// Copyright 2021 the V8 project authors. All rights reserved. // Use of this source code is governed by a BSD-style license that can be // found in the LICENSE file. #include "src/regexp/regexp.h" #include #include #include #include "include/v8-context.h" #include "include/v8-initialization.h" #include "include/v8-isolate.h" #include "include/v8-local-handle.h" #include "src/api/api-inl.h" #include "src/ast/ast.h" #include "src/base/strings.h" #include "src/codegen/assembler-arch.h" #include "src/codegen/macro-assembler.h" #include "src/init/v8.h" #include "src/objects/js-regexp-inl.h" #include "src/objects/objects-inl.h" #include "src/regexp/regexp-ast-printer.h" #include "src/regexp/regexp-bytecode-generator.h" #include "src/regexp/regexp-bytecodes-inl.h" #include "src/regexp/regexp-compiler.h" #include "src/regexp/regexp-interpreter.h" #include "src/regexp/regexp-macro-assembler-arch.h" #include "src/regexp/regexp-parser.h" #include "src/regexp/regexp-stack.h" #include "src/strings/char-predicates-inl.h" #include "src/strings/string-stream.h" #include "src/strings/unicode-inl.h" #include "src/utils/ostreams.h" #include "src/zone/zone-list-inl.h" #include "test/common/flag-utils.h" #include "test/unittests/test-utils.h" namespace v8 { namespace internal { using REB = regexp::Bytecode; namespace { regexp::Bytecode GetBytecode(DirectHandle bytecode_array, int i) { return regexp::Bytecodes::FromByte(bytecode_array->get(i)); } } // namespace TEST_F(TestWithNativeContext, ConvertFlagsToString) { RunJS("let regexp = new RegExp(/ab+c/ig);"); DirectHandle regexp = RunJS("regexp"); DirectHandle flags = RunJS("regexp.flags"); DirectHandle converted_flags = JSRegExp::StringFromFlags(isolate(), regexp->flags()); EXPECT_TRUE(String::Equals(isolate(), flags, converted_flags)); } TEST_F(TestWithNativeContext, ConvertFlagsToStringNoFlags) { RunJS("let regexp = new RegExp(/ab+c/);"); DirectHandle regexp = RunJS("regexp"); DirectHandle flags = RunJS("regexp.flags"); DirectHandle converted_flags = JSRegExp::StringFromFlags(isolate(), regexp->flags()); EXPECT_TRUE(String::Equals(isolate(), flags, converted_flags)); } TEST_F(TestWithNativeContext, ConvertFlagsToStringAllFlags) { RunJS("let regexp = new RegExp(/ab+c/dgimsuy);"); DirectHandle regexp = RunJS("regexp"); DirectHandle flags = RunJS("regexp.flags"); DirectHandle converted_flags = JSRegExp::StringFromFlags(isolate(), regexp->flags()); EXPECT_TRUE(String::Equals(isolate(), flags, converted_flags)); } static DirectHandle CreateRegExpData(Isolate* isolate, DirectHandle source) { // All IrRegExpData field except original_source are unused for these tests. constexpr JSRegExp::Flags flags{}; constexpr int capture_count = 0; constexpr uint32_t backtrack_limit = JSRegExp::kNoBacktrackLimit; constexpr uint32_t bit_field = 0; return isolate->factory()->NewIrRegExpData( source, isolate->factory()->empty_string(), flags, capture_count, backtrack_limit, bit_field); } using RegExpTest = TestWithIsolate; static bool CheckParse(const char* input) { Isolate* isolate = reinterpret_cast(v8::Isolate::GetCurrent()); v8::HandleScope scope(v8::Isolate::GetCurrent()); Zone zone(isolate->allocator(), ZONE_NAME); DirectHandle str = isolate->factory()->NewStringFromAsciiChecked(input); regexp::CompileData