OLD | NEW |
---|---|
1 // Copyright 2012 the V8 project authors. All rights reserved. | 1 // Copyright 2012 the V8 project authors. All rights reserved. |
2 // Redistribution and use in source and binary forms, with or without | 2 // Redistribution and use in source and binary forms, with or without |
3 // modification, are permitted provided that the following conditions are | 3 // modification, are permitted provided that the following conditions are |
4 // met: | 4 // met: |
5 // | 5 // |
6 // * Redistributions of source code must retain the above copyright | 6 // * Redistributions of source code must retain the above copyright |
7 // notice, this list of conditions and the following disclaimer. | 7 // notice, this list of conditions and the following disclaimer. |
8 // * Redistributions in binary form must reproduce the above | 8 // * Redistributions in binary form must reproduce the above |
9 // copyright notice, this list of conditions and the following | 9 // copyright notice, this list of conditions and the following |
10 // disclaimer in the documentation and/or other materials provided | 10 // disclaimer in the documentation and/or other materials provided |
(...skipping 2556 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
2567 FixedArrayBuilder array_builder_; | 2567 FixedArrayBuilder array_builder_; |
2568 Handle<String> subject_; | 2568 Handle<String> subject_; |
2569 int character_count_; | 2569 int character_count_; |
2570 bool is_ascii_; | 2570 bool is_ascii_; |
2571 }; | 2571 }; |
2572 | 2572 |
2573 | 2573 |
2574 class CompiledReplacement { | 2574 class CompiledReplacement { |
2575 public: | 2575 public: |
2576 explicit CompiledReplacement(Zone* zone) | 2576 explicit CompiledReplacement(Zone* zone) |
2577 : parts_(1, zone), replacement_substrings_(0, zone), zone_(zone) {} | 2577 : parts_(1, zone), replacement_substrings_(0, zone), |
2578 simple_hint_(false), | |
2579 zone_(zone) {} | |
2578 | 2580 |
2579 // Return whether the replacement is simple. | 2581 void Compile(Handle<String> replacement, |
2580 bool Compile(Handle<String> replacement, | |
2581 int capture_count, | 2582 int capture_count, |
2582 int subject_length); | 2583 int subject_length); |
2583 | 2584 |
2584 // Use Apply only if Compile returned false. | |
2585 void Apply(ReplacementStringBuilder* builder, | 2585 void Apply(ReplacementStringBuilder* builder, |
2586 int match_from, | 2586 int match_from, |
2587 int match_to, | 2587 int match_to, |
2588 int32_t* match); | 2588 Handle<JSArray> last_match_info); |
2589 | 2589 |
2590 // Number of distinct parts of the replacement pattern. | 2590 // Number of distinct parts of the replacement pattern. |
2591 int parts() { | 2591 int parts() { |
2592 return parts_.length(); | 2592 return parts_.length(); |
2593 } | 2593 } |
2594 | 2594 |
2595 bool simple_hint() { | |
2596 return simple_hint_; | |
2597 } | |
2598 | |
2595 Zone* zone() const { return zone_; } | 2599 Zone* zone() const { return zone_; } |
2596 | 2600 |
2597 private: | 2601 private: |
2598 enum PartType { | 2602 enum PartType { |
2599 SUBJECT_PREFIX = 1, | 2603 SUBJECT_PREFIX = 1, |
2600 SUBJECT_SUFFIX, | 2604 SUBJECT_SUFFIX, |
2601 SUBJECT_CAPTURE, | 2605 SUBJECT_CAPTURE, |
2602 REPLACEMENT_SUBSTRING, | 2606 REPLACEMENT_SUBSTRING, |
2603 REPLACEMENT_STRING, | 2607 REPLACEMENT_STRING, |
2604 | 2608 |
(...skipping 40 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
2645 // tag == REPLACEMENT_STRING: data is index into array of substrings | 2649 // tag == REPLACEMENT_STRING: data is index into array of substrings |
2646 // of the replacement string. | 2650 // of the replacement string. |
2647 // tag <= 0: Temporary representation of the substring of the replacement | 2651 // tag <= 0: Temporary representation of the substring of the replacement |
2648 // string ranging over -tag .. data. | 2652 // string ranging over -tag .. data. |
2649 // Is replaced by REPLACEMENT_{SUB,}STRING when we create the | 2653 // Is replaced by REPLACEMENT_{SUB,}STRING when we create the |
2650 // substring objects. | 2654 // substring objects. |
2651 int data; | 2655 int data; |
2652 }; | 2656 }; |
2653 | 2657 |
2654 template<typename Char> | 2658 template<typename Char> |
2655 bool ParseReplacementPattern(ZoneList<ReplacementPart>* parts, | 2659 static bool ParseReplacementPattern(ZoneList<ReplacementPart>* parts, |
2656 Vector<Char> characters, | 2660 Vector<Char> characters, |
2657 int capture_count, | 2661 int capture_count, |
2658 int subject_length, | 2662 int subject_length, |
2659 Zone* zone) { | 2663 Zone* zone) { |
2660 int length = characters.length(); | 2664 int length = characters.length(); |
2661 int last = 0; | 2665 int last = 0; |
2662 for (int i = 0; i < length; i++) { | 2666 for (int i = 0; i < length; i++) { |
2663 Char c = characters[i]; | 2667 Char c = characters[i]; |
2664 if (c == '$') { | 2668 if (c == '$') { |
2665 int next_index = i + 1; | 2669 int next_index = i + 1; |
2666 if (next_index == length) { // No next character! | 2670 if (next_index == length) { // No next character! |
2667 break; | 2671 break; |
2668 } | 2672 } |
2669 Char c2 = characters[next_index]; | 2673 Char c2 = characters[next_index]; |
(...skipping 73 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
2743 break; | 2747 break; |
2744 } | 2748 } |
2745 default: | 2749 default: |
2746 i = next_index; | 2750 i = next_index; |
2747 break; | 2751 break; |
2748 } | 2752 } |
2749 } | 2753 } |
2750 } | 2754 } |
2751 if (length > last) { | 2755 if (length > last) { |
2752 if (last == 0) { | 2756 if (last == 0) { |
2753 // Replacement is simple. Do not use Apply to do the replacement. | 2757 parts->Add(ReplacementPart::ReplacementString(), zone); |
2754 return true; | 2758 return true; |
2755 } else { | 2759 } else { |
2756 parts->Add(ReplacementPart::ReplacementSubString(last, length), zone); | 2760 parts->Add(ReplacementPart::ReplacementSubString(last, length), zone); |
2757 } | 2761 } |
2758 } | 2762 } |
2759 return false; | 2763 return false; |
2760 } | 2764 } |
2761 | 2765 |
2762 ZoneList<ReplacementPart> parts_; | 2766 ZoneList<ReplacementPart> parts_; |
2763 ZoneList<Handle<String> > replacement_substrings_; | 2767 ZoneList<Handle<String> > replacement_substrings_; |
2768 bool simple_hint_; | |
2764 Zone* zone_; | 2769 Zone* zone_; |
2765 }; | 2770 }; |
2766 | 2771 |
2767 | 2772 |
2768 bool CompiledReplacement::Compile(Handle<String> replacement, | 2773 void CompiledReplacement::Compile(Handle<String> replacement, |
2769 int capture_count, | 2774 int capture_count, |
2770 int subject_length) { | 2775 int subject_length) { |
2771 { | 2776 { |
2772 AssertNoAllocation no_alloc; | 2777 AssertNoAllocation no_alloc; |
2773 String::FlatContent content = replacement->GetFlatContent(); | 2778 String::FlatContent content = replacement->GetFlatContent(); |
2774 ASSERT(content.IsFlat()); | 2779 ASSERT(content.IsFlat()); |
2775 bool simple = false; | |
2776 if (content.IsAscii()) { | 2780 if (content.IsAscii()) { |
2777 simple = ParseReplacementPattern(&parts_, | 2781 simple_hint_ = ParseReplacementPattern(&parts_, |
2778 content.ToAsciiVector(), | 2782 content.ToAsciiVector(), |
2779 capture_count, | 2783 capture_count, |
2780 subject_length, | 2784 subject_length, |
2781 zone()); | 2785 zone()); |
2782 } else { | 2786 } else { |
2783 ASSERT(content.IsTwoByte()); | 2787 ASSERT(content.IsTwoByte()); |
2784 simple = ParseReplacementPattern(&parts_, | 2788 simple_hint_ = ParseReplacementPattern(&parts_, |
2785 content.ToUC16Vector(), | 2789 content.ToUC16Vector(), |
2786 capture_count, | 2790 capture_count, |
2787 subject_length, | 2791 subject_length, |
2788 zone()); | 2792 zone()); |
2789 } | 2793 } |
2790 if (simple) return true; | |
2791 } | 2794 } |
2792 | |
2793 Isolate* isolate = replacement->GetIsolate(); | 2795 Isolate* isolate = replacement->GetIsolate(); |
2794 // Find substrings of replacement string and create them as String objects. | 2796 // Find substrings of replacement string and create them as String objects. |
2795 int substring_index = 0; | 2797 int substring_index = 0; |
2796 for (int i = 0, n = parts_.length(); i < n; i++) { | 2798 for (int i = 0, n = parts_.length(); i < n; i++) { |
2797 int tag = parts_[i].tag; | 2799 int tag = parts_[i].tag; |
2798 if (tag <= 0) { // A replacement string slice. | 2800 if (tag <= 0) { // A replacement string slice. |
2799 int from = -tag; | 2801 int from = -tag; |
2800 int to = parts_[i].data; | 2802 int to = parts_[i].data; |
2801 replacement_substrings_.Add( | 2803 replacement_substrings_.Add( |
2802 isolate->factory()->NewSubString(replacement, from, to), zone()); | 2804 isolate->factory()->NewSubString(replacement, from, to), zone()); |
2803 parts_[i].tag = REPLACEMENT_SUBSTRING; | 2805 parts_[i].tag = REPLACEMENT_SUBSTRING; |
2804 parts_[i].data = substring_index; | 2806 parts_[i].data = substring_index; |
2805 substring_index++; | 2807 substring_index++; |
2806 } else if (tag == REPLACEMENT_STRING) { | 2808 } else if (tag == REPLACEMENT_STRING) { |
2807 replacement_substrings_.Add(replacement, zone()); | 2809 replacement_substrings_.Add(replacement, zone()); |
2808 parts_[i].data = substring_index; | 2810 parts_[i].data = substring_index; |
2809 substring_index++; | 2811 substring_index++; |
2810 } | 2812 } |
2811 } | 2813 } |
2812 return false; | |
2813 } | 2814 } |
2814 | 2815 |
2815 | 2816 |
2816 void CompiledReplacement::Apply(ReplacementStringBuilder* builder, | 2817 void CompiledReplacement::Apply(ReplacementStringBuilder* builder, |
2817 int match_from, | 2818 int match_from, |
2818 int match_to, | 2819 int match_to, |
2819 int32_t* match) { | 2820 Handle<JSArray> last_match_info) { |
2820 ASSERT_LT(0, parts_.length()); | |
2821 for (int i = 0, n = parts_.length(); i < n; i++) { | 2821 for (int i = 0, n = parts_.length(); i < n; i++) { |
2822 ReplacementPart part = parts_[i]; | 2822 ReplacementPart part = parts_[i]; |
2823 switch (part.tag) { | 2823 switch (part.tag) { |
2824 case SUBJECT_PREFIX: | 2824 case SUBJECT_PREFIX: |
2825 if (match_from > 0) builder->AddSubjectSlice(0, match_from); | 2825 if (match_from > 0) builder->AddSubjectSlice(0, match_from); |
2826 break; | 2826 break; |
2827 case SUBJECT_SUFFIX: { | 2827 case SUBJECT_SUFFIX: { |
2828 int subject_length = part.data; | 2828 int subject_length = part.data; |
2829 if (match_to < subject_length) { | 2829 if (match_to < subject_length) { |
2830 builder->AddSubjectSlice(match_to, subject_length); | 2830 builder->AddSubjectSlice(match_to, subject_length); |
2831 } | 2831 } |
2832 break; | 2832 break; |
2833 } | 2833 } |
2834 case SUBJECT_CAPTURE: { | 2834 case SUBJECT_CAPTURE: { |
2835 int capture = part.data; | 2835 int capture = part.data; |
2836 int from = match[capture * 2]; | 2836 FixedArray* match_info = FixedArray::cast(last_match_info->elements()); |
2837 int to = match[capture * 2 + 1]; | 2837 int from = RegExpImpl::GetCapture(match_info, capture * 2); |
2838 int to = RegExpImpl::GetCapture(match_info, capture * 2 + 1); | |
2838 if (from >= 0 && to > from) { | 2839 if (from >= 0 && to > from) { |
2839 builder->AddSubjectSlice(from, to); | 2840 builder->AddSubjectSlice(from, to); |
2840 } | 2841 } |
2841 break; | 2842 break; |
2842 } | 2843 } |
2843 case REPLACEMENT_SUBSTRING: | 2844 case REPLACEMENT_SUBSTRING: |
2844 case REPLACEMENT_STRING: | 2845 case REPLACEMENT_STRING: |
2845 builder->AddString(replacement_substrings_[part.data]); | 2846 builder->AddString(replacement_substrings_[part.data]); |
2846 break; | 2847 break; |
2847 default: | 2848 default: |
(...skipping 101 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
2949 pattern_content.ToUC16Vector(), | 2950 pattern_content.ToUC16Vector(), |
2950 indices, | 2951 indices, |
2951 limit, | 2952 limit, |
2952 zone); | 2953 zone); |
2953 } | 2954 } |
2954 } | 2955 } |
2955 } | 2956 } |
2956 } | 2957 } |
2957 | 2958 |
2958 | 2959 |
2960 // Two smis before and after the match, for very long strings. | |
2961 const int kMaxBuilderEntriesPerRegExpMatch = 5; | |
2962 | |
2963 | |
2964 static void SetLastMatchInfoNoCaptures(Handle<String> subject, | |
2965 Handle<JSArray> last_match_info, | |
2966 int match_start, | |
2967 int match_end) { | |
2968 // Fill last_match_info with a single capture. | |
2969 last_match_info->EnsureSize(2 + RegExpImpl::kLastMatchOverhead); | |
2970 AssertNoAllocation no_gc; | |
2971 FixedArray* elements = FixedArray::cast(last_match_info->elements()); | |
2972 RegExpImpl::SetLastCaptureCount(elements, 2); | |
2973 RegExpImpl::SetLastInput(elements, *subject); | |
2974 RegExpImpl::SetLastSubject(elements, *subject); | |
2975 RegExpImpl::SetCapture(elements, 0, match_start); | |
2976 RegExpImpl::SetCapture(elements, 1, match_end); | |
2977 } | |
2978 | |
2979 | |
2980 template <typename SubjectChar, typename PatternChar> | |
2981 static bool SearchStringMultiple(Isolate* isolate, | |
2982 Vector<const SubjectChar> subject, | |
2983 Vector<const PatternChar> pattern, | |
2984 String* pattern_string, | |
2985 FixedArrayBuilder* builder, | |
2986 int* match_pos) { | |
2987 int pos = *match_pos; | |
2988 int subject_length = subject.length(); | |
2989 int pattern_length = pattern.length(); | |
2990 int max_search_start = subject_length - pattern_length; | |
2991 StringSearch<PatternChar, SubjectChar> search(isolate, pattern); | |
2992 while (pos <= max_search_start) { | |
2993 if (!builder->HasCapacity(kMaxBuilderEntriesPerRegExpMatch)) { | |
2994 *match_pos = pos; | |
2995 return false; | |
2996 } | |
2997 // Position of end of previous match. | |
2998 int match_end = pos + pattern_length; | |
2999 int new_pos = search.Search(subject, match_end); | |
3000 if (new_pos >= 0) { | |
3001 // A match. | |
3002 if (new_pos > match_end) { | |
3003 ReplacementStringBuilder::AddSubjectSlice(builder, | |
3004 match_end, | |
3005 new_pos); | |
3006 } | |
3007 pos = new_pos; | |
3008 builder->Add(pattern_string); | |
3009 } else { | |
3010 break; | |
3011 } | |
3012 } | |
3013 | |
3014 if (pos < max_search_start) { | |
3015 ReplacementStringBuilder::AddSubjectSlice(builder, | |
3016 pos + pattern_length, | |
3017 subject_length); | |
3018 } | |
3019 *match_pos = pos; | |
3020 return true; | |
3021 } | |
3022 | |
3023 | |
3024 | |
3025 | |
2959 template<typename ResultSeqString> | 3026 template<typename ResultSeqString> |
2960 MUST_USE_RESULT static MaybeObject* StringReplaceAtomRegExpWithString( | 3027 MUST_USE_RESULT static MaybeObject* StringReplaceAtomRegExpWithString( |
2961 Isolate* isolate, | 3028 Isolate* isolate, |
2962 Handle<String> subject, | 3029 Handle<String> subject, |
2963 Handle<JSRegExp> pattern_regexp, | 3030 Handle<JSRegExp> pattern_regexp, |
2964 Handle<String> replacement, | 3031 Handle<String> replacement, |
2965 Handle<JSArray> last_match_info) { | 3032 Handle<JSArray> last_match_info, |
3033 Zone* zone) { | |
2966 ASSERT(subject->IsFlat()); | 3034 ASSERT(subject->IsFlat()); |
2967 ASSERT(replacement->IsFlat()); | 3035 ASSERT(replacement->IsFlat()); |
2968 | 3036 |
2969 Zone* zone = isolate->runtime_zone(); | 3037 ZoneScope zone_space(isolate->runtime_zone(), DELETE_ON_EXIT); |
2970 ZoneScope zone_space(zone, DELETE_ON_EXIT); | 3038 ZoneList<int> indices(8, isolate->runtime_zone()); |
2971 ZoneList<int> indices(8, zone); | |
2972 ASSERT_EQ(JSRegExp::ATOM, pattern_regexp->TypeTag()); | 3039 ASSERT_EQ(JSRegExp::ATOM, pattern_regexp->TypeTag()); |
