Skip to content

V
V8
Commit 4213b46b authored by christian.plesner.hansen@gmail.com's avatar christian.plesner.hansen@gmail.com
Browse files
Options

Fixed some assertion propagation issues and added non-multiline $ 

propagation.


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@856 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
parent ef59fde8
Hide whitespace changes
Inline Side-by-side
Showing with 145 additions and 88 deletions
src/jsregexp.cc
View file @ 4213b46b
...@@ -218,7 +218,7 @@ Handle<Object> RegExpImpl::Compile(Handle<JSRegExp> re, ...@@ -218,7 +218,7 @@ Handle<Object> RegExpImpl::Compile(Handle<JSRegExp> re,
FlattenString(pattern); FlattenString(pattern);
RegExpParseResult parse_result; RegExpParseResult parse_result;
FlatStringReader reader(pattern); FlatStringReader reader(pattern);
if (!ParseRegExp(&reader, &parse_result)) { if (!ParseRegExp(&reader, flags.is_multiline(), &parse_result)) {
// Throw an exception if we fail to parse the pattern. // Throw an exception if we fail to parse the pattern.
ThrowRegExpException(re, ThrowRegExpException(re,
pattern, pattern,
...@@ -241,7 +241,8 @@ Handle<Object> RegExpImpl::Compile(Handle<JSRegExp> re, ...@@ -241,7 +241,8 @@ Handle<Object> RegExpImpl::Compile(Handle<JSRegExp> re,
Handle<FixedArray> irregexp_data = Handle<FixedArray> irregexp_data =
RegExpEngine::Compile(&parse_result, RegExpEngine::Compile(&parse_result,
&node, &node,
flags.is_ignore_case()); flags.is_ignore_case(),
flags.is_multiline());
if (irregexp_data.is_null()) { if (irregexp_data.is_null()) {
if (FLAG_disable_jscre) { if (FLAG_disable_jscre) {
UNIMPLEMENTED(); UNIMPLEMENTED();
...@@ -980,7 +981,8 @@ bool RegExpNode::GoTo(RegExpCompiler* compiler) { ...@@ -980,7 +981,8 @@ bool RegExpNode::GoTo(RegExpCompiler* compiler) {
// TODO(erikcorry): Implement support. // TODO(erikcorry): Implement support.
if (info_.follows_word_interest || if (info_.follows_word_interest ||
info_.follows_newline_interest || info_.follows_newline_interest ||
info_.follows_start_interest) { info_.follows_start_interest ||
info_.at_end) {
return false; return false;
} }
if (label_.is_bound()) { if (label_.is_bound()) {
...@@ -1004,7 +1006,8 @@ bool RegExpNode::GoTo(RegExpCompiler* compiler) { ...@@ -1004,7 +1006,8 @@ bool RegExpNode::GoTo(RegExpCompiler* compiler) {
bool EndNode::GoTo(RegExpCompiler* compiler) { bool EndNode::GoTo(RegExpCompiler* compiler) {
if (info()->follows_word_interest || if (info()->follows_word_interest ||
info()->follows_newline_interest || info()->follows_newline_interest ||
info()->follows_start_interest) { info()->follows_start_interest ||
info()->at_end) {
return false; return false;
} }
if (!label()->is_bound()) { if (!label()->is_bound()) {
...@@ -1629,14 +1632,21 @@ void DotPrinter::PrintAttributes(RegExpNode* that) { ...@@ -1629,14 +1632,21 @@ void DotPrinter::PrintAttributes(RegExpNode* that) {
"fontcolor=lightgrey, margin=0.1, fontsize=10, label=\"{", "fontcolor=lightgrey, margin=0.1, fontsize=10, label=\"{",
that); that);
NodeInfo* info = that->info(); NodeInfo* info = that->info();
stream()->Add("{NI|%i}|{SI|%i}|{WI|%i}", stream()->Add("{NI|%i}|{WI|%i}|{SI|%i}",
info->follows_newline_interest, info->follows_newline_interest,
info->follows_start_interest, info->follows_word_interest,
info->follows_word_interest); info->follows_start_interest);
stream()->Add("|{DN|%i}|{DS|%i}|{DW|%i}", stream()->Add("|{DN|%i}|{DW|%i}|{DS|%i}|{AE|%i}",
info->determine_newline, info->determine_newline,
info->determine_word,
info->determine_start, info->determine_start,
info->determine_word); info->at_end);
if (info->follows_newline != NodeInfo::UNKNOWN)