result; regexp::Flags flags; return regexp::Parser::ParseRegExpFromHeapString(isolate, &zone, str, flags, &result); } static void CheckParseEq(const char* input, const char* expected, bool unicode = false) { Isolate* isolate = reinterpret_cast(v8::Isolate::GetCurrent()); v8::HandleScope scope(v8::Isolate::GetCurrent()); Zone zone(isolate->allocator(), ZONE_NAME); DirectHandle str = isolate->factory()->NewStringFromAsciiChecked(input); regexp::CompileData result; regexp::Flags flags; if (unicode) flags |= regexp::Flag::kUnicode; CHECK(regexp::Parser::ParseRegExpFromHeapString(isolate, &zone, str, flags, &result)); CHECK_NOT_NULL(result.tree); CHECK_EQ(regexp::Error::kNone, result.error); #ifdef V8_ENABLE_REGEXP_DIAGNOSTICS std::ostringstream os; regexp::AstNodePrinter printer(os, nullptr, &zone); printer.Print(result.tree); if (strcmp(expected, os.str().c_str()) != 0) { printf("%s | %s\n", expected, os.str().c_str()); } CHECK_EQ(0, strcmp(expected, os.str().c_str())); #endif } static bool CheckSimple(const char* input) { Isolate* isolate = reinterpret_cast(v8::Isolate::GetCurrent()); v8::HandleScope scope(v8::Isolate::GetCurrent()); Zone zone(isolate->allocator(), ZONE_NAME); DirectHandle str = isolate->factory()->NewStringFromAsciiChecked(input); regexp::CompileData result; CHECK(regexp::Parser::ParseRegExpFromHeapString(isolate, &zone, str, {}, &result)); CHECK_NOT_NULL(result.tree); CHECK_EQ(regexp::Error::kNone, result.error); return result.simple; } struct MinMaxPair { int min_match; int max_match; }; static MinMaxPair CheckMinMaxMatch(const char* input) { Isolate* isolate = reinterpret_cast(v8::Isolate::GetCurrent()); v8::HandleScope scope(v8::Isolate::GetCurrent()); Zone zone(isolate->allocator(), ZONE_NAME); DirectHandle str = isolate->factory()->NewStringFromAsciiChecked(input); regexp::CompileData result; CHECK(regexp::Parser::ParseRegExpFromHeapString(isolate, &zone, str, {}, &result)); CHECK_NOT_NULL(result.tree); CHECK_EQ(regexp::Error::kNone, result.error); int min_match = result.tree->min_match(); int max_match = result.tree->max_match(); MinMaxPair pair = {min_match, max_match}; return pair; } #define CHECK_PARSE_ERROR(input) CHECK(!CheckParse(input)) #define CHECK_SIMPLE(input, simple) CHECK_EQ(simple, CheckSimple(input)); #define CHECK_MIN_MAX(input, min, max) \ { \ MinMaxPair min_max = CheckMinMaxMatch(input); \ CHECK_EQ(min, min_max.min_match); \ CHECK_EQ(max, min_max.max_match); \ } TEST_F(RegExpTest, Parser) { CHECK_PARSE_ERROR("?"); CheckParseEq("abc", "'abc'"); CheckParseEq("", "%"); CheckParseEq("abc|def", "(| 'abc' 'def')"); CheckParseEq("abc|def|ghi", "(| 'abc' 'def' 'ghi')"); CheckParseEq("^xxx$", "(: @^i 'xxx' @$i)"); CheckParseEq("ab\\b\\d\\bcd", "(: 'ab' @b [0-9] @b 'cd')"); CheckParseEq("\\w|\\d", "(| [0-9 A-Z _ a-z] [0-9])"); CheckParseEq("a*", "(# 0 - g 'a')"); CheckParseEq("a*?", "(# 0 - n 'a')"); CheckParseEq("abc+", "(: 'ab' (# 1 - g 'c'))"); CheckParseEq("abc+?", "(: 'ab' (# 1 - n 'c'))"); CheckParseEq("xyz?", "(: 'xy' (# 0 1 g 'z'))"); CheckParseEq("xyz??", "(: 'xy' (# 0 1 n 'z'))"); CheckParseEq("xyz{0,1}", "(: 'xy' (# 0 1 g 'z'))"); CheckParseEq("xyz{0,1}?", "(: 'xy' (# 0 1 n 'z'))"); CheckParseEq("xyz{93}", "(: 'xy' (# 