2973 String* pattern = | 3040 String* pattern = |
2974 String::cast(pattern_regexp->DataAt(JSRegExp::kAtomPatternIndex)); | 3041 String::cast(pattern_regexp->DataAt(JSRegExp::kAtomPatternIndex)); |
2975 int subject_len = subject->length(); | 3042 int subject_len = subject->length(); |
2976 int pattern_len = pattern->length(); | 3043 int pattern_len = pattern->length(); |
2977 int replacement_len = replacement->length(); | 3044 int replacement_len = replacement->length(); |
2978 | 3045 |
2979 FindStringIndicesDispatch( | 3046 FindStringIndicesDispatch(isolate, *subject, pattern, &indices, 0xffffffff, |
2980 isolate, *subject, pattern, &indices, 0xffffffff, zone); | 3047 zone); |
2981 | 3048 |
2982 int matches = indices.length(); | 3049 int matches = indices.length(); |
2983 if (matches == 0) return *subject; | 3050 if (matches == 0) return *subject; |
2984 | 3051 |
2985 // Detect integer overflow. | 3052 // Detect integer overflow. |
2986 int64_t result_len_64 = | 3053 int64_t result_len_64 = |
2987 (static_cast<int64_t>(replacement_len) - | 3054 (static_cast<int64_t>(replacement_len) - |
2988 static_cast<int64_t>(pattern_len)) * | 3055 static_cast<int64_t>(pattern_len)) * |
2989 static_cast<int64_t>(matches) + | 3056 static_cast<int64_t>(matches) + |
2990 static_cast<int64_t>(subject_len); | 3057 static_cast<int64_t>(subject_len); |
(...skipping 34 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
3025 subject_pos = indices.at(i) + pattern_len; | 3092 subject_pos = indices.at(i) + pattern_len; |
3026 } | 3093 } |
3027 // Add remaining subject content at the end. | 3094 // Add remaining subject content at the end. |
3028 if (subject_pos < subject_len) { | 3095 if (subject_pos < subject_len) { |
3029 String::WriteToFlat(*subject, | 3096 String::WriteToFlat(*subject, |
3030 result->GetChars() + result_pos, | 3097 result->GetChars() + result_pos, |
3031 subject_pos, | 3098 subject_pos, |
3032 subject_len); | 3099 subject_len); |
3033 } | 3100 } |
3034 | 3101 |
3035 int32_t match_indices[] = { indices.at(matches - 1), | 3102 SetLastMatchInfoNoCaptures(subject, |
3036 indices.at(matches - 1) + pattern_len }; | 3103 last_match_info, |
3037 RegExpImpl::SetLastMatchInfo(last_match_info, subject, 0, match_indices); | 3104 indices.at(matches - 1), |
3105 indices.at(matches - 1) + pattern_len); | |
3038 | 3106 |
3039 return *result; | 3107 return *result; |
3040 } | 3108 } |
3041 | 3109 |
3042 | 3110 |
3043 MUST_USE_RESULT static MaybeObject* StringReplaceRegExpWithString( | 3111 MUST_USE_RESULT static MaybeObject* StringReplaceRegExpWithString( |
3044 Isolate* isolate, | 3112 Isolate* isolate, |
3045 Handle<String> subject, | 3113 String* subject, |
3046 Handle<JSRegExp> regexp, | 3114 JSRegExp* regexp, |
3047 Handle<String> replacement, | 3115 String* replacement, |
3048 Handle<JSArray> last_match_info) { | 3116 JSArray* last_match_info, |
3117 Zone* zone) { | |
3049 ASSERT(subject->IsFlat()); | 3118 ASSERT(subject->IsFlat()); |
3050 ASSERT(replacement->IsFlat()); | 3119 ASSERT(replacement->IsFlat()); |
3051 | 3120 |
3052 bool is_global = regexp->GetFlags().is_global(); | 3121 HandleScope handles(isolate); |
ulan
2012/08/21 09:43:40
Empty line is missing here.
Yang
2012/08/21 09:46:10
Done.
| |
3053 int capture_count = regexp->CaptureCount(); | 3122 int length = subject->length(); |
3054 int subject_length = subject->length(); | 3123 Handle<String> subject_handle(subject); |
3124 Handle<JSRegExp> regexp_handle(regexp); | |
3125 Handle<String> replacement_handle(replacement); | |
3126 Handle<JSArray> last_match_info_handle(last_match_info); | |
3127 Handle<Object> match = RegExpImpl::Exec(regexp_handle, | |
3128 subject_handle, | |
3129 0, | |
3130 last_match_info_handle); | |
3131 if (match.is_null()) { | |
3132 return Failure::Exception(); | |
3133 } | |
3134 if (match->IsNull()) { | |
3135 return *subject_handle; | |
3136 } | |
3137 | |
3138 int capture_count = regexp_handle->CaptureCount(); | |
3055 | 3139 |
3056 // CompiledReplacement uses zone allocation. | 3140 // CompiledReplacement uses zone allocation. |
3057 Zone* zone = isolate->runtime_zone(); | |
3058 ZoneScope zonescope(zone, DELETE_ON_EXIT); | 3141 ZoneScope zonescope(zone, DELETE_ON_EXIT); |
3059 CompiledReplacement compiled_replacement(zone); | 3142 CompiledReplacement compiled_replacement(zone); |
3060 bool simple_replace = compiled_replacement.Compile(replacement, | 3143 |
3061 capture_count, | 3144 compiled_replacement.Compile(replacement_handle, |
3062 subject_length); | 3145 capture_count, |
3146 length); | |
3147 | |
3148 bool is_global = regexp_handle->GetFlags().is_global(); | |
3063 | 3149 |
3064 // Shortcut for simple non-regexp global replacements | 3150 // Shortcut for simple non-regexp global replacements |
3065 if (is_global && | 3151 if (is_global && |
3066 regexp->TypeTag() == JSRegExp::ATOM && | 3152 regexp_handle->TypeTag() == JSRegExp::ATOM && |
3067 simple_replace) { | 3153 compiled_replacement.simple_hint()) { |
3068 if (subject->HasOnlyAsciiChars()) { | 3154 if (subject_handle->HasOnlyAsciiChars() && |
3155 replacement_handle->HasOnlyAsciiChars()) { | |
3069 return StringReplaceAtomRegExpWithString<SeqAsciiString>( | 3156 return StringReplaceAtomRegExpWithString<SeqAsciiString>( |
3070 isolate, subject, regexp, replacement, last_match_info); | 3157 isolate, |
3071 } else { | 3158 subject_handle, |
3159 regexp_handle, | |
3160 replacement_handle, | |
3161 last_match_info_handle, | |
3162 zone); | |
3163 } else { | |
3072 return StringReplaceAtomRegExpWithString<SeqTwoByteString>( | 3164 return StringReplaceAtomRegExpWithString<SeqTwoByteString>( |
3073 isolate, subject, regexp, replacement, last_match_info); | 3165 isolate, |
3166 subject_handle, | |
3167 regexp_handle, | |
3168 replacement_handle, | |
3169 last_match_info_handle, | |
3170 zone); | |
3074 } | 3171 } |
3075 } | 3172 } |
3076 | 3173 |
3077 RegExpImpl::GlobalCache global_cache(regexp, subject, is_global, isolate); | |
3078 if (global_cache.HasException()) return Failure::Exception(); | |
3079 | |
3080 int32_t* current_match = global_cache.FetchNext(); | |
3081 if (current_match == NULL) { | |
3082 if (global_cache.HasException()) return Failure::Exception(); | |
3083 return *subject; | |
3084 } | |
3085 | |
3086 // Guessing the number of parts that the final result string is built | 3174 // Guessing the number of parts that the final result string is built |
3087 // from. Global regexps can match any number of times, so we guess | 3175 // from. Global regexps can match any number of times, so we guess |
3088 // conservatively. | 3176 // conservatively. |
3089 int expected_parts = | 3177 int expected_parts = |
3090 (compiled_replacement.parts() + 1) * (is_global ? 4 : 1) + 1; | 3178 (compiled_replacement.parts() + 1) * (is_global ? 4 : 1) + 1; |
3091 ReplacementStringBuilder builder(isolate->heap(), | 3179 ReplacementStringBuilder builder(isolate->heap(), |
3092 subject, | 3180 subject_handle, |
3093 expected_parts); | 3181 expected_parts); |
3094 | 3182 |
3183 // Index of end of last match. | |
3184 int prev = 0; | |
3185 | |
3186 | |
3095 // Number of parts added by compiled replacement plus preceeding | 3187 // Number of parts added by compiled replacement plus preceeding |
3096 // string and possibly suffix after last match. It is possible for | 3188 // string and possibly suffix after last match. It is possible for |
3097 // all components to use two elements when encoded as two smis. | 3189 // all components to use two elements when encoded as two smis. |
3098 const int parts_added_per_loop = 2 * (compiled_replacement.parts() + 2); | 3190 const int parts_added_per_loop = 2 * (compiled_replacement.parts() + 2); |
3099 | 3191 bool matched = true; |