stream()->Add("|{FN|%i}", info->follows_newline);
if (info->follows_word != NodeInfo::UNKNOWN)
stream()->Add("|{FW|%i}", info->follows_word);
if (info->follows_start != NodeInfo::UNKNOWN)
stream()->Add("|{FS|%i}", info->follows_start);
Label* label = that->label(); Label* label = that->label();
if (label->is_bound()) if (label->is_bound())
stream()->Add("|{@|%x}", label->pos()); stream()->Add("|{@|%x}", label->pos());
...@@ -1924,11 +1934,14 @@ RegExpNode* RegExpAssertion::ToNode(RegExpCompiler* compiler, ...@@ -1924,11 +1934,14 @@ RegExpNode* RegExpAssertion::ToNode(RegExpCompiler* compiler,
case BOUNDARY: case NON_BOUNDARY: case BOUNDARY: case NON_BOUNDARY:
info.follows_word_interest = true; info.follows_word_interest = true;
break; break;
case END_OF_LINE: case END_OF_INPUT: case END_OF_INPUT:
info.at_end = true;
break;
case END_OF_LINE:
// This is wrong but has the effect of making the compiler abort. // This is wrong but has the effect of making the compiler abort.
info.follows_start_interest = true; info.at_end = true;
} }
return on_success->PropagateInterest(&info); return on_success->PropagateForward(&info);
} }
...@@ -2216,7 +2229,7 @@ void CharacterRange::AddCaseEquivalents(ZoneList<CharacterRange>* ranges) { ...@@ -2216,7 +2229,7 @@ void CharacterRange::AddCaseEquivalents(ZoneList<CharacterRange>* ranges) {
RegExpNode* RegExpNode::GetSibling(NodeInfo* info) { RegExpNode* RegExpNode::GetSibling(NodeInfo* info) {
for (int i = 0; i < siblings_.length(); i++) { for (int i = 0; i < siblings_.length(); i++) {
RegExpNode* sibling = siblings_.Get(i); RegExpNode* sibling = siblings_.Get(i);
if (sibling->info()->SameInterests(info)) if (sibling->info()->HasSameForwardInterests(info))
return sibling; return sibling;
} }
return NULL; return NULL;
...@@ -2225,58 +2238,64 @@ RegExpNode* RegExpNode::GetSibling(NodeInfo* info) { ...@@ -2225,58 +2238,64 @@ RegExpNode* RegExpNode::GetSibling(NodeInfo* info) {
template <class C> template <class C>
static RegExpNode* PropagateToEndpoint(C* node, NodeInfo* info) { static RegExpNode* PropagateToEndpoint(C* node, NodeInfo* info) {
RegExpNode* sibling = node->GetSibling(info); NodeInfo full_info(*node->info());
full_info.AddFromPreceding(info);
RegExpNode* sibling = node->GetSibling(&full_info);
if (sibling != NULL) return sibling; if (sibling != NULL) return sibling;
node->EnsureSiblings(); node->EnsureSiblings();
sibling = new C(*node); sibling = new C(*node);
sibling->info()->AdoptInterests(info); sibling->info()->AddFromPreceding(&full_info);
node->AddSibling(sibling); node->AddSibling(sibling);
return sibling; return sibling;
} }
RegExpNode* ActionNode::PropagateInterest(NodeInfo* info) { RegExpNode* ActionNode::PropagateForward(NodeInfo* info) {
RegExpNode* sibling = GetSibling(info); NodeInfo full_info(*this->info());
full_info.AddFromPreceding(info);
RegExpNode* sibling = GetSibling(&full_info);
if (sibling != NULL) return sibling; if (sibling != NULL) return sibling;
EnsureSiblings(); EnsureSiblings();
ActionNode* action = new ActionNode(*this); ActionNode* action = new ActionNode(*this);
action->info()->AdoptInterests(info); action->info()->AddFromPreceding(&full_info);
AddSibling(action); AddSibling(action);
action->set_on_success(action->on_success()->PropagateInterest(info)); action->set_on_success(action->on_success()->PropagateForward(info));
return action; return action;
} }
RegExpNode* ChoiceNode::PropagateInterest(NodeInfo* info) { RegExpNode* ChoiceNode::PropagateForward(NodeInfo* info) {
RegExpNode* sibling = GetSibling(info); NodeInfo full_info(*this->info());
full_info.AddFromPreceding(info);
