pcrecpp_unittest.cc

php-4.4.7学习linux时下载的源代码

  // Get one string that passes (sometimes), one that never does.
  string text_good("abcdefghijk");
  string text_bad("acdefghijkl");

  // According to pcretest, matching text_good against (\w+)*b
  // requires match_limit of at least 8192, and match_recursion_limit
  // of at least 37.

  RE_Options options_ml;
  options_ml.set_match_limit(8192);
  RE re("(\\w+)*b", options_ml);
  CHECK(re.PartialMatch(text_good) == true);
  CHECK(re.PartialMatch(text_bad) == false);
  CHECK(re.FullMatch(text_good) == false);
  CHECK(re.FullMatch(text_bad) == false);

  options_ml.set_match_limit(1024);
  RE re2("(\\w+)*b", options_ml);
  CHECK(re2.PartialMatch(text_good) == false);   // because of match_limit
  CHECK(re2.PartialMatch(text_bad) == false);
  CHECK(re2.FullMatch(text_good) == false);
  CHECK(re2.FullMatch(text_bad) == false);

  RE_Options options_mlr;
  options_mlr.set_match_limit_recursion(50);
  RE re3("(\\w+)*b", options_mlr);
  CHECK(re3.PartialMatch(text_good) == true);
  CHECK(re3.PartialMatch(text_bad) == false);
  CHECK(re3.FullMatch(text_good) == false);
  CHECK(re3.FullMatch(text_bad) == false);

  options_mlr.set_match_limit_recursion(10);
  RE re4("(\\w+)*b", options_mlr);
  CHECK(re4.PartialMatch(text_good) == false);
  CHECK(re4.PartialMatch(text_bad) == false);
  CHECK(re4.FullMatch(text_good) == false);
  CHECK(re4.FullMatch(text_bad) == false);
}

// A meta-quoted string, interpreted as a pattern, should always match
// the original unquoted string.
static void TestQuoteMeta(string unquoted, RE_Options options = RE_Options()) {
  string quoted = RE::QuoteMeta(unquoted);
  RE re(quoted, options);
  CHECK(re.FullMatch(unquoted));
}

// A string containing meaningful regexp characters, which is then meta-
// quoted, should not generally match a string the unquoted string does.
static void NegativeTestQuoteMeta(string unquoted, string should_not_match,
                                  RE_Options options = RE_Options()) {
  string quoted = RE::QuoteