Alexander Gutkin | 0d4c523 | 2013-02-28 13:47:27 +0000 | [diff] [blame] | 1 | // Copyright 2007 The RE2 Authors. All Rights Reserved. |
| 2 | // Use of this source code is governed by a BSD-style |
| 3 | // license that can be found in the LICENSE file. |
| 4 | |
| 5 | // Test prog.cc, compile.cc |
| 6 | |
| 7 | #include <string> |
| 8 | #include <vector> |
| 9 | #include "util/test.h" |
| 10 | #include "re2/regexp.h" |
| 11 | #include "re2/prog.h" |
| 12 | |
| 13 | DEFINE_string(show, "", "regular expression to compile and dump"); |
| 14 | |
| 15 | namespace re2 { |
| 16 | |
| 17 | // Simple input/output tests checking that |
| 18 | // the regexp compiles to the expected code. |
| 19 | // These are just to sanity check the basic implementation. |
| 20 | // The real confidence tests happen by testing the NFA/DFA |
| 21 | // that run the compiled code. |
| 22 | |
| 23 | struct Test { |
| 24 | const char* regexp; |
| 25 | const char* code; |
| 26 | }; |
| 27 | |
| 28 | static Test tests[] = { |
| 29 | { "a", |
| 30 | "1. byte [61-61] -> 2\n" |
| 31 | "2. match! 0\n" }, |
| 32 | { "ab", |
| 33 | "1. byte [61-61] -> 2\n" |
| 34 | "2. byte [62-62] -> 3\n" |
| 35 | "3. match! 0\n" }, |
| 36 | { "a|c", |
| 37 | "3. alt -> 1 | 2\n" |
| 38 | "1. byte [61-61] -> 4\n" |
| 39 | "2. byte [63-63] -> 4\n" |
| 40 | "4. match! 0\n" }, |
| 41 | { "a|b", |
| 42 | "1. byte [61-62] -> 2\n" |
| 43 | "2. match! 0\n" }, |
| 44 | { "[ab]", |
| 45 | "1. byte [61-62] -> 2\n" |
| 46 | "2. match! 0\n" }, |
| 47 | { "a+", |
| 48 | "1. byte [61-61] -> 2\n" |
| 49 | "2. alt -> 1 | 3\n" |
| 50 | "3. match! 0\n" }, |
| 51 | { "a+?", |
| 52 | "1. byte [61-61] -> 2\n" |
| 53 | "2. alt -> 3 | 1\n" |
| 54 | "3. match! 0\n" }, |
| 55 | { "a*", |
| 56 | "2. alt -> 1 | 3\n" |
| 57 | "1. byte [61-61] -> 2\n" |
| 58 | "3. match! 0\n" }, |
| 59 | { "a*?", |
| 60 | "2. alt -> 3 | 1\n" |
| 61 | "3. match! 0\n" |
| 62 | "1. byte [61-61] -> 2\n" }, |
| 63 | { "a?", |
| 64 | "2. alt -> 1 | 3\n" |
| 65 | "1. byte [61-61] -> 3\n" |
| 66 | "3. match! 0\n" }, |
| 67 | { "a??", |
| 68 | "2. alt -> 3 | 1\n" |
| 69 | "3. match! 0\n" |
| 70 | "1. byte [61-61] -> 3\n" }, |
| 71 | { "a{4}", |
| 72 | "1. byte [61-61] -> 2\n" |
| 73 | "2. byte [61-61] -> 3\n" |
| 74 | "3. byte [61-61] -> 4\n" |
| 75 | "4. byte [61-61] -> 5\n" |
| 76 | "5. match! 0\n" }, |
| 77 | { "(a)", |
| 78 | "2. capture 2 -> 1\n" |
| 79 | "1. byte [61-61] -> 3\n" |
| 80 | "3. capture 3 -> 4\n" |
| 81 | "4. match! 0\n" }, |
| 82 | { "(?:a)", |
| 83 | "1. byte [61-61] -> 2\n" |
| 84 | "2. match! 0\n" }, |
| 85 | { "", |
| 86 | "2. match! 0\n" }, |
| 87 | { ".", |
| 88 | "3. alt -> 1 | 2\n" |