93 93 g 'z'))"); CheckParseEq("xyz{93}?", "(: 'xy' (# 93 93 n 'z'))"); CheckParseEq("xyz{1,32}", "(: 'xy' (# 1 32 g 'z'))"); CheckParseEq("xyz{1,32}?", "(: 'xy' (# 1 32 n 'z'))"); CheckParseEq("xyz{1,}", "(: 'xy' (# 1 - g 'z'))"); CheckParseEq("xyz{1,}?", "(: 'xy' (# 1 - n 'z'))"); CheckParseEq("a\\fb\\nc\\rd\\te\\vf", "'a\\x0cb\\x0ac\\x0dd\\x09e\\x0bf'"); CheckParseEq("a\\nb\\bc", "(: 'a\\x0ab' @b 'c')"); CheckParseEq("(?:foo)", "(?: 'foo')"); CheckParseEq("(?: foo )", "(?: ' foo ')"); CheckParseEq("(foo|bar|baz)", "(^ (| 'foo' 'bar' 'baz'))"); CheckParseEq("foo|(bar|baz)|quux", "(| 'foo' (^ (| 'bar' 'baz')) 'quux')"); CheckParseEq("foo(?=bar)baz", "(: 'foo' (-> + 'bar') 'baz')"); CheckParseEq("foo(?!bar)baz", "(: 'foo' (-> - 'bar') 'baz')"); CheckParseEq("foo(?<=bar)baz", "(: 'foo' (<- + 'bar') 'baz')"); CheckParseEq("foo(? + %)"); CheckParseEq("[]", "^[\\x00-\\u{10ffff}]"); // Doesn't compile on windows CheckParseEq("[^]", "[\\x00-\\u{10ffff}]"); // \uffff isn't in codepage 1252 CheckParseEq("[x]", "[x]"); CheckParseEq("[xyz]", "[x y z]"); CheckParseEq("[a-zA-Z0-9]", "[a-z A-Z 0-9]"); CheckParseEq("[-123]", "[- 1 2 3]"); CheckParseEq("[^123]", "^[1 2 3]"); CheckParseEq("]", "']'"); CheckParseEq("}", "'}'"); CheckParseEq("[a-b-c]", "[a-b - c]"); CheckParseEq("[\\d]", "[0-9]"); CheckParseEq("[x\\dz]", "[x 0-9 z]"); CheckParseEq("[\\d-z]", "[0-9 - z]"); CheckParseEq("[\\d-\\d]", "[0-9 0-9 -]"); CheckParseEq("[z-\\d]", "[0-9 z -]"); // Control character outside character class. CheckParseEq("\\cj\\cJ\\ci\\cI\\ck\\cK", "'\\x0a\\x0a\\x09\\x09\\x0b\\x0b'"); CheckParseEq("\\c!", "'\\c!'"); CheckParseEq("\\c_", "'\\c_'"); CheckParseEq("\\c~", "'\\c~'"); CheckParseEq("\\c1", "'\\c1'"); // Control character inside character class. CheckParseEq("[\\c!]", "[\\ c !]"); CheckParseEq("[\\c_]", "[\\x1f]"); CheckParseEq("[\\c~]", "[\\ c ~]"); CheckParseEq("[\\ca]", "[\\x01]"); CheckParseEq("[\\cz]", "[\\x1a]"); CheckParseEq("[\\cA]", "[\\x01]"); CheckParseEq("[\\cZ]", "[\\x1a]"); CheckParseEq("[\\c1]", "[\\x11]"); CheckParseEq("[a\\]c]", "[a ] c]"); CheckParseEq("\\[\\]\\{\\}\$\$\\%\\^\\#\\ ", "'[]{}()%^# '"); CheckParseEq("[\\[\\]\\{\\}\$\$\\%\\^\\#\\ ]", "[[ ] { } ( ) % ^ # ]"); CheckParseEq("\\0", "'\\x00'"); CheckParseEq("\\8", "'8'"); CheckParseEq("\\9", "'9'"); CheckParseEq("\\11", "'\\x09'"); CheckParseEq("\\11a", "'\\x09a'"); CheckParseEq("\\011", "'\\x09'"); CheckParseEq("\\00011", "'\\x0011'"); CheckParseEq("\\118", "'\\x098'"); CheckParseEq("\\111", "'I'"); CheckParseEq("\\1111", "'I1'"); CheckParseEq("(x)(x)(x)\\1", "(: (^ 'x') (^ 'x') (^ 'x') (<- 1))"); CheckParseEq("(x)(x)(x)\\2", "(: (^ 'x') (^ 'x') (^ 'x') (<- 2))"); CheckParseEq("(x)(x)(x)\\3", "(: (^ 'x') (^ 'x') (^ 'x') (<- 3))"); CheckParseEq("(x)(x)(x)\\4", "(: (^ 'x') (^ 'x') (^ 'x') '\\x04')"); CheckParseEq("(x)(x)(x)\\1*", "(: (^ 'x') (^ 'x') (^ 'x')" " (# 0 - g (<- 1)))"); CheckParseEq("(x)(x)(x)\\2*", "(: (^ 'x') (^ 'x') (^ 'x')" " (# 0 - g (<- 2)))"); CheckParseEq("(x)(x)(x)\\3*", "(: (^ 'x') (^ 'x') (^ 'x')" " (# 0 - g (<- 3)))"); CheckParseEq("(x)(x)(x)\\4*", "(: (^ 'x') (^ 'x') (^ 'x')" " (# 0 - g '\\x04'))"); CheckParseEq("(x)(x)(x)(x)(x)(x)(x)(x)(x)(x)\\10", "(: (^ 'x') (^ 'x') (^ 'x') (^ 'x') (^ 'x') (^ 'x')" " (^ 'x') (^ 'x') (^ 'x') (^ 'x') (<- 10))"); CheckParseEq("(x)(x)(x)(x)(x)(x)(x)(x)(x)(x)\\11", "(: (^ 'x') (^ 'x') (^ 'x') (^ 'x') (^ 'x') (^ 'x')" " (^ 'x') (^ 'x') (^ 'x') (^ 'x') '\\x09')"); CheckParseEq("(a)\\1", "(: (^ 'a') (<- 1))"); CheckParseEq("(a\\1)", "(^ 'a')"); CheckParseEq("(\\1a)", "(^ 'a')"); CheckParseEq("(\\2)(\\1)", "(: (^ (<- 2)) (^ (<- 1)))"); CheckParseEq("(?