RegExpNode* sibling = GetSibling(&full_info);
if (sibling != NULL) return sibling; if (sibling != NULL) return sibling;
EnsureSiblings(); EnsureSiblings();
ChoiceNode* choice = new ChoiceNode(*this); ChoiceNode* choice = new ChoiceNode(*this);
choice->info()->AdoptInterests(info); choice->info()->AddFromPreceding(&full_info);
AddSibling(choice); AddSibling(choice);
ZoneList<GuardedAlternative>* old_alternatives = alternatives(); ZoneList<GuardedAlternative>* old_alternatives = alternatives();
int count = old_alternatives->length(); int count = old_alternatives->length();
choice->alternatives_ = new ZoneList<GuardedAlternative>(count); choice->alternatives_ = new ZoneList<GuardedAlternative>(count);
for (int i = 0; i < count; i++) { for (int i = 0; i < count; i++) {
GuardedAlternative alternative = old_alternatives->at(i); GuardedAlternative alternative = old_alternatives->at(i);
alternative.set_node(alternative.node()->PropagateInterest(info)); alternative.set_node(alternative.node()->PropagateForward(info));
choice->alternatives()->Add(alternative); choice->alternatives()->Add(alternative);
} }
return choice; return choice;
} }
RegExpNode* EndNode::PropagateInterest(NodeInfo* info) { RegExpNode* EndNode::PropagateForward(NodeInfo* info) {
return PropagateToEndpoint(this, info); return PropagateToEndpoint(this, info);
} }
RegExpNode* BackReferenceNode::PropagateInterest(NodeInfo* info) { RegExpNode* BackReferenceNode::PropagateForward(NodeInfo* info) {
return PropagateToEndpoint(this, info); return PropagateToEndpoint(this, info);
} }
RegExpNode* TextNode::PropagateInterest(NodeInfo* info) { RegExpNode* TextNode::PropagateForward(NodeInfo* info) {
return PropagateToEndpoint(this, info); return PropagateToEndpoint(this, info);
} }
...@@ -2455,15 +2474,21 @@ void Analysis::VisitText(TextNode* that) { ...@@ -2455,15 +2474,21 @@ void Analysis::VisitText(TextNode* that) {
} }
EnsureAnalyzed(that->on_success()); EnsureAnalyzed(that->on_success());
EnsureAnalyzed(that->on_failure()); EnsureAnalyzed(that->on_failure());
NodeInfo* info = that->info();
NodeInfo* next_info = that->on_success()->info();
// If the following node is interested in what it follows then this
// node must determine it.
info->determine_newline = next_info->follows_newline_interest;
info->determine_word = next_info->follows_word_interest;
info->determine_start = next_info->follows_start_interest;
} }
void Analysis::VisitAction(ActionNode* that) { void Analysis::VisitAction(ActionNode* that) {
RegExpNode* next = that->on_success(); EnsureAnalyzed(that->on_success());
EnsureAnalyzed(next); // If the next node is interested in what it follows then this node
that->info()->determine_newline = next->info()->prev_determine_newline(); // has to be interested too so it can pass the information on.
that->info()->determine_word = next->info()->prev_determine_word(); that->info()->AddFromFollowing(that->on_success()->info());
that->info()->determine_start = next->info()->prev_determine_start();
} }
...@@ -2472,9 +2497,9 @@ void Analysis::VisitChoice(ChoiceNode* that) { ...@@ -2472,9 +2497,9 @@ void Analysis::VisitChoice(ChoiceNode* that) {
for (int i = 0; i < that->alternatives()->length(); i++) { for (int i = 0; i < that->alternatives()->length(); i++) {
RegExpNode* node = that->alternatives()->at(i).node(); RegExpNode* node = that->alternatives()->at(i).node();
EnsureAnalyzed(node); EnsureAnalyzed(node);
info->determine_newline |= node->info()->prev_determine_newline(); // Anything the following nodes need to know has to be known by
info->determine_word |= node->info()->prev_determine_word(); // this node also, so it can pass it on.