3100 int prev = 0; | |
3101 | |
3102 do { | 3192 do { |
3193 ASSERT(last_match_info_handle->HasFastObjectElements()); | |
3194 // Increase the capacity of the builder before entering local handle-scope, | |
3195 // so its internal buffer can safely allocate a new handle if it grows. | |
3103 builder.EnsureCapacity(parts_added_per_loop); | 3196 builder.EnsureCapacity(parts_added_per_loop); |
3104 | 3197 |
3105 int start = current_match[0]; | 3198 HandleScope loop_scope(isolate); |
3106 int end = current_match[1]; | 3199 int start, end; |
3200 { | |
3201 AssertNoAllocation match_info_array_is_not_in_a_handle; | |
3202 FixedArray* match_info_array = | |
3203 FixedArray::cast(last_match_info_handle->elements()); | |
3204 | |
3205 ASSERT_EQ(capture_count * 2 + 2, | |
3206 RegExpImpl::GetLastCaptureCount(match_info_array)); | |
3207 start = RegExpImpl::GetCapture(match_info_array, 0); | |
3208 end = RegExpImpl::GetCapture(match_info_array, 1); | |
3209 } | |
3107 | 3210 |
3108 if (prev < start) { | 3211 if (prev < start) { |
3109 builder.AddSubjectSlice(prev, start); | 3212 builder.AddSubjectSlice(prev, start); |
3110 } | 3213 } |
3214 compiled_replacement.Apply(&builder, | |
3215 start, | |
3216 end, | |
3217 last_match_info_handle); | |
3111 | 3218 |
3112 if (simple_replace) { | |
3113 builder.AddString(replacement); | |
3114 } else { | |
3115 compiled_replacement.Apply(&builder, | |
3116 start, | |
3117 end, | |
3118 current_match); | |
3119 } | |
3120 prev = end; | 3219 prev = end; |
3121 | 3220 |
3122 // Only continue checking for global regexps. | 3221 // Only continue checking for global regexps. |
3123 if (!is_global) break; | 3222 if (!is_global) break; |
3124 | 3223 |
3125 current_match = global_cache.FetchNext(); | 3224 // Continue from where the match ended, unless it was an empty match. |
3126 } while (current_match != NULL); | 3225 int next = end; |
3226 if (start == end) { | |
3227 next = end + 1; | |
3228 if (next > length) break; | |
3229 } | |
3127 | 3230 |
3128 if (global_cache.HasException()) return Failure::Exception(); | 3231 match = RegExpImpl::Exec(regexp_handle, |
3232 subject_handle, | |
3233 next, | |
3234 last_match_info_handle); | |
3235 if (match.is_null()) { | |
3236 return Failure::Exception(); | |
3237 } | |
3238 matched = !match->IsNull(); | |
3239 } while (matched); | |
3129 | 3240 |
3130 if (prev < subject_length) { | 3241 if (prev < length) { |
3131 builder.EnsureCapacity(2); | 3242 builder.AddSubjectSlice(prev, length); |
3132 builder.AddSubjectSlice(prev, subject_length); | |
3133 } | 3243 } |
3134 | 3244 |
3135 RegExpImpl::SetLastMatchInfo(last_match_info, | |
3136 subject, | |
3137 capture_count, | |
3138 global_cache.LastSuccessfulMatch()); | |
3139 | |
3140 return *(builder.ToString()); | 3245 return *(builder.ToString()); |
3141 } | 3246 } |
3142 | 3247 |
3143 | 3248 |
3144 template <typename ResultSeqString> | 3249 template <typename ResultSeqString> |
3145 MUST_USE_RESULT static MaybeObject* StringReplaceRegExpWithEmptyString( | 3250 MUST_USE_RESULT static MaybeObject* StringReplaceRegExpWithEmptyString( |
3146 Isolate* isolate, | 3251 Isolate* isolate, |
3147 Handle<String> subject, | 3252 String* subject, |
3148 Handle<JSRegExp> regexp, | 3253 JSRegExp* regexp, |
3149 Handle<JSArray> last_match_info) { | 3254 JSArray* last_match_info, |
3255 Zone* zone) { | |
3150 ASSERT(subject->IsFlat()); | 3256 ASSERT(subject->IsFlat()); |
3151 | 3257 |
3152 bool is_global = regexp->GetFlags().is_global(); | 3258 HandleScope handles(isolate); |
3259 | |
3260 Handle<String> subject_handle(subject); | |
3261 Handle<JSRegExp> regexp_handle(regexp); | |
3262 Handle<JSArray> last_match_info_handle(last_match_info); | |
3153 | 3263 |
3154 // Shortcut for simple non-regexp global replacements | 3264 // Shortcut for simple non-regexp global replacements |
3155 if (is_global && | 3265 if (regexp_handle->GetFlags().is_global() && |
3156 regexp->TypeTag() == JSRegExp::ATOM) { | 3266 regexp_handle->TypeTag() == JSRegExp::ATOM) { |
3157 Handle<String> empty_string(HEAP->empty_string()); | 3267 Handle<String> empty_string_handle(HEAP->empty_string()); |
3158 if (subject->HasOnlyAsciiChars()) { | 3268 if (subject_handle->HasOnlyAsciiChars()) { |
3159 return StringReplaceAtomRegExpWithString<SeqAsciiString>( | 3269 return StringReplaceAtomRegExpWithString<SeqAsciiString>( |
3160 isolate, | 3270 isolate, |
3161 subject, | 3271 subject_handle, |
3162 regexp, | 3272 regexp_handle, |
3163 empty_string, | 3273 empty_string_handle, |
3164 last_match_info); | 3274 last_match_info_handle, |
3275 zone); | |
3165 } else { | 3276 } else { |
3166 return StringReplaceAtomRegExpWithString<SeqTwoByteString>( | 3277 return StringReplaceAtomRegExpWithString<SeqTwoByteString>( |
3167 isolate, | 3278 isolate, |
3168 subject, | 3279 subject_handle, |
3169 regexp, | 3280 regexp_handle, |
3170 empty_string, | 3281 empty_string_handle, |
3171 last_match_info); | 3282 last_match_info_handle, |
3283 zone); | |
3172 } | 3284 } |
3173 } | 3285 } |
3174 | 3286 |
3175 RegExpImpl::GlobalCache global_cache(regexp, subject, is_global, isolate); | 3287 Handle<Object> match = RegExpImpl::Exec(regexp_handle, |
3176 if (global_cache.HasException()) return Failure::Exception(); | 3288 subject_handle, |
3289 0, | |
3290 last_match_info_handle); | |
3291 if (match.is_null()) return Failure::Exception(); | |
3292 if (match->IsNull()) return *subject_handle; | |
3177 | 3293 |
3178 int32_t* current_match = global_cache.FetchNext(); | 3294 ASSERT(last_match_info_handle->HasFastObjectElements()); |
3179 if (current_match == NULL) { | 3295 |
3180 if (global_cache.HasException()) return Failure::Exception(); | 3296 int start, end; |
3181 return *subject; | 3297 { |
3298 AssertNoAllocation match_info_array_is_not_in_a_handle; | |
3299 FixedArray* match_info_array = | |
3300 FixedArray::cast(last_match_info_handle->elements()); | |
3301 | |
3302 start = RegExpImpl::GetCapture(match_info_array, 0); | |
3303 end = RegExpImpl::GetCapture(match_info_array, 1); | |
3182 } | 3304 } |
3183 | 3305 |
3184 int start = current_match[0]; | 3306 bool global = regexp_handle->GetFlags().is_global(); |
3185 int end = current_match[1]; | |
3186 int capture_count = regexp->CaptureCount(); | |
3187 int subject_length = subject->length(); | |
3188 | 3307 |
3189 int new_length = subject_length - (end - start); | 3308 if (start == end && !global) return *subject_handle; |
3190 if (new_length == 0) return isolate->heap()->empty_string(); | |
3191 | 3309 |
3310 int length = subject_handle->length(); | |
3311 int new_length = length - (end - start); | |
3312 if (new_length == 0) { | |
3313 return isolate->heap()->empty_string(); | |
3314 } | |
3192 Handle<ResultSeqString> answer; | 3315 Handle<ResultSeqString> answer; |
3193 if (ResultSeqString::kHasAsciiEncoding) { | 3316 if (ResultSeqString::kHasAsciiEncoding) { |
3194 answer = Handle<ResultSeqString>::cast( | 3317 answer = Handle<ResultSeqString>::cast( |
3195 isolate->factory()->NewRawAsciiString(new_length)); | 3318 isolate->factory()->NewRawAsciiString(new_length)); |
3196 } else { | 3319 } else { |
3197 answer = Handle<ResultSeqString>::cast( | 3320 answer = Handle<ResultSeqString>::cast( |
3198 isolate->factory()->NewRawTwoByteString(new_length)); | 3321 isolate->factory()->NewRawTwoByteString(new_length)); |
3199 } | 3322 } |
3200 | 3323 |
3201 if (!is_global) { | 3324 // If the regexp isn't global, only match once. |
3202 RegExpImpl::SetLastMatchInfo( | 3325 if (!global) { |
3203 last_match_info, subject, capture_count, current_match); | 3326 if (start > 0) { |
3204 if (start == end) { | 3327 String::WriteToFlat(*subject_handle, |