=a)?a", "'a'"); CheckParseEq("(?=a){0,10}a", "'a'"); CheckParseEq("(?=a){1,10}a", "(: (-> + 'a') 'a')"); CheckParseEq("(?=a){9,10}a", "(: (-> + 'a') 'a')"); CheckParseEq("(?!a)?a", "'a'"); CheckParseEq("\\1(a)", "(: (<- 1) (^ 'a'))"); CheckParseEq("(?!(a))\\1", "(: (-> - (^ 'a')) (<- 1))"); CheckParseEq("(?!\\1(a\\1)\\1)\\1", "(: (-> - (: (<- 1) (^ 'a') (<- 1))) (<- 1))"); CheckParseEq("\\1\\2(a(?:\\1(b\\1\\2))\\2)\\1", "(: (<- 1) (<- 2) (^ (: 'a' (?: (^ 'b')) (<- 2))) (<- 1))"); CheckParseEq("\\1\\2(a(?<=\\1(b\\1\\2))\\2)\\1", "(: (<- 1) (<- 2) (^ (: 'a' (<- + (^ 'b')) (<- 2))) (<- 1))"); CheckParseEq("[\\0]", "[\\x00]"); CheckParseEq("[\\11]", "[\\x09]"); CheckParseEq("[\\11a]", "[\\x09 a]"); CheckParseEq("[\\011]", "[\\x09]"); CheckParseEq("[\\00011]", "[\\x00 1 1]"); CheckParseEq("[\\118]", "[\\x09 8]"); CheckParseEq("[\\111]", "[I]"); CheckParseEq("[\\1111]", "[I 1]"); CheckParseEq("\\x34", "'\x34'"); CheckParseEq("\\x60", "'\x60'"); CheckParseEq("\\x3z", "'x3z'"); CheckParseEq("\\c", "'\\c'"); CheckParseEq("\\u0034", "'\x34'"); CheckParseEq("\\u003z", "'u003z'"); CheckParseEq("foo[z]*", "(: 'foo' (# 0 - g [z]))"); CheckParseEq("^^^$$$\\b\\b\\b\\b", "(: @^i @^i @^i @$i @$i @$i @b @b @b @b)"); CheckParseEq("\\b\\b\\b\\b\\B\\B\\B\\B\\b\\b\\b\\b", "(: @b @b @b @b @B @B @B @B @b @b @b @b)"); CheckParseEq("\\b\\B\\b", "(: @b @B @b)"); // Lookarounds are not simple Assertions and must not be merged. CheckParseEq("(?=a)(?=a)", "(: (-> + 'a') (-> + 'a'))"); // Unicode regexps CheckParseEq("\\u{12345}", "'\\ud808\\udf45'", true); CheckParseEq("\\u{12345}\\u{23456}", "(! '\\ud808\\udf45' '\\ud84d\\udc56')", true); CheckParseEq("\\u{12345}|\\u{23456}", "(| '\\ud808\\udf45' '\\ud84d\\udc56')", true); CheckParseEq("\\u{12345}{3}", "(# 3 3 g '\\ud808\\udf45')", true); CheckParseEq("\\u{12345}*", "(# 0 - g '\\ud808\\udf45')", true); CheckParseEq("\\ud808\\udf45*", "(# 0 - g '\\ud808\\udf45')", true); CheckParseEq("[\\ud808\\udf45-\\ud809\\udccc]", "[\\u{012345}-\\u{0124cc}]", true); CHECK_SIMPLE("", false); CHECK_SIMPLE("a", true); CHECK_SIMPLE("a|b", false); CHECK_SIMPLE("a\\n", false); CHECK_SIMPLE("^a", false); CHECK_SIMPLE("a$", false); CHECK_SIMPLE("a\\b!", false); CHECK_SIMPLE("a\\Bb", false); CHECK_SIMPLE("a*", false); CHECK_SIMPLE("a*?", false); CHECK_SIMPLE("a?", false); CHECK_SIMPLE("a??", false); CHECK_SIMPLE("a{0,1}?", false); CHECK_SIMPLE("a{1,1}?", false); CHECK_SIMPLE("a{1,2}?", false); CHECK_SIMPLE("a+?", false); CHECK_SIMPLE("(a)", false); CHECK_SIMPLE("(a)\\1", false); CHECK_SIMPLE("(\\1a)", false); CHECK_SIMPLE("\\1(a)", false); CHECK_SIMPLE("a\\s", false); CHECK_SIMPLE("a\\S", false); CHECK_SIMPLE("a\\d", false); CHECK_SIMPLE("a\\D", false); CHECK_SIMPLE("a\\w", false); CHECK_SIMPLE("a\\W", false); CHECK_SIMPLE("a.", false); CHECK_SIMPLE("a\\q", false); CHECK_SIMPLE("a[a]", false); CHECK_SIMPLE("a[^a]", false); CHECK_SIMPLE("a[a-z]", false); CHECK_SIMPLE("a[\\q]", false); CHECK_SIMPLE("a(?:b)", false); CHECK_SIMPLE("a(?