| 89 | "1. byte [00-09] -> 4\n" |
| 90 | "2. byte [0b-ff] -> 4\n" |
| 91 | "4. match! 0\n" }, |
| 92 | { "[^ab]", |
| 93 | "5. alt -> 3 | 4\n" |
| 94 | "3. alt -> 1 | 2\n" |
| 95 | "4. byte [63-ff] -> 6\n" |
| 96 | "1. byte [00-09] -> 6\n" |
| 97 | "2. byte [0b-60] -> 6\n" |
| 98 | "6. match! 0\n" }, |
| 99 | { "[Aa]", |
| 100 | "1. byte/i [61-61] -> 2\n" |
| 101 | "2. match! 0\n" }, |
| 102 | }; |
| 103 | |
| 104 | TEST(TestRegexpCompileToProg, Simple) { |
| 105 | int failed = 0; |
| 106 | for (int i = 0; i < arraysize(tests); i++) { |
| 107 | const re2::Test& t = tests[i]; |
| 108 | Regexp* re = Regexp::Parse(t.regexp, Regexp::PerlX|Regexp::Latin1, NULL); |
| 109 | if (re == NULL) { |
| 110 | LOG(ERROR) << "Cannot parse: " << t.regexp; |
| 111 | failed++; |
| 112 | continue; |
| 113 | } |
| 114 | Prog* prog = re->CompileToProg(0); |
| 115 | if (prog == NULL) { |
| 116 | LOG(ERROR) << "Cannot compile: " << t.regexp; |
| 117 | re->Decref(); |
| 118 | failed++; |
| 119 | continue; |
| 120 | } |
| 121 | CHECK(re->CompileToProg(1) == NULL); |
| 122 | string s = prog->Dump(); |
| 123 | if (s != t.code) { |
| 124 | LOG(ERROR) << "Incorrect compiled code for: " << t.regexp; |
| 125 | LOG(ERROR) << "Want:\n" << t.code; |
| 126 | LOG(ERROR) << "Got:\n" << s; |
| 127 | failed++; |
| 128 | } |
| 129 | delete prog; |
| 130 | re->Decref(); |
| 131 | } |
| 132 | EXPECT_EQ(failed, 0); |
| 133 | } |
| 134 | |
| 135 | // The distinct byte ranges involved in the UTF-8 dot ([^\n]). |
| 136 | // Once, erroneously split between 0x3f and 0x40 because it is |
| 137 | // a 6-bit boundary. |
| 138 | static struct UTF8ByteRange { |
| 139 | int lo; |
| 140 | int hi; |
| 141 | } utf8ranges[] = { |
| 142 | { 0x00, 0x09 }, |
| 143 | { 0x0A, 0x0A }, |
| 144 | { 0x10, 0x7F }, |
| 145 | { 0x80, 0x8F }, |
| 146 | { 0x90, 0x9F }, |
| 147 | { 0xA0, 0xBF }, |
| 148 | { 0xC0, 0xC1 }, |
| 149 | { 0xC2, 0xDF }, |
| 150 | { 0xE0, 0xE0 }, |
| 151 | { 0xE1, 0xEF }, |
| 152 | { 0xF0, 0xF0 }, |
| 153 | { 0xF1, 0xF3 }, |
| 154 | { 0xF4, 0xF4 }, |
| 155 | { 0xF5, 0xFF }, |
| 156 | }; |
| 157 | |
| 158 | TEST(TestCompile, ByteRanges) { |
| 159 | Regexp* re = Regexp::Parse(".", Regexp::PerlX, NULL); |
| 160 | EXPECT_TRUE(re != NULL); |
| 161 | Prog* prog = re->CompileToProg(0); |
| 162 | EXPECT_TRUE(prog != NULL); |
| 163 | EXPECT_EQ(prog->bytemap_range(), arraysize(utf8ranges)); |
| 164 | for (int i = 0; i < arraysize(utf8ranges); i++) |
| 165 | for (int j = utf8ranges[i].lo; j <= utf8ranges[i].hi; j++) |
| 166 | EXPECT_EQ(prog->bytemap()[j], i) << " byte " << j; |
| 167 | delete prog; |
| 168 | re->Decref(); |
| 169 | } |
| 170 | |
| 171 | } // namespace re2 |