3205 return *subject; | 3328 answer->GetChars(), |
3206 } else { | 3329 0, |
3207 if (start > 0) { | 3330 start); |
3208 String::WriteToFlat(*subject, answer->GetChars(), 0, start); | |
3209 } | |
3210 if (end < subject_length) { | |
3211 String::WriteToFlat( | |
3212 *subject, answer->GetChars() + start, end, subject_length); | |
3213 } | |
3214 return *answer; | |
3215 } | 3331 } |
3332 if (end < length) { | |
3333 String::WriteToFlat(*subject_handle, | |
3334 answer->GetChars() + start, | |
3335 end, | |
3336 length); | |
3337 } | |
3338 return *answer; | |
3216 } | 3339 } |
3217 | 3340 |
3218 int prev = 0; | 3341 int prev = 0; // Index of end of last match. |
3342 int next = 0; // Start of next search (prev unless last match was empty). | |
3219 int position = 0; | 3343 int position = 0; |
3220 | 3344 |
3221 do { | 3345 do { |
3222 start = current_match[0]; | |
3223 end = current_match[1]; | |
3224 if (prev < start) { | 3346 if (prev < start) { |
3225 // Add substring subject[prev;start] to answer string. | 3347 // Add substring subject[prev;start] to answer string. |
3226 String::WriteToFlat( | 3348 String::WriteToFlat(*subject_handle, |
3227 *subject, answer->GetChars() + position, prev, start); | 3349 answer->GetChars() + position, |
3350 prev, | |
3351 start); | |
3228 position += start - prev; | 3352 position += start - prev; |
3229 } | 3353 } |
3230 prev = end; | 3354 prev = end; |
3355 next = end; | |
3356 // Continue from where the match ended, unless it was an empty match. | |
3357 if (start == end) { | |
3358 next++; | |
3359 if (next > length) break; | |
3360 } | |
3361 match = RegExpImpl::Exec(regexp_handle, | |
3362 subject_handle, | |
3363 next, | |
3364 last_match_info_handle); | |
3365 if (match.is_null()) return Failure::Exception(); | |
3366 if (match->IsNull()) break; | |
3231 | 3367 |
3232 current_match = global_cache.FetchNext(); | 3368 ASSERT(last_match_info_handle->HasFastObjectElements()); |
3233 } while (current_match != NULL); | 3369 HandleScope loop_scope(isolate); |
3370 { | |
3371 AssertNoAllocation match_info_array_is_not_in_a_handle; | |
3372 FixedArray* match_info_array = | |
3373 FixedArray::cast(last_match_info_handle->elements()); | |
3374 start = RegExpImpl::GetCapture(match_info_array, 0); | |
3375 end = RegExpImpl::GetCapture(match_info_array, 1); | |
3376 } | |
3377 } while (true); | |
3234 | 3378 |
3235 if (global_cache.HasException()) return Failure::Exception(); | 3379 if (prev < length) { |
3236 | |
3237 RegExpImpl::SetLastMatchInfo(last_match_info, | |
3238 subject, | |
3239 capture_count, | |
3240 global_cache.LastSuccessfulMatch()); | |
3241 | |
3242 if (prev < subject_length) { | |
3243 // Add substring subject[prev;length] to answer string. | 3380 // Add substring subject[prev;length] to answer string. |
3244 String::WriteToFlat( | 3381 String::WriteToFlat(*subject_handle, |
3245 *subject, answer->GetChars() + position, prev, subject_length); | 3382 answer->GetChars() + position, |
3246 position += subject_length - prev; | 3383 prev, |
3384 length); | |
3385 position += length - prev; | |
3247 } | 3386 } |
3248 | 3387 |
3249 if (position == 0) return isolate->heap()->empty_string(); | 3388 if (position == 0) { |
3389 return isolate->heap()->empty_string(); | |
3390 } | |
3250 | 3391 |
3251 // Shorten string and fill | 3392 // Shorten string and fill |
3252 int string_size = ResultSeqString::SizeFor(position); | 3393 int string_size = ResultSeqString::SizeFor(position); |
3253 int allocated_string_size = ResultSeqString::SizeFor(new_length); | 3394 int allocated_string_size = ResultSeqString::SizeFor(new_length); |
3254 int delta = allocated_string_size - string_size; | 3395 int delta = allocated_string_size - string_size; |
3255 | 3396 |
3256 answer->set_length(position); | 3397 answer->set_length(position); |
3257 if (delta == 0) return *answer; | 3398 if (delta == 0) return *answer; |
3258 | 3399 |
3259 Address end_of_string = answer->address() + string_size; | 3400 Address end_of_string = answer->address() + string_size; |
3260 isolate->heap()->CreateFillerObjectAt(end_of_string, delta); | 3401 isolate->heap()->CreateFillerObjectAt(end_of_string, delta); |
3261 if (Marking::IsBlack(Marking::MarkBitFrom(*answer))) { | 3402 if (Marking::IsBlack(Marking::MarkBitFrom(*answer))) { |
3262 MemoryChunk::IncrementLiveBytesFromMutator(answer->address(), -delta); | 3403 MemoryChunk::IncrementLiveBytesFromMutator(answer->address(), -delta); |
3263 } | 3404 } |
3264 | 3405 |
3265 return *answer; | 3406 return *answer; |
3266 } | 3407 } |
3267 | 3408 |
3268 | 3409 |
3269 RUNTIME_FUNCTION(MaybeObject*, Runtime_StringReplaceRegExpWithString) { | 3410 RUNTIME_FUNCTION(MaybeObject*, Runtime_StringReplaceRegExpWithString) { |
3270 ASSERT(args.length() == 4); | 3411 ASSERT(args.length() == 4); |
3271 | 3412 |
3272 HandleScope scope(isolate); | 3413 CONVERT_ARG_CHECKED(String, subject, 0); |
3414 if (!subject->IsFlat()) { | |
3415 Object* flat_subject; | |
3416 { MaybeObject* maybe_flat_subject = subject->TryFlatten(); | |
3417 if (!maybe_flat_subject->ToObject(&flat_subject)) { | |
3418 return maybe_flat_subject; | |
3419 } | |
3420 } | |
3421 subject = String::cast(flat_subject); | |
3422 } | |
3273 | 3423 |
3274 CONVERT_ARG_HANDLE_CHECKED(String, subject, 0); | 3424 CONVERT_ARG_CHECKED(String, replacement, 2); |
3275 CONVERT_ARG_HANDLE_CHECKED(String, replacement, 2); | 3425 if (!replacement->IsFlat()) { |
3276 CONVERT_ARG_HANDLE_CHECKED(JSRegExp, regexp, 1); | 3426 Object* flat_replacement; |
3277 CONVERT_ARG_HANDLE_CHECKED(JSArray, last_match_info, 3); | 3427 { MaybeObject* maybe_flat_replacement = replacement->TryFlatten(); |
3428 if (!maybe_flat_replacement->ToObject(&flat_replacement)) { | |
3429 return maybe_flat_replacement; | |
3430 } | |
3431 } | |
3432 replacement = String::cast(flat_replacement); | |
3433 } | |
3278 | 3434 |
3279 if (!subject->IsFlat()) subject = FlattenGetString(subject); | 3435 CONVERT_ARG_CHECKED(JSRegExp, regexp, 1); |
3280 | 3436 CONVERT_ARG_CHECKED(JSArray, last_match_info, 3); |
3281 if (!replacement->IsFlat()) replacement = FlattenGetString(replacement); | |
3282 | 3437 |
3283 ASSERT(last_match_info->HasFastObjectElements()); | 3438 ASSERT(last_match_info->HasFastObjectElements()); |
3284 | 3439 |
3440 Zone* zone = isolate->runtime_zone(); | |
3285 if (replacement->length() == 0) { | 3441 if (replacement->length() == 0) { |
3286 if (subject->HasOnlyAsciiChars()) { | 3442 if (subject->HasOnlyAsciiChars()) { |
3287 return StringReplaceRegExpWithEmptyString<SeqAsciiString>( | 3443 return StringReplaceRegExpWithEmptyString<SeqAsciiString>( |
3288 isolate, subject, regexp, last_match_info); | 3444 isolate, subject, regexp, last_match_info, zone); |
3289 } else { | 3445 } else { |
3290 return StringReplaceRegExpWithEmptyString<SeqTwoByteString>( | 3446 return StringReplaceRegExpWithEmptyString<SeqTwoByteString>( |
3291 isolate, subject, regexp, last_match_info); | 3447 isolate, subject, regexp, last_match_info, zone); |
3292 } | 3448 } |
3293 } | 3449 } |
3294 | 3450 |
3295 return StringReplaceRegExpWithString( | 3451 return StringReplaceRegExpWithString(isolate, |
3296 isolate, subject, regexp, replacement, last_match_info); | 3452 subject, |
3453 regexp, | |
3454 replacement, | |
3455 last_match_info, | |
3456 zone); | |
3297 } | 3457 } |
3298 | 3458 |
3299 | 3459 |
3300 Handle<String> Runtime::StringReplaceOneCharWithString(Isolate* isolate, | 3460 Handle<String> Runtime::StringReplaceOneCharWithString(Isolate* isolate, |
3301 Handle<String> subject, | 3461 Handle<String> subject, |
3302 Handle<String> search, | 3462 Handle<String> search, |
3303 Handle<String> replace, | 3463 Handle<String> replace, |