info->determine_start |= node->info()->prev_determine_start(); info->AddFromFollowing(node->info());
} }
if (!that->table_calculated()) { if (!that->table_calculated()) {
DispatchTableConstructor cons(that->table()); DispatchTableConstructor cons(that->table());
...@@ -2607,7 +2632,8 @@ void DispatchTableConstructor::VisitAction(ActionNode* that) { ...@@ -2607,7 +2632,8 @@ void DispatchTableConstructor::VisitAction(ActionNode* that) {
Handle<FixedArray> RegExpEngine::Compile(RegExpParseResult* input, Handle<FixedArray> RegExpEngine::Compile(RegExpParseResult* input,
RegExpNode** node_return, RegExpNode** node_return,
bool ignore_case) { bool ignore_case,
bool is_multiline) {
RegExpCompiler compiler(input->capture_count, ignore_case); RegExpCompiler compiler(input->capture_count, ignore_case);
// Wrap the body of the regexp in capture #0. // Wrap the body of the regexp in capture #0.
RegExpNode* captured_body = RegExpCapture::ToNode(input->tree, RegExpNode* captured_body = RegExpCapture::ToNode(input->tree,
......
src/jsregexp.h
View file @ 4213b46b
...@@ -437,6 +437,10 @@ class TextElement { ...@@ -437,6 +437,10 @@ class TextElement {
struct NodeInfo { struct NodeInfo {
enum PrecedingInfo {
UNKNOWN = -1, FALSE = 0, TRUE = 1
};
NodeInfo() NodeInfo()
: being_analyzed(false), : being_analyzed(false),
been_analyzed(false), been_analyzed(false),
...@@ -445,34 +449,59 @@ struct NodeInfo { ...@@ -445,34 +449,59 @@ struct NodeInfo {
determine_start(false), determine_start(false),
follows_word_interest(false), follows_word_interest(false),
follows_newline_interest(false), follows_newline_interest(false),
follows_start_interest(false) { } follows_start_interest(false),
bool SameInterests(NodeInfo* that) { at_end(false),
return (follows_word_interest == that->follows_word_interest) follows_word(UNKNOWN),
follows_newline(UNKNOWN),
follows_start(UNKNOWN) { }
bool HasSameForwardInterests(NodeInfo* that) {
return (at_end == that->at_end)
&& (follows_word_interest == that->follows_word_interest)
&& (follows_newline_interest == that->follows_newline_interest) && (follows_newline_interest == that->follows_newline_interest)
&& (follows_start_interest == that->follows_start_interest); && (follows_start_interest == that->follows_start_interest);
} }
void AdoptInterests(NodeInfo* that) {
follows_word_interest = that->follows_word_interest; // Updates the interests of this node given the interests of the
follows_newline_interest = that->follows_newline_interest; // node preceding it.
follows_start_interest = that->follows_start_interest; void AddFromPreceding(NodeInfo* that) {
} at_end |= that->at_end;
bool prev_determine_word() { follows_word_interest |= that->follows_word_interest;
return determine_word || follows_word_interest; follows_newline_interest |= that->follows_newline_interest;
} follows_start_interest |= that->follows_start_interest;
bool prev_determine_newline() {
return determine_newline || follows_newline_interest;
} }
bool prev_determine_start() {
return determine_start || follows_start_interest; // Sets the interests of this node to include the interests of the
// following node.
void AddFromFollowing(NodeInfo* that) {
follows_word_interest |= that->follows_word_interest;
follows_newline_interest |= that->follows_newline_interest;
follows_start_interest |= that->follows_start_interest;
} }
bool being_analyzed: 1; bool being_analyzed: 1;
bool been_analyzed: 1; bool been_analyzed: 1;
// These bits are set if this node must propagate forward information
// about the last character it consumed (or, in the case of 'start',
// if it is at the start of the input).
bool determine_word: 1; bool determine_word: 1;
bool determine_newline: 1; bool determine_newline: 1;
bool determine_start: 1; bool determine_start: 1;
// These bits are set of this node has to know what the preceding
// character was.
bool follows_word_interest: 1; bool follows_word_interest: 1;
bool follows_newline_interest: 1; bool follows_newline_interest: 1;
bool follows_start_interest: 1; bool follows_start_interest: 1;
bool at_end: 1;
// These bits are set if the node can make assumptions about what
// the previous character was.
