
- for (intptr_t i = 0; !done; i++) {
- uint32_t c = current();
- intptr_t d = HexValue(c);
+ for (int i = 0; !done; i++) {
+ uc32 c = current();
+ int d = HexValue(c);
if (d < 0) {
Reset(start);
return false;
@@ -863,9 +796,9 @@ bool RegExpParser::ParseHexEscape(intptr_t length, uint32_t *value) {
}
-uint32_t RegExpParser::ParseClassCharacterEscape() {
- ASSERT(current() == '\\');
- DEBUG_ASSERT(has_next() && !IsSpecialClassEscape(Next()));
+uc32 RegExpParser::ParseClassCharacterEscape() {
+ DCHECK(current() == '\\');
+ DCHECK(has_next() && !IsSpecialClassEscape(Next()));
Advance();
switch (current()) {
case 'b':
@@ -889,8 +822,8 @@ uint32_t RegExpParser::ParseClassCharacterEscape() {
Advance();
return '\v';
case 'c': {
- uint32_t controlLetter = Next();
- uint32_t letter = controlLetter & ~('A' ^ 'a');
+ uc32 controlLetter = Next();
+ uc32 letter = controlLetter & ~('A' ^ 'a');
// For compatibility with JSC, inside a character class
// we also accept digits and underscore as control characters.
if ((controlLetter >= '0' && controlLetter <= '9') ||
@@ -913,7 +846,7 @@ uint32_t RegExpParser::ParseClassCharacterEscape() {
return ParseOctalLiteral();
case 'x': {
Advance();
- uint32_t value;
+ uc32 value;
if (ParseHexEscape(2, &value)) {
return value;
}
@@ -923,7 +856,7 @@ uint32_t RegExpParser::ParseClassCharacterEscape() {
}
case 'u': {
Advance();
- uint32_t value;
+ uc32 value;
if (ParseHexEscape(4, &value)) {
return value;
}
@@ -935,7 +868,7 @@ uint32_t RegExpParser::ParseClassCharacterEscape() {
// Extended identity escape. We accept any character that hasn't
// been matched by a more specific case, not just the subset required
// by the ECMAScript specification.
- uint32_t result = current();
+ uc32 result = current();
Advance();
return result;
}
@@ -944,9 +877,9 @@ uint32_t RegExpParser::ParseClassCharacterEscape() {
}
-CharacterRange RegExpParser::ParseClassAtom(uint16_t* char_class) {
- ASSERT(0 == *char_class);
- uint32_t first = current();
+CharacterRange RegExpParser::ParseClassAtom(uc16* char_class) {
+ DCHECK_EQ(0, *char_class);
+ uc32 first = current();
if (first == '\\') {
switch (Next()) {
case 'w': case 'W': case 'd': case 'D': case 's': case 'S': {
@@ -955,10 +888,9 @@ CharacterRange RegExpParser::ParseClassAtom(uint16_t* char_class) {
return CharacterRange::Singleton(0); // Return dummy value.
}
case kEndMarker:
- ReportError("\\ at end of pattern");
- UNREACHABLE();
+ return ReportError(CStrVector("\\ at end of pattern"));
default:
- uint32_t c = ParseClassCharacterEscape();
+ uc32 c = ParseClassCharacterEscape(CHECK_FAILED);
return CharacterRange::Singleton(c);
}
} else {
@@ -968,18 +900,19 @@ CharacterRange RegExpParser::ParseClassAtom(uint16_t* char_class) {
}
-static const uint16_t kNoCharClass = 0;
+static const uc16 kNoCharClass = 0;
// Adds range or pre-defined character class to character ranges.
// If char_class is not kInvalidClass, it's interpreted as a class
// escape (i.e., 's' means whitespace, from '\s').
-static inline void AddRangeOrEscape(ZoneGrowableArray<CharacterRange>* ranges,
- uint16_t char_class,
- CharacterRange range) {
+static inline void AddRangeOrEscape(ZoneList<CharacterRange>* ranges,
+ uc16 char_class,
+ CharacterRange range,
+ Zone* zone) {
if (char_class != kNoCharClass) {
- CharacterRange::AddClassEscape(char_class, ranges);
+ CharacterRange::AddClassEscape(char_class, ranges, zone);
} else {
- ranges->Add(range);
+ ranges->Add(range, zone);
}
}
@@ -988,18 +921,18 @@ RegExpTree* RegExpParser::ParseCharacterClass() {
static const char* kUnterminated = "Unterminated character class";
static const char* kRangeOutOfOrder = "Range out of order in character class";
- ASSERT(current() == '[');
+ DCHECK_EQ(current(), '[');
Advance();
bool is_negated = false;
if (current() == '^') {
is_negated = true;
Advance();
}
- ZoneGrowableArray<CharacterRange>* ranges =
- new(I) ZoneGrowableArray<CharacterRange>(2);
+ ZoneList<CharacterRange>* ranges =
+ new(zone()) ZoneList<CharacterRange>(2, zone());
while (has_more() && current() != ']') {
- uint16_t char_class = kNoCharClass;
- CharacterRange first = ParseClassAtom(&char_class);
+ uc16 char_class = kNoCharClass;
+ CharacterRange first = ParseClassAtom(&char_class CHECK_FAILED);
if (current() == '-') {
Advance();
if (current() == kEndMarker) {
@@ -1007,72 +940,64 @@ RegExpTree* RegExpParser::ParseCharacterClass() {
// following code report an error.