=b)", false); CHECK_SIMPLE("a(?!b)", false); CHECK_SIMPLE("\\x60", false); CHECK_SIMPLE("\\u0060", false); CHECK_SIMPLE("\\cA", false); CHECK_SIMPLE("\\q", false); CHECK_SIMPLE("\\1112", false); CHECK_SIMPLE("\\0", false); CHECK_SIMPLE("(a)\\1", false); CHECK_SIMPLE("(?=a)?a", false); CHECK_SIMPLE("(?!a)?a\\1", false); CHECK_SIMPLE("(?:(?=a))a\\1", false); CheckParseEq("a{}", "'a{}'"); CheckParseEq("a{,}", "'a{,}'"); CheckParseEq("a{", "'a{'"); CheckParseEq("a{z}", "'a{z}'"); CheckParseEq("a{1z}", "'a{1z}'"); CheckParseEq("a{12z}", "'a{12z}'"); CheckParseEq("a{12,", "'a{12,'"); CheckParseEq("a{12,3b", "'a{12,3b'"); CheckParseEq("{}", "'{}'"); CheckParseEq("{,}", "'{,}'"); CheckParseEq("{", "'{'"); CheckParseEq("{z}", "'{z}'"); CheckParseEq("{1z}", "'{1z}'"); CheckParseEq("{12z}", "'{12z}'"); CheckParseEq("{12,", "'{12,'"); CheckParseEq("{12,3b", "'{12,3b'"); CHECK_MIN_MAX("a", 1, 1); CHECK_MIN_MAX("abc", 3, 3); CHECK_MIN_MAX("a[bc]d", 3, 3); CHECK_MIN_MAX("a|bc", 1, 2); CHECK_MIN_MAX("ab|c", 1, 2); CHECK_MIN_MAX("a||bc", 0, 2); CHECK_MIN_MAX("|", 0, 0); CHECK_MIN_MAX("(?:ab)", 2, 2); CHECK_MIN_MAX("(?:ab|cde)", 2, 3); CHECK_MIN_MAX("(?:ab)|cde", 2, 3); CHECK_MIN_MAX("(ab)", 2, 2); CHECK_MIN_MAX("(ab|cde)", 2, 3); CHECK_MIN_MAX("(ab)\\1", 2, regexp::Tree::kInfinity); CHECK_MIN_MAX("(ab|cde)\\1", 2, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:ab)?", 0, 2); CHECK_MIN_MAX("(?:ab)*", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:ab)+", 2, regexp::Tree::kInfinity); CHECK_MIN_MAX("a?", 0, 1); CHECK_MIN_MAX("a*", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("a+", 1, regexp::Tree::kInfinity); CHECK_MIN_MAX("a??", 0, 1); CHECK_MIN_MAX("a*?", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("a+?", 1, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:a?)?", 0, 1); CHECK_MIN_MAX("(?:a*)?", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:a+)?", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:a?)+", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:a*)+", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:a+)+", 1, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:a?)*", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:a*)*", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:a+)*", 0, regexp::Tree::kInfinity); CHECK_MIN_MAX("a{0}", 0, 0); CHECK_MIN_MAX("(?:a+){0}", 0, 0); CHECK_MIN_MAX("(?:a+){0,0}", 0, 0); CHECK_MIN_MAX("a*b", 1, regexp::Tree::kInfinity); CHECK_MIN_MAX("a+b", 2, regexp::Tree::kInfinity); CHECK_MIN_MAX("a*b|c", 1, regexp::Tree::kInfinity); CHECK_MIN_MAX("a+b|c", 1, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:a{5,1000000}){3,1000000}", 15, regexp::Tree::kInfinity); CHECK_MIN_MAX("(?:ab){4,7}", 8, 14); CHECK_MIN_MAX("a\\bc", 2, 2); CHECK_MIN_MAX("a\\Bc", 2, 2); CHECK_MIN_MAX("a\\sc", 3, 3); CHECK_MIN_MAX("a\\Sc", 3, 3); CHECK_MIN_MAX("a(?=b)c", 2, 2); CHECK_MIN_MAX("a(?=bbb|bb)c", 2, 2); CHECK_MIN_MAX("a(?!bbb|bb)c", 2, 2); CheckParseEq("(?x)(?x)(?x)\\k", "(: (^ 'x') (^ 'x') (^ 'x') (<- 1))", true); CheckParseEq("(?x)(?x)(?x)\\k", "(: (^ 'x') (^ 'x') (^ 'x') (<- 2))", true); CheckParseEq("(?x)(?x)(?x)\\k", "(: (^ 'x') (^ 'x') (^ 'x') (<- 3))", true); CheckParseEq("(?a)\\k ", "(: (^ 'a') (<- 1))", true); CheckParseEq("(?a\\k )", "(^ 'a')", true); CheckParseEq("(?