PrecedingInfo follows_word: 2;
PrecedingInfo follows_newline: 2;
PrecedingInfo follows_start: 2;
}; };
...@@ -511,7 +540,13 @@ class RegExpNode: public ZoneObject { ...@@ -511,7 +540,13 @@ class RegExpNode: public ZoneObject {
// Until the implementation is complete we will return true for success and // Until the implementation is complete we will return true for success and
// false for failure. // false for failure.
virtual bool Emit(RegExpCompiler* compiler) = 0; virtual bool Emit(RegExpCompiler* compiler) = 0;
virtual RegExpNode* PropagateInterest(NodeInfo* info) = 0;
// Propagates the given interest information forward. When seeing
// \bfoo for instance, the \b is implemented by propagating forward
// to the 'foo' string that it should only succeed if its first
// character is a letter xor the previous character was a letter.
virtual RegExpNode* PropagateForward(NodeInfo* info) = 0;
NodeInfo* info() { return &info_; } NodeInfo* info() { return &info_; }
virtual bool IsBacktrack() { return false; } virtual bool IsBacktrack() { return false; }
RegExpNode* GetSibling(NodeInfo* info); RegExpNode* GetSibling(NodeInfo* info);
...@@ -558,7 +593,7 @@ class ActionNode: public SeqRegExpNode { ...@@ -558,7 +593,7 @@ class ActionNode: public SeqRegExpNode {
static ActionNode* EscapeSubmatch(int reg, RegExpNode* on_success); static ActionNode* EscapeSubmatch(int reg, RegExpNode* on_success);
virtual void Accept(NodeVisitor* visitor); virtual void Accept(NodeVisitor* visitor);
virtual bool Emit(RegExpCompiler* compiler); virtual bool Emit(RegExpCompiler* compiler);
virtual RegExpNode* PropagateInterest(NodeInfo* info); virtual RegExpNode* PropagateForward(NodeInfo* info);
private: private:
union { union {
struct { struct {
...@@ -592,7 +627,7 @@ class TextNode: public SeqRegExpNode { ...@@ -592,7 +627,7 @@ class TextNode: public SeqRegExpNode {
on_failure_(on_failure), on_failure_(on_failure),
elms_(elms) { } elms_(elms) { }
virtual void Accept(NodeVisitor* visitor); virtual void Accept(NodeVisitor* visitor);
virtual RegExpNode* PropagateInterest(NodeInfo* info); virtual RegExpNode* PropagateForward(NodeInfo* info);
RegExpNode* on_failure() { return on_failure_; } RegExpNode* on_failure() { return on_failure_; }
virtual bool Emit(RegExpCompiler* compiler); virtual bool Emit(RegExpCompiler* compiler);
ZoneList<TextElement>* elements() { return elms_; } ZoneList<TextElement>* elements() { return elms_; }
...@@ -618,7 +653,7 @@ class BackReferenceNode: public SeqRegExpNode { ...@@ -618,7 +653,7 @@ class BackReferenceNode: public SeqRegExpNode {
int start_register() { return start_reg_; } int start_register() { return start_reg_; }
int end_register() { return end_reg_; } int end_register() { return end_reg_; }
virtual bool Emit(RegExpCompiler* compiler); virtual bool Emit(RegExpCompiler* compiler);
virtual RegExpNode* PropagateInterest(NodeInfo* info); virtual RegExpNode* PropagateForward(NodeInfo* info);
private: private:
RegExpNode* on_failure_; RegExpNode* on_failure_;
int start_reg_; int start_reg_;
...@@ -632,7 +667,7 @@ class EndNode: public RegExpNode { ...@@ -632,7 +667,7 @@ class EndNode: public RegExpNode {
explicit EndNode(Action action) : action_(action) { } explicit EndNode(Action action) : action_(action) { }
virtual void Accept(NodeVisitor* visitor); virtual void Accept(NodeVisitor* visitor);