break;
} else if (current() == ']') {
- AddRangeOrEscape(ranges, char_class, first);
- ranges->Add(CharacterRange::Singleton('-'));
+ AddRangeOrEscape(ranges, char_class, first, zone());
+ ranges->Add(CharacterRange::Singleton('-'), zone());
break;
}
- uint16_t char_class_2 = kNoCharClass;
- CharacterRange next = ParseClassAtom(&char_class_2);
+ uc16 char_class_2 = kNoCharClass;
+ CharacterRange next = ParseClassAtom(&char_class_2 CHECK_FAILED);
if (char_class != kNoCharClass || char_class_2 != kNoCharClass) {
// Either end is an escaped character class. Treat the '-' verbatim.
- AddRangeOrEscape(ranges, char_class, first);
- ranges->Add(CharacterRange::Singleton('-'));
- AddRangeOrEscape(ranges, char_class_2, next);
+ AddRangeOrEscape(ranges, char_class, first, zone());
+ ranges->Add(CharacterRange::Singleton('-'), zone());
+ AddRangeOrEscape(ranges, char_class_2, next, zone());
continue;
}
if (first.from() > next.to()) {
- ReportError(kRangeOutOfOrder);
- UNREACHABLE();
+ return ReportError(CStrVector(kRangeOutOfOrder) CHECK_FAILED);
}
- ranges->Add(CharacterRange::Range(first.from(), next.to()));
+ ranges->Add(CharacterRange::Range(first.from(), next.to()), zone());
} else {
- AddRangeOrEscape(ranges, char_class, first);
+ AddRangeOrEscape(ranges, char_class, first, zone());
}
}
if (!has_more()) {
- ReportError(kUnterminated);
- UNREACHABLE();
+ return ReportError(CStrVector(kUnterminated) CHECK_FAILED);
}
Advance();
if (ranges->length() == 0) {
- ranges->Add(CharacterRange::Everything());
+ ranges->Add(CharacterRange::Everything(), zone());
is_negated = !is_negated;
}
- return new(I) RegExpCharacterClass(ranges, is_negated);
+ return new(zone()) RegExpCharacterClass(ranges, is_negated);
}
// ----------------------------------------------------------------------------
// The Parser interface.
-bool RegExpParser::ParseRegExp(const String& input,
+bool RegExpParser::ParseRegExp(FlatStringReader* input,
bool multiline,
- RegExpCompileData* result) {
- ASSERT(result != NULL);
- LongJumpScope jump;
- RegExpParser parser(input, &result->error, multiline);
- if (setjmp(*jump.Set()) == 0) {
- RegExpTree* tree = parser.ParsePattern();
- ASSERT(tree != NULL);
- ASSERT(result->error.IsNull());
+ RegExpCompileData* result,
+ Zone* zone) {
+ DCHECK(result != NULL);
+ RegExpParser parser(input, &result->error, multiline, zone);
+ RegExpTree* tree = parser.ParsePattern();
+ if (parser.failed()) {
+ DCHECK(tree == NULL);
+ DCHECK(!result->error.is_null());
+ } else {
+ DCHECK(tree != NULL);
+ DCHECK(result->error.is_null());
result->tree = tree;
- intptr_t capture_count = parser.captures_started();
+ int capture_count = parser.captures_started();
result->simple = tree->IsAtom() && parser.simple() && capture_count == 0;
result->contains_anchor = parser.contains_anchor();
result->capture_count = capture_count;
- } else {
- ASSERT(!result->error.IsNull());
- Isolate::Current()->object_store()->clear_sticky_error();
-
- // Throw a FormatException on parsing failures.
- const String& message = String::Handle(
- String::Concat(result->error, input));
- const Array& args = Array::Handle(Array::New(1));
- args.SetAt(0, message);
-
- Exceptions::ThrowByType(Exceptions::kFormat, args);
}
return !parser.failed();
}
+// SNIP
+
} // namespace dart
« no previous file with comments | « runtime/vm/regexp_parser.h ('k') | runtime/vm/symbols.h » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698