3304 bool* found, | 3464 bool* found, |
3305 int recursion_limit) { | 3465 int recursion_limit) { |
3306 if (recursion_limit == 0) return Handle<String>::null(); | 3466 if (recursion_limit == 0) return Handle<String>::null(); |
(...skipping 302 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
3609 | 3769 |
3610 | 3770 |
3611 RUNTIME_FUNCTION(MaybeObject*, Runtime_StringMatch) { | 3771 RUNTIME_FUNCTION(MaybeObject*, Runtime_StringMatch) { |
3612 ASSERT_EQ(3, args.length()); | 3772 ASSERT_EQ(3, args.length()); |
3613 | 3773 |
3614 CONVERT_ARG_HANDLE_CHECKED(String, subject, 0); | 3774 CONVERT_ARG_HANDLE_CHECKED(String, subject, 0); |
3615 CONVERT_ARG_HANDLE_CHECKED(JSRegExp, regexp, 1); | 3775 CONVERT_ARG_HANDLE_CHECKED(JSRegExp, regexp, 1); |
3616 CONVERT_ARG_HANDLE_CHECKED(JSArray, regexp_info, 2); | 3776 CONVERT_ARG_HANDLE_CHECKED(JSArray, regexp_info, 2); |
3617 HandleScope handles; | 3777 HandleScope handles; |
3618 | 3778 |
3619 RegExpImpl::GlobalCache global_cache(regexp, subject, true, isolate); | 3779 Handle<Object> match = RegExpImpl::Exec(regexp, subject, 0, regexp_info); |
3620 if (global_cache.HasException()) return Failure::Exception(); | |
3621 | 3780 |
3622 int capture_count = regexp->CaptureCount(); | 3781 if (match.is_null()) { |
3782 return Failure::Exception(); | |
3783 } | |
3784 if (match->IsNull()) { | |
3785 return isolate->heap()->null_value(); | |
3786 } | |
3787 int length = subject->length(); | |
3623 | 3788 |
3624 Zone* zone = isolate->runtime_zone(); | 3789 Zone* zone = isolate->runtime_zone(); |
3625 ZoneScope zone_space(zone, DELETE_ON_EXIT); | 3790 ZoneScope zone_space(zone, DELETE_ON_EXIT); |
3626 ZoneList<int> offsets(8, zone); | 3791 ZoneList<int> offsets(8, zone); |
3627 | 3792 int start; |
3628 while (true) { | 3793 int end; |
3629 int32_t* match = global_cache.FetchNext(); | 3794 do { |
3630 if (match == NULL) break; | 3795 { |
3631 offsets.Add(match[0], zone); // start | 3796 AssertNoAllocation no_alloc; |
3632 offsets.Add(match[1], zone); // end | 3797 FixedArray* elements = FixedArray::cast(regexp_info->elements()); |
3633 } | 3798 start = Smi::cast(elements->get(RegExpImpl::kFirstCapture))->value(); |
3634 | 3799 end = Smi::cast(elements->get(RegExpImpl::kFirstCapture + 1))->value(); |
3635 if (global_cache.HasException()) return Failure::Exception(); | 3800 } |
3636 | 3801 offsets.Add(start, zone); |
3637 if (offsets.length() == 0) { | 3802 offsets.Add(end, zone); |
3638 // Not a single match. | 3803 if (start == end) if (++end > length) break; |
3639 return isolate->heap()->null_value(); | 3804 match = RegExpImpl::Exec(regexp, subject, end, regexp_info); |
3640 } | 3805 if (match.is_null()) { |
3641 | 3806 return Failure::Exception(); |
3642 RegExpImpl::SetLastMatchInfo(regexp_info, | 3807 } |
3643 subject, | 3808 } while (!match->IsNull()); |
3644 capture_count, | |
3645 global_cache.LastSuccessfulMatch()); | |
3646 | |
3647 int matches = offsets.length() / 2; | 3809 int matches = offsets.length() / 2; |
3648 Handle<FixedArray> elements = isolate->factory()->NewFixedArray(matches); | 3810 Handle<FixedArray> elements = isolate->factory()->NewFixedArray(matches); |
3649 Handle<String> substring = | 3811 Handle<String> substring = isolate->factory()-> |
3650 isolate->factory()->NewSubString(subject, offsets.at(0), offsets.at(1)); | 3812 NewSubString(subject, offsets.at(0), offsets.at(1)); |
3651 elements->set(0, *substring); | 3813 elements->set(0, *substring); |
3652 for (int i = 1; i < matches; i++) { | 3814 for (int i = 1; i < matches ; i++) { |
3653 HandleScope temp_scope(isolate); | |
3654 int from = offsets.at(i * 2); | 3815 int from = offsets.at(i * 2); |
3655 int to = offsets.at(i * 2 + 1); | 3816 int to = offsets.at(i * 2 + 1); |
3656 Handle<String> substring = | 3817 Handle<String> substring = isolate->factory()-> |
3657 isolate->factory()->NewProperSubString(subject, from, to); | 3818 NewProperSubString(subject, from, to); |
3658 elements->set(i, *substring); | 3819 elements->set(i, *substring); |
3659 } | 3820 } |
3660 Handle<JSArray> result = isolate->factory()->NewJSArrayWithElements(elements); | 3821 Handle<JSArray> result = isolate->factory()->NewJSArrayWithElements(elements); |
3661 result->set_length(Smi::FromInt(matches)); | 3822 result->set_length(Smi::FromInt(matches)); |
3662 return *result; | 3823 return *result; |
3663 } | 3824 } |
3664 | 3825 |
3665 | 3826 |
3666 // Only called from Runtime_RegExpExecMultiple so it doesn't need to maintain | 3827 static bool SearchStringMultiple(Isolate* isolate, |
3667 // separate last match info. See comment on that function. | 3828 Handle<String> subject, |
3668 template<bool has_capture> | 3829 Handle<String> pattern, |
3669 static int SearchRegExpMultiple( | 3830 Handle<JSArray> last_match_info, |
3831 FixedArrayBuilder* builder) { | |
3832 ASSERT(subject->IsFlat()); | |
3833 ASSERT(pattern->IsFlat()); | |
3834 | |
3835 // Treating as if a previous match was before first character. | |
3836 int match_pos = -pattern->length(); | |
3837 | |
3838 for (;;) { // Break when search complete. | |
3839 builder->EnsureCapacity(kMaxBuilderEntriesPerRegExpMatch); | |
3840 AssertNoAllocation no_gc; | |
3841 String::FlatContent subject_content = subject->GetFlatContent(); | |
3842 String::FlatContent pattern_content = pattern->GetFlatContent(); | |
3843 if (subject_content.IsAscii()) { | |
3844 Vector<const char> subject_vector = subject_content.ToAsciiVector(); | |
3845 if (pattern_content.IsAscii()) { | |
3846 if (SearchStringMultiple(isolate, | |
3847 subject_vector, | |
3848 pattern_content.ToAsciiVector(), | |
3849 *pattern, | |
3850 builder, | |
3851 &match_pos)) break; | |
3852 } else { | |
3853 if (SearchStringMultiple(isolate, | |
3854 subject_vector, | |
3855 pattern_content.ToUC16Vector(), | |
3856 *pattern, | |
3857 builder, | |
3858 &match_pos)) break; | |
3859 } | |
3860 } else { | |
3861 Vector<const uc16> subject_vector = subject_content.ToUC16Vector(); | |
3862 if (pattern_content.IsAscii()) { | |
3863 if (SearchStringMultiple(isolate, | |
3864 subject_vector, | |
3865 pattern_content.ToAsciiVector(), | |
3866 *pattern, | |
3867 builder, | |
3868 &match_pos)) break; | |
3869 } else { | |
3870 if (SearchStringMultiple(isolate, | |
3871 subject_vector, | |
3872 pattern_content.ToUC16Vector(), | |
3873 *pattern, | |
3874 builder, | |
3875 &match_pos)) break; | |
3876 } | |
3877 } | |
3878 } | |
3879 | |
3880 if (match_pos >= 0) { | |
3881 SetLastMatchInfoNoCaptures(subject, | |
3882 last_match_info, | |
3883 match_pos, | |
3884 match_pos + pattern->length()); | |
3885 return true; | |
3886 } | |
3887 return false; // No matches at all. | |
3888 } | |
3889 | |
3890 | |
3891 static int SearchRegExpNoCaptureMultiple( | |
3670 Isolate* isolate, | 3892 Isolate* isolate, |
3671 Handle<String> subject, | 3893 Handle<String> subject, |
3672 Handle<JSRegExp> regexp, | 3894 Handle<JSRegExp> regexp, |
3673 Handle<JSArray> last_match_array, | 3895 Handle<JSArray> last_match_array, |
3674 FixedArrayBuilder* builder) { | 3896 FixedArrayBuilder* builder) { |
3675 ASSERT(subject->IsFlat()); | 3897 ASSERT(subject->IsFlat()); |
3676 ASSERT_NE(has_capture, regexp->CaptureCount() == 0); | 3898 ASSERT(regexp->CaptureCount() == 0); |
3677 | |
3678 RegExpImpl::GlobalCache global_cache(regexp, subject, true, isolate); | |
3679 if (global_cache.HasException()) return RegExpImpl::RE_EXCEPTION; | |
3680 | |
3681 int capture_count = regexp->CaptureCount(); | |
3682 int subject_length = subject->length(); | |
3683 | |
3684 // Position to search from. | |
3685 int match_start = -1; | 3899 int match_start = -1; |
3686 int match_end = 0; | 3900 int match_end = 0; |
3901 int pos = 0; | |