virtual bool Emit(RegExpCompiler* compiler); virtual bool Emit(RegExpCompiler* compiler);
virtual RegExpNode* PropagateInterest(NodeInfo* info); virtual RegExpNode* PropagateForward(NodeInfo* info);
virtual bool IsBacktrack() { return action_ == BACKTRACK; } virtual bool IsBacktrack() { return action_ == BACKTRACK; }
virtual bool GoTo(RegExpCompiler* compiler); virtual bool GoTo(RegExpCompiler* compiler);
private: private:
...@@ -683,7 +718,7 @@ class ChoiceNode: public RegExpNode { ...@@ -683,7 +718,7 @@ class ChoiceNode: public RegExpNode {
DispatchTable* table() { return &table_; } DispatchTable* table() { return &table_; }
RegExpNode* on_failure() { return on_failure_; } RegExpNode* on_failure() { return on_failure_; }
virtual bool Emit(RegExpCompiler* compiler); virtual bool Emit(RegExpCompiler* compiler);
virtual RegExpNode* PropagateInterest(NodeInfo* info); virtual RegExpNode* PropagateForward(NodeInfo* info);
bool table_calculated() { return table_calculated_; } bool table_calculated() { return table_calculated_; }
void set_table_calculated(bool b) { table_calculated_ = b; } void set_table_calculated(bool b) { table_calculated_ = b; }
bool being_calculated() { return being_calculated_; } bool being_calculated() { return being_calculated_; }
...@@ -770,7 +805,8 @@ class RegExpEngine: public AllStatic { ...@@ -770,7 +805,8 @@ class RegExpEngine: public AllStatic {
public: public:
static Handle<FixedArray> Compile(RegExpParseResult* input, static Handle<FixedArray> Compile(RegExpParseResult* input,
RegExpNode** node_return, RegExpNode** node_return,
bool ignore_case); bool ignore_case,
bool multiline);
static void DotPrint(const char* label, RegExpNode* node); static void DotPrint(const char* label, RegExpNode* node);
}; };
......
src/parser.cc
View file @ 4213b46b
...@@ -545,7 +545,7 @@ class RegExpParser { ...@@ -545,7 +545,7 @@ class RegExpParser {
bool CaptureAvailable(int index); bool CaptureAvailable(int index);
uc32 current_; uc32 current_;
bool has_more_; bool has_more_;
bool multiline_mode_; bool multiline_;
int next_pos_; int next_pos_;
FlatStringReader* in_; FlatStringReader* in_;
Handle<String>* error_; Handle<String>* error_;
...@@ -3499,10 +3499,10 @@ Expression* Parser::NewThrowError(Handle<String> constructor, ...@@ -3499,10 +3499,10 @@ Expression* Parser::NewThrowError(Handle<String> constructor,
RegExpParser::RegExpParser(FlatStringReader* in, RegExpParser::RegExpParser(FlatStringReader* in,
Handle<String>* error, Handle<String>* error,
bool multiline_mode) bool multiline)
: current_(kEndMarker), : current_(kEndMarker),
has_more_(true), has_more_(true),
multiline_mode_(multiline_mode), multiline_(multiline),
next_pos_(0), next_pos_(0),
in_(in), in_(in),
error_(error), error_(error),
...@@ -3617,16 +3617,16 @@ RegExpTree* RegExpParser::ParseDisjunction(bool* ok) { ...@@ -3617,16 +3617,16 @@ RegExpTree* RegExpParser::ParseDisjunction(bool* ok) {
case '^': { case '^': {
Advance(); Advance();
RegExpAssertion::Type type = RegExpAssertion::Type type =
multiline_mode_ ? RegExpAssertion::START_OF_LINE : multiline_ ? RegExpAssertion::START_OF_LINE :
RegExpAssertion::START_OF_INPUT; RegExpAssertion::START_OF_INPUT;
builder.AddAssertion(new RegExpAssertion(type)); builder.AddAssertion(new RegExpAssertion(type));
continue; continue;
} }