\\k a)", "(^ 'a')", true); CheckParseEq("(?\\k)(?\\k)", "(: (^ (<- 2)) (^ (<- 1)))", true); CheckParseEq("\\k (?a)", "(: (<- 1) (^ 'a'))", true); CheckParseEq("(?<\\u{03C0}>a)", "(^ 'a')", true); CheckParseEq("(?<\\u03C0>a)", "(^ 'a')", true); } TEST_F(RegExpTest, ParserRegression) { CheckParseEq("[A-Z$-][x]", "(! [A-Z $ -] [x])"); CheckParseEq("a{3,4*}", "(: 'a{3,' (# 0 - g '4') '}')"); CheckParseEq("{", "'{'"); CheckParseEq("a|", "(| 'a' %)"); } static void ExpectError(const char* input, const char* expected, bool unicode = false) { Isolate* isolate = reinterpret_cast(v8::Isolate::GetCurrent()); v8::HandleScope scope(v8::Isolate::GetCurrent()); Zone zone(isolate->allocator(), ZONE_NAME); DirectHandle str = isolate->factory()->NewStringFromAsciiChecked(input); regexp::CompileData result; regexp::Flags flags; if (unicode) flags |= regexp::Flag::kUnicode; CHECK(!regexp::Parser::ParseRegExpFromHeapString(isolate, &zone, str, flags, &result)); CHECK_NULL(result.tree); CHECK_NE(regexp::Error::kNone, result.error); CHECK_EQ(0, strcmp(expected, ErrorString(result.error))); } TEST_F(RegExpTest, Errors) { const char* kEndBackslash = "\\ at end of pattern"; ExpectError("\\", kEndBackslash); const char* kUnterminatedGroup = "Unterminated group"; ExpectError("(foo", kUnterminatedGroup); const char* kInvalidGroup = "Invalid group"; ExpectError("(?", kInvalidGroup); const char* kUnterminatedCharacterClass = "Unterminated character class"; ExpectError("[", kUnterminatedCharacterClass); ExpectError("[a-", kUnterminatedCharacterClass); const char* kNothingToRepeat = "Nothing to repeat"; ExpectError("*", kNothingToRepeat); ExpectError("?", kNothingToRepeat); ExpectError("+", kNothingToRepeat); ExpectError("{1}", kNothingToRepeat); ExpectError("{1,2}", kNothingToRepeat); ExpectError("{1,}", kNothingToRepeat); // Check that we don't allow more than kMaxCapture captures const int kMaxCaptures = 1 << 16; // Must match regexp::Parser::kMaxCaptures. const char* kTooManyCaptures = "Too many captures"; std::ostringstream os; for (int i = 0; i <= kMaxCaptures; i++) { os << "()"; } ExpectError(os.str().c_str(), kTooManyCaptures); const char* kInvalidCaptureName = "Invalid capture group name"; ExpectError("(?<>.)", kInvalidCaptureName, true); ExpectError("(?<1>.)", kInvalidCaptureName, true); ExpectError("(?<_%>.)", kInvalidCaptureName, true); ExpectError("\\k.)(?.)", kDuplicateCaptureName, true); const char* kInvalidUnicodeEscape = "Invalid Unicode escape"; ExpectError("(?<\\u{FISK}", kInvalidUnicodeEscape, true); const char* kInvalidCaptureReferenced = "Invalid named capture referenced"; ExpectError("\\k ", kInvalidCaptureReferenced, true); ExpectError("(?)