3902 int registers_per_match = RegExpImpl::IrregexpPrepare(regexp, subject); | |
3903 if (registers_per_match < 0) return RegExpImpl::RE_EXCEPTION; | |
3904 | |
3905 int max_matches; | |
3906 int num_registers = RegExpImpl::GlobalOffsetsVectorSize(regexp, | |
3907 registers_per_match, | |
3908 &max_matches); | |
3909 OffsetsVector registers(num_registers, isolate); | |
3910 Vector<int32_t> register_vector(registers.vector(), registers.length()); | |
3911 int subject_length = subject->length(); | |
3687 bool first = true; | 3912 bool first = true; |
3688 | 3913 for (;;) { // Break on failure, return on exception. |
3689 // Two smis before and after the match, for very long strings. | 3914 int num_matches = RegExpImpl::IrregexpExecRaw(regexp, |
3690 static const int kMaxBuilderEntriesPerRegExpMatch = 5; | 3915 subject, |
3691 | 3916 pos, |
3692 while (true) { | 3917 register_vector); |
3693 int32_t* current_match = global_cache.FetchNext(); | 3918 if (num_matches > 0) { |
3694 if (current_match == NULL) break; | 3919 for (int match_index = 0; match_index < num_matches; match_index++) { |
3695 match_start = current_match[0]; | 3920 int32_t* current_match = ®ister_vector[match_index * 2]; |
3696 builder->EnsureCapacity(kMaxBuilderEntriesPerRegExpMatch); | 3921 match_start = current_match[0]; |
3697 if (match_end < match_start) { | 3922 builder->EnsureCapacity(kMaxBuilderEntriesPerRegExpMatch); |
3698 ReplacementStringBuilder::AddSubjectSlice(builder, | 3923 if (match_end < match_start) { |
3699 match_end, | 3924 ReplacementStringBuilder::AddSubjectSlice(builder, |
3700 match_start); | 3925 match_end, |
3701 } | 3926 match_start); |
3702 match_end = current_match[1]; | 3927 } |
3703 { | 3928 match_end = current_match[1]; |
3704 // Avoid accumulating new handles inside loop. | 3929 HandleScope loop_scope(isolate); |
3705 HandleScope temp_scope(isolate); | 3930 if (!first) { |
3706 Handle<String> match; | 3931 builder->Add(*isolate->factory()->NewProperSubString(subject, |
3707 if (!first) { | 3932 match_start, |
3708 match = isolate->factory()->NewProperSubString(subject, | 3933 match_end)); |
3709 match_start, | 3934 } else { |
3710 match_end); | 3935 builder->Add(*isolate->factory()->NewSubString(subject, |
3711 } else { | 3936 match_start, |
3712 match = isolate->factory()->NewSubString(subject, | 3937 match_end)); |
3713 match_start, | 3938 first = false; |
3714 match_end); | 3939 } |
3715 first = false; | |
3716 } | 3940 } |
3717 | 3941 |
3718 if (has_capture) { | 3942 // If we did not get the maximum number of matches, we can stop here |
3719 // Arguments array to replace function is match, captures, index and | 3943 // since there are no matches left. |
3720 // subject, i.e., 3 + capture count in total. | 3944 if (num_matches < max_matches) break; |
3721 Handle<FixedArray> elements = | |
3722 isolate->factory()->NewFixedArray(3 + capture_count); | |
3723 | 3945 |
3724 elements->set(0, *match); | 3946 if (match_start != match_end) { |
3725 for (int i = 1; i <= capture_count; i++) { | 3947 pos = match_end; |
3726 int start = current_match[i * 2]; | |
3727 if (start >= 0) { | |
3728 int end = current_match[i * 2 + 1]; | |
3729 ASSERT(start <= end); | |
3730 Handle<String> substring = | |
3731 isolate->factory()->NewSubString(subject, start, end); | |
3732 elements->set(i, *substring); | |
3733 } else { | |
3734 ASSERT(current_match[i * 2 + 1] < 0); | |
3735 elements->set(i, isolate->heap()->undefined_value()); | |
3736 } | |
3737 } | |
3738 elements->set(capture_count + 1, Smi::FromInt(match_start)); | |
3739 elements->set(capture_count + 2, *subject); | |
3740 builder->Add(*isolate->factory()->NewJSArrayWithElements(elements)); | |
3741 } else { | 3948 } else { |
3742 builder->Add(*match); | 3949 pos = match_end + 1; |
3950 if (pos > subject_length) break; | |
3743 } | 3951 } |
3952 } else if (num_matches == 0) { | |
3953 break; | |
3954 } else { | |
3955 ASSERT_EQ(num_matches, RegExpImpl::RE_EXCEPTION); | |
3956 return RegExpImpl::RE_EXCEPTION; | |
3744 } | 3957 } |
3745 } | 3958 } |
3746 | 3959 |
3747 if (global_cache.HasException()) return RegExpImpl::RE_EXCEPTION; | |
3748 | |
3749 if (match_start >= 0) { | 3960 if (match_start >= 0) { |
3750 // Finished matching, with at least one match. | |
3751 if (match_end < subject_length) { | 3961 if (match_end < subject_length) { |
3752 ReplacementStringBuilder::AddSubjectSlice(builder, | 3962 ReplacementStringBuilder::AddSubjectSlice(builder, |
3753 match_end, | 3963 match_end, |
3754 subject_length); | 3964 subject_length); |
3755 } | 3965 } |
3756 | 3966 SetLastMatchInfoNoCaptures(subject, |
3757 RegExpImpl::SetLastMatchInfo( | 3967 last_match_array, |
3758 last_match_array, subject, capture_count, NULL); | 3968 match_start, |
3759 | 3969 match_end); |
3760 return RegExpImpl::RE_SUCCESS; | 3970 return RegExpImpl::RE_SUCCESS; |
3761 } else { | 3971 } else { |
3762 return RegExpImpl::RE_FAILURE; // No matches at all. | 3972 return RegExpImpl::RE_FAILURE; // No matches at all. |
3763 } | 3973 } |
3764 } | 3974 } |
3765 | 3975 |
3766 | 3976 |
3977 // Only called from Runtime_RegExpExecMultiple so it doesn't need to maintain | |
3978 // separate last match info. See comment on that function. | |
3979 static int SearchRegExpMultiple( | |
3980 Isolate* isolate, | |
3981 Handle<String> subject, | |
3982 Handle<JSRegExp> regexp, | |
3983 Handle<JSArray> last_match_array, | |
3984 FixedArrayBuilder* builder, | |
3985 Zone* zone) { | |
3986 | |
3987 ASSERT(subject->IsFlat()); | |
3988 int registers_per_match = RegExpImpl::IrregexpPrepare(regexp, subject); | |
3989 if (registers_per_match < 0) return RegExpImpl::RE_EXCEPTION; | |
3990 | |
3991 int max_matches; | |
3992 int num_registers = RegExpImpl::GlobalOffsetsVectorSize(regexp, | |
3993 registers_per_match, | |
3994 &max_matches); | |
3995 OffsetsVector registers(num_registers, isolate); | |
3996 Vector<int32_t> register_vector(registers.vector(), registers.length()); | |
3997 | |
3998 int num_matches = RegExpImpl::IrregexpExecRaw(regexp, | |
3999 subject, | |
4000 0, | |
4001 register_vector); | |
4002 | |
4003 int capture_count = regexp->CaptureCount(); | |
4004 int subject_length = subject->length(); | |
4005 | |
4006 // Position to search from. | |
4007 int pos = 0; | |
4008 // End of previous match. Differs from pos if match was empty. | |
4009 int match_end = 0; | |
4010 bool first = true; | |
4011 | |
4012 if (num_matches > 0) { | |
4013 do { | |
4014 int match_start = 0; | |
4015 for (int match_index = 0; match_index < num_matches; match_index++) { | |
4016 int32_t* current_match = | |
4017 ®ister_vector[match_index * registers_per_match]; | |
4018 match_start = current_match[0]; | |
4019 builder->EnsureCapacity(kMaxBuilderEntriesPerRegExpMatch); | |
4020 if (match_end < match_start) { | |
4021 ReplacementStringBuilder::AddSubjectSlice(builder, | |
4022 match_end, | |
4023 match_start); | |
4024 } | |
4025 match_end = current_match[1]; | |
4026 | |
4027 { | |
4028 // Avoid accumulating new handles inside loop. | |
4029 HandleScope temp_scope(isolate); | |
4030 // Arguments array to replace function is match, captures, index and | |
4031 // subject, i.e., 3 + capture count in total. | |
4032 Handle<FixedArray> elements = | |
4033 isolate->factory()->NewFixedArray(3 + capture_count); | |
4034 Handle<String> match; | |
4035 if (!first) { | |
4036 match = isolate->factory()->NewProperSubString(subject, | |
4037 match_start, | |
4038 match_end); | |
4039 } else { | |
4040 match = isolate->factory()->NewSubString(subject, | |
4041 match_start, | |
4042 match_end); | |
4043 } | |
4044 elements->set(0, *match); | |
4045 for (int i = 1; i <= capture_count; i++) { | |
4046 int start = current_match[i * 2]; | |