case '$': { case '$': {
Advance(); Advance();
RegExpAssertion::Type type = RegExpAssertion::Type type =
multiline_mode_ ? RegExpAssertion::END_OF_LINE : multiline_ ? RegExpAssertion::END_OF_LINE :
RegExpAssertion::END_OF_INPUT; RegExpAssertion::END_OF_INPUT;
builder.AddAssertion(new RegExpAssertion(type)); builder.AddAssertion(new RegExpAssertion(type));
continue; continue;
} }
...@@ -4294,10 +4294,11 @@ ScriptDataImpl* PreParse(unibrow::CharacterStream* stream, ...@@ -4294,10 +4294,11 @@ ScriptDataImpl* PreParse(unibrow::CharacterStream* stream,
} }
bool ParseRegExp(FlatStringReader* input, RegExpParseResult* result) { bool ParseRegExp(FlatStringReader* input,
bool multiline,
RegExpParseResult* result) {
ASSERT(result != NULL); ASSERT(result != NULL);
// TODO(plesner): Get multiline flag somehow RegExpParser parser(input, &result->error, multiline);
RegExpParser parser(input, &result->error, false);
bool ok = true; bool ok = true;
result->tree = parser.ParsePattern(&ok); result->tree = parser.ParsePattern(&ok);
if (!ok) { if (!ok) {
......
src/parser.h
View file @ 4213b46b
...@@ -145,7 +145,9 @@ ScriptDataImpl* PreParse(unibrow::CharacterStream* stream, ...@@ -145,7 +145,9 @@ ScriptDataImpl* PreParse(unibrow::CharacterStream* stream,
v8::Extension* extension); v8::Extension* extension);
bool ParseRegExp(FlatStringReader* input, RegExpParseResult* result); bool ParseRegExp(FlatStringReader* input,
bool multiline,
RegExpParseResult* result);
// Support for doing lazy compilation. The script is the script containing full // Support for doing lazy compilation. The script is the script containing full
......
test/cctest/test-regexp.cc
View file @ 4213b46b
...@@ -56,7 +56,7 @@ static SmartPointer<const char> Parse(const char* input) { ...@@ -56,7 +56,7 @@ static SmartPointer<const char> Parse(const char* input) {
ZoneScope zone_scope(DELETE_ON_EXIT); ZoneScope zone_scope(DELETE_ON_EXIT);
FlatStringReader reader(CStrVector(input)); FlatStringReader reader(CStrVector(input));
RegExpParseResult result; RegExpParseResult result;
CHECK(v8::internal::ParseRegExp(&reader, &result)); CHECK(v8::internal::ParseRegExp(&reader, false, &result));
CHECK(result.tree != NULL); CHECK(result.tree != NULL);
CHECK(result.error.is_null()); CHECK(result.error.is_null());
SmartPointer<const char> output = result.tree->ToString(); SmartPointer<const char> output = result.tree->ToString();
...@@ -69,7 +69,7 @@ static bool ParseEscapes(const char* input) { ...@@ -69,7 +69,7 @@ static bool ParseEscapes(const char* input) {
ZoneScope zone_scope(DELETE_ON_EXIT); ZoneScope zone_scope(DELETE_ON_EXIT);
FlatStringReader reader(CStrVector(input)); FlatStringReader reader(CStrVector(input));
RegExpParseResult result; RegExpParseResult result;
CHECK(v8::internal::ParseRegExp(&reader, &result)); CHECK(v8::internal::ParseRegExp(&reader, false, &result));
CHECK(result.tree != NULL); CHECK(result.tree != NULL);
CHECK(result.error.is_null()); CHECK(result.error.is_null());
return result.has_character_escapes; return result.has_character_escapes;
...@@ -257,7 +257,7 @@ static void ExpectError(const char* input, ...@@ -257,7 +257,7 @@ static void ExpectError(const char* input,
ZoneScope zone_scope(DELETE_ON_EXIT); ZoneScope zone_scope(DELETE_ON_EXIT);
FlatStringReader reader(CStrVector(input)); FlatStringReader reader(CStrVector(input));