\\k", kInvalidCaptureReferenced, true); const char* kInvalidNamedReference = "Invalid named reference"; ExpectError("\\ka", kInvalidNamedReference, true); } static bool IsDigit(base::uc32 c) { return ('0' <= c && c <= '9'); } static bool NotDigit(base::uc32 c) { return !IsDigit(c); } static bool NotWhiteSpaceNorLineTermiantor(base::uc32 c) { return !IsWhiteSpaceOrLineTerminator(c); } static bool NotWord(base::uc32 c) { return !IsRegExpWord(c); } static bool NotLineTerminator(base::uc32 c) { return !unibrow::IsLineTerminator(c); } static void TestCharacterClassEscapes(regexp::StandardCharacterSet c, bool(pred)(base::uc32 c)) { Zone zone( reinterpret_cast(v8::Isolate::GetCurrent())->allocator(), ZONE_NAME); ZoneList* ranges = zone.New>(2, &zone); regexp::CharacterRange::AddClassEscape(c, ranges, false, &zone); for (base::uc32 i = 0; i < (1 << 16); i++) { bool in_class = false; for (int j = 0; !in_class && j < ranges->length(); j++) { regexp::CharacterRange& range = ranges->at(j); in_class = (range.from() <= i && i <= range.to()); } CHECK_EQ(pred(i), in_class); } } TEST_F(RegExpTest, CharacterClassEscapes) { TestCharacterClassEscapes(regexp::StandardCharacterSet::kNotLineTerminator, NotLineTerminator); TestCharacterClassEscapes(regexp::StandardCharacterSet::kDigit, IsDigit); TestCharacterClassEscapes(regexp::StandardCharacterSet::kNotDigit, NotDigit); TestCharacterClassEscapes(regexp::StandardCharacterSet::kWhitespace, IsWhiteSpaceOrLineTerminator); TestCharacterClassEscapes(regexp::StandardCharacterSet::kNotWhitespace, NotWhiteSpaceNorLineTermiantor); TestCharacterClassEscapes(regexp::StandardCharacterSet::kWord, IsRegExpWord); TestCharacterClassEscapes(regexp::StandardCharacterSet::kNotWord, NotWord); } static regexp::Node* Compile(const char* input, bool multiline, bool unicode, bool is_one_byte, Zone* zone) { Isolate* isolate = reinterpret_cast(v8::Isolate::GetCurrent()); DirectHandle str = isolate->factory()->NewStringFromAsciiChecked(input); regexp::CompileData compile_data; compile_data.compilation_target = regexp::CompilationTarget::kNative; regexp::Flags flags; if (multiline) flags |= regexp::Flag::kMultiline; if (unicode) flags |= regexp::Flag::kUnicode; if (!regexp::Parser::ParseRegExpFromHeapString(isolate, zone, str, flags, &compile_data)) { return nullptr; } DirectHandle pattern = isolate->factory() ->NewStringFromUtf8(base::CStrVector(input)) .ToHandleChecked(); DirectHandle sample_subject = isolate->factory() ->NewStringFromUtf8(base::CStrVector("")) .ToHandleChecked(); DirectHandle re_data = TrustedCast(isolate->factory()->NewIrRegExpData( pattern, pattern, JSRegExp::AsJSRegExpFlags(flags), compile_data.capture_count, JSRegExp::kNoBacktrackLimit, 0)); RegExp::CompileForTesting(isolate, zone, &compile_data, flags, pattern, sample_subject, re_data, is_one_byte); return compile_data.node; } static void Execute(const char* input, bool multiline, bool unicode, bool is_one_byte, bool dot_output = false) { v8::HandleScope scope(v8::Isolate::GetCurrent()); Zone zone( reinterpret_cast(v8::Isolate::GetCurrent())->allocator(), ZONE_NAME); regexp::Node* node = Compile(input, multiline, unicode, is_one_byte, &zone); USE(node); #ifdef DEBUG if (dot_output) RegExp::DotPrintForTesting(input, node); #endif // DEBUG } // Test of debug-only syntax. #ifdef DEBUG TEST_F(RegExpTest, ParsePossessiveRepetition) { bool old_flag_value = v8_flags.regexp_possessive_quantifier; // Enable possessive quantifier syntax. v8_flags.regexp_possessive_quantifier = true; CheckParseEq("a*+", "(# 0 - p 'a')"); CheckParseEq("a++", "(# 1 - p 'a')"); CheckParseEq("a?+", "(# 0 1 p 'a')"); CheckParseEq("a{10,20}+", "(# 10 20 p 'a')"); CheckParseEq("za{10,20}+b", "(: 'z' (# 10 20 p 'a') 'b')"); // Disable possessive quantifier syntax. v8_flags.regexp_possessive_quantifier = false; CHECK_PARSE_ERROR("a*+"); CHECK_PARSE_ERROR("a++"); CHECK_PARSE_ERROR("a?+"); CHECK_PARSE_ERROR("a{10,20}+"); CHECK_PARSE_ERROR("a{10,20}+b"); v8_flags.regexp_possessive_quantifier = old_flag_value; } #endif // Tests of interpreter. #if V8_TARGET_ARCH_IA32 using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerIA32; #elif V8_TARGET_ARCH_X64 using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerX64; #elif V8_TARGET_ARCH_ARM using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerARM; #elif V8_TARGET_ARCH_ARM64 using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerARM64; #elif V8_TARGET_ARCH_S390X using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerS390; #elif V8_TARGET_ARCH_PPC64 using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerPPC; #elif V8_TARGET_ARCH_MIPS64 using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerMIPS; #elif V8_TARGET_ARCH_LOONG64 using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerLOONG64; #elif V8_TARGET_ARCH_RISCV64 using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerRISCV; #elif V8_TARGET_ARCH_RISCV32 using ArchRegExpMacroAssembler = regexp::RegExpMacroAssemblerRISCV; #endif class ContextInitializer { public: ContextInitializer() : scope_(v8::Isolate::GetCurrent()), env_(v8::Context::New(v8::Isolate::GetCurrent())) { env_->Enter(); } ~ContextInitializer() { env_->Exit(); } private: v8::HandleScope scope_; v8::Local env_; }; // Create new JSRegExp object with only necessary fields (for this tests) // initialized. static DirectHandle CreateJSRegExp(DirectHandle source, DirectHandlecode, bool is_unicode = false) { Isolate* isolate = reinterpret_cast(v8::Isolate::GetCurrent()); Factory* factory = isolate->factory(); DirectHandle constructor = isolate->regexp_function(); DirectHandle regexp = Cast(factory->NewJSObject(constructor)); regexp->set_flags(Smi::FromInt(0)); factory->SetRegExpIrregexpData(regexp, source, source, {}, 0, JSRegExp::kNoBacktrackLimit, true); Tagged data = TrustedCast(regexp->data(isolate)); const bool is_latin1 = !is_unicode; data->set_code(is_latin1, *code); return regexp; } static ArchRegExpMacroAssembler::Result Execute( Tagged regexp, Tagged input, int start_offset, Address input_start, Address input_end, int* captures) { // For testing, we don't bother to pass in the `captures` size. This is okay // as long as the caller knows what they're doing, and avoids having the // engine write OOB or exiting a global execution loop early. static constexpr int kCapturesSize = 0; return static_cast( regexp::NativeRegExpMacroAssembler::ExecuteForTesting( input, start_offset, reinterpret_cast(input_start), reinterpret_cast(input_end), captures, kCapturesSize, reinterpret_cast(v8::Isolate::GetCurrent()), regexp)); } TEST_F(RegExpTest, MacroAssemblerNativeSuccess) { ContextInitializer initializer; Factory* factory = i_isolate()->factory(); Zone zone(i_isolate()->allocator(), ZONE_NAME); ArchRegExpMacroAssembler m(i_isolate(), &zone, regexp::NativeRegExpMacroAssembler::LATIN1, 4); m.Succeed(); DirectHandle source = factory->NewStringFromStaticChars(""); DirectHandle