4047 if (start >= 0) { | |
4048 int end = current_match[i * 2 + 1]; | |
4049 ASSERT(start <= end); | |
4050 Handle<String> substring; | |
4051 if (!first) { | |
4052 substring = | |
4053 isolate->factory()->NewProperSubString(subject, start, end); | |
4054 } else { | |
4055 substring = | |
4056 isolate->factory()->NewSubString(subject, start, end); | |
4057 } | |
4058 elements->set(i, *substring); | |
4059 } else { | |
4060 ASSERT(current_match[i * 2 + 1] < 0); | |
4061 elements->set(i, isolate->heap()->undefined_value()); | |
4062 } | |
4063 } | |
4064 elements->set(capture_count + 1, Smi::FromInt(match_start)); | |
4065 elements->set(capture_count + 2, *subject); | |
4066 builder->Add(*isolate->factory()->NewJSArrayWithElements(elements)); | |
4067 } | |
4068 first = false; | |
4069 } | |
4070 | |
4071 // If we did not get the maximum number of matches, we can stop here | |
4072 // since there are no matches left. | |
4073 if (num_matches < max_matches) break; | |
4074 | |
4075 if (match_end > match_start) { | |
4076 pos = match_end; | |
4077 } else { | |
4078 pos = match_end + 1; | |
4079 if (pos > subject_length) { | |
4080 break; | |
4081 } | |
4082 } | |
4083 | |
4084 num_matches = RegExpImpl::IrregexpExecRaw(regexp, | |
4085 subject, | |
4086 pos, | |
4087 register_vector); | |
4088 } while (num_matches > 0); | |
4089 | |
4090 if (num_matches != RegExpImpl::RE_EXCEPTION) { | |
4091 // Finished matching, with at least one match. | |
4092 if (match_end < subject_length) { | |
4093 ReplacementStringBuilder::AddSubjectSlice(builder, | |
4094 match_end, | |
4095 subject_length); | |
4096 } | |
4097 | |
4098 int last_match_capture_count = (capture_count + 1) * 2; | |
4099 int last_match_array_size = | |
4100 last_match_capture_count + RegExpImpl::kLastMatchOverhead; | |
4101 last_match_array->EnsureSize(last_match_array_size); | |
4102 AssertNoAllocation no_gc; | |
4103 FixedArray* elements = FixedArray::cast(last_match_array->elements()); | |
4104 // We have to set this even though the rest of the last match array is | |
4105 // ignored. | |
4106 RegExpImpl::SetLastCaptureCount(elements, last_match_capture_count); | |
4107 // These are also read without consulting the override. | |
4108 RegExpImpl::SetLastSubject(elements, *subject); | |
4109 RegExpImpl::SetLastInput(elements, *subject); | |
4110 return RegExpImpl::RE_SUCCESS; | |
4111 } | |
4112 } | |
4113 // No matches at all, return failure or exception result directly. | |
4114 return num_matches; | |
4115 } | |
4116 | |
4117 | |
3767 // This is only called for StringReplaceGlobalRegExpWithFunction. This sets | 4118 // This is only called for StringReplaceGlobalRegExpWithFunction. This sets |
3768 // lastMatchInfoOverride to maintain the last match info, so we don't need to | 4119 // lastMatchInfoOverride to maintain the last match info, so we don't need to |
3769 // set any other last match array info. | 4120 // set any other last match array info. |
3770 RUNTIME_FUNCTION(MaybeObject*, Runtime_RegExpExecMultiple) { | 4121 RUNTIME_FUNCTION(MaybeObject*, Runtime_RegExpExecMultiple) { |
3771 ASSERT(args.length() == 4); | 4122 ASSERT(args.length() == 4); |
3772 HandleScope handles(isolate); | 4123 HandleScope handles(isolate); |
3773 | 4124 |
3774 CONVERT_ARG_HANDLE_CHECKED(String, subject, 1); | 4125 CONVERT_ARG_HANDLE_CHECKED(String, subject, 1); |
3775 if (!subject->IsFlat()) FlattenString(subject); | 4126 if (!subject->IsFlat()) FlattenString(subject); |
3776 CONVERT_ARG_HANDLE_CHECKED(JSRegExp, regexp, 0); | 4127 CONVERT_ARG_HANDLE_CHECKED(JSRegExp, regexp, 0); |
3777 CONVERT_ARG_HANDLE_CHECKED(JSArray, last_match_info, 2); | 4128 CONVERT_ARG_HANDLE_CHECKED(JSArray, last_match_info, 2); |
3778 CONVERT_ARG_HANDLE_CHECKED(JSArray, result_array, 3); | 4129 CONVERT_ARG_HANDLE_CHECKED(JSArray, result_array, 3); |
3779 | 4130 |
3780 ASSERT(last_match_info->HasFastObjectElements()); | 4131 ASSERT(last_match_info->HasFastObjectElements()); |
3781 ASSERT(regexp->GetFlags().is_global()); | 4132 ASSERT(regexp->GetFlags().is_global()); |
3782 Handle<FixedArray> result_elements; | 4133 Handle<FixedArray> result_elements; |
3783 if (result_array->HasFastObjectElements()) { | 4134 if (result_array->HasFastObjectElements()) { |
3784 result_elements = | 4135 result_elements = |
3785 Handle<FixedArray>(FixedArray::cast(result_array->elements())); | 4136 Handle<FixedArray>(FixedArray::cast(result_array->elements())); |
3786 } | 4137 } |
3787 if (result_elements.is_null() || result_elements->length() < 16) { | 4138 if (result_elements.is_null() || result_elements->length() < 16) { |
3788 result_elements = isolate->factory()->NewFixedArrayWithHoles(16); | 4139 result_elements = isolate->factory()->NewFixedArrayWithHoles(16); |
3789 } | 4140 } |
3790 FixedArrayBuilder builder(result_elements); | 4141 FixedArrayBuilder builder(result_elements); |
3791 | 4142 |
4143 if (regexp->TypeTag() == JSRegExp::ATOM) { | |
4144 Handle<String> pattern( | |
4145 String::cast(regexp->DataAt(JSRegExp::kAtomPatternIndex))); | |
4146 ASSERT(pattern->IsFlat()); | |
4147 if (SearchStringMultiple(isolate, subject, pattern, | |
4148 last_match_info, &builder)) { | |
4149 return *builder.ToJSArray(result_array); | |
4150 } | |
4151 return isolate->heap()->null_value(); | |
4152 } | |
4153 | |
4154 ASSERT_EQ(regexp->TypeTag(), JSRegExp::IRREGEXP); | |
4155 | |
3792 int result; | 4156 int result; |
3793 if (regexp->CaptureCount() == 0) { | 4157 if (regexp->CaptureCount() == 0) { |
3794 result = SearchRegExpMultiple<false>( | 4158 result = SearchRegExpNoCaptureMultiple(isolate, |
3795 isolate, subject, regexp, last_match_info, &builder); | 4159 subject, |
4160 regexp, | |
4161 last_match_info, | |
4162 &builder); | |
3796 } else { | 4163 } else { |
3797 result = SearchRegExpMultiple<true>( | 4164 result = SearchRegExpMultiple(isolate, |
3798 isolate, subject, regexp, last_match_info, &builder); | 4165 subject, |
4166 regexp, | |
4167 last_match_info, | |
4168 &builder, | |
4169 isolate->runtime_zone()); | |
3799 } | 4170 } |
3800 | |
3801 if (result == RegExpImpl::RE_SUCCESS) return *builder.ToJSArray(result_array); | 4171 if (result == RegExpImpl::RE_SUCCESS) return *builder.ToJSArray(result_array); |
3802 if (result == RegExpImpl::RE_FAILURE) return isolate->heap()->null_value(); | 4172 if (result == RegExpImpl::RE_FAILURE) return isolate->heap()->null_value(); |
3803 ASSERT_EQ(result, RegExpImpl::RE_EXCEPTION); | 4173 ASSERT_EQ(result, RegExpImpl::RE_EXCEPTION); |
3804 return Failure::Exception(); | 4174 return Failure::Exception(); |
3805 } | 4175 } |
3806 | 4176 |
3807 | 4177 |
3808 RUNTIME_FUNCTION(MaybeObject*, Runtime_NumberToRadixString) { | 4178 RUNTIME_FUNCTION(MaybeObject*, Runtime_NumberToRadixString) { |
3809 NoHandleAllocation ha; | 4179 NoHandleAllocation ha; |
3810 ASSERT(args.length() == 2); | 4180 ASSERT(args.length() == 2); |
(...skipping 9536 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
13347 // Handle last resort GC and make sure to allow future allocations | 13717 // Handle last resort GC and make sure to allow future allocations |
13348 // to grow the heap without causing GCs (if possible). | 13718 // to grow the heap without causing GCs (if possible). |
13349 isolate->counters()->gc_last_resort_from_js()->Increment(); | 13719 isolate->counters()->gc_last_resort_from_js()->Increment(); |
13350 isolate->heap()->CollectAllGarbage(Heap::kNoGCFlags, | 13720 isolate->heap()->CollectAllGarbage(Heap::kNoGCFlags, |
13351 "Runtime::PerformGC"); | 13721 "Runtime::PerformGC"); |
13352 } | 13722 } |
13353 } | 13723 } |
13354 | 13724 |
13355 | 13725 |
13356 } } // namespace v8::internal | 13726 } } // namespace v8::internal |
OLD | NEW |