RegExpParseResult result; RegExpParseResult result;
CHECK_EQ(false, v8::internal::ParseRegExp(&reader, &result)); CHECK_EQ(false, v8::internal::ParseRegExp(&reader, false, &result));
CHECK(result.tree == NULL); CHECK(result.tree == NULL);
CHECK(!result.error.is_null()); CHECK(!result.error.is_null());
SmartPointer<char> str = result.error->ToCString(ALLOW_NULLS); SmartPointer<char> str = result.error->ToCString(ALLOW_NULLS);
...@@ -373,23 +373,23 @@ TEST(CharacterClassEscapes) { ...@@ -373,23 +373,23 @@ TEST(CharacterClassEscapes) {
} }
static RegExpNode* Compile(const char* input) { static RegExpNode* Compile(const char* input, bool multiline) {
FlatStringReader reader(CStrVector(input)); FlatStringReader reader(CStrVector(input));
RegExpParseResult result; RegExpParseResult result;
if (!v8::internal::ParseRegExp(&reader, &result)) if (!v8::internal::ParseRegExp(&reader, multiline, &result))
return NULL; return NULL;
RegExpNode* node = NULL; RegExpNode* node = NULL;
RegExpEngine::Compile(&result, &node, false); RegExpEngine::Compile(&result, &node, false, multiline);
return node; return node;
} }
static void Execute(const char* input, static void Execute(const char* input,
const char* str, bool multiline,
bool dot_output = false) { bool dot_output = false) {
v8::HandleScope scope; v8::HandleScope scope;
ZoneScope zone_scope(DELETE_ON_EXIT); ZoneScope zone_scope(DELETE_ON_EXIT);
RegExpNode* node = Compile(input); RegExpNode* node = Compile(input, multiline);
USE(node); USE(node);
#ifdef DEBUG #ifdef DEBUG
if (dot_output) { if (dot_output) {
...@@ -400,14 +400,6 @@ static void Execute(const char* input, ...@@ -400,14 +400,6 @@ static void Execute(const char* input,
} }
TEST(Execution) {
V8::Initialize(NULL);
Execute(".*?(?:a[bc]d|e[fg]h)", "xxxabbegh");
Execute(".*?(?:a[bc]d|e[fg]h)", "xxxabbefh");
Execute(".*?(?:a[bc]d|e[fg]h)", "xxxabbefd");
}
class TestConfig { class TestConfig {
public: public:
typedef int Key; typedef int Key;
...@@ -1043,8 +1035,8 @@ TEST(LatinCanonicalize) { ...@@ -1043,8 +1035,8 @@ TEST(LatinCanonicalize) {
TEST(SimplePropagation) { TEST(SimplePropagation) {
v8::HandleScope scope; v8::HandleScope scope;
ZoneScope zone_scope(DELETE_ON_EXIT); ZoneScope zone_scope(DELETE_ON_EXIT);
RegExpNode* node = Compile("(a|^b|c)"); RegExpNode* node = Compile("(a|^b|c)", false);
CHECK(node->info()->determine_start); CHECK(node->info()->follows_start_interest);
} }
...@@ -1061,7 +1053,7 @@ static uc32 CanonRange(uc32 c) { ...@@ -1061,7 +1053,7 @@ static uc32 CanonRange(uc32 c) {
TEST(RangeCanonicalization) { TEST(RangeCanonicalization) {
ASSERT((CanonRange(0) & CharacterRange::kStartMarker) != 0); CHECK((CanonRange(0) & CharacterRange::kStartMarker) != 0);
// Check that we arrive at the same result when using the basic // Check that we arrive at the same result when using the basic
// range canonicalization primitives as when using immediate // range canonicalization primitives as when using immediate
// canonicalization. // canonicalization.
...@@ -1175,5 +1167,5 @@ TEST(CharacterRangeCaseIndependence) { ...@@ -1175,5 +1167,5 @@ TEST(CharacterRangeCaseIndependence) {
TEST(Graph) { TEST(Graph) {
V8::Initialize(NULL); V8::Initialize(NULL);
Execute("\\w+", "", true); Execute("(?:foo|bar$)", false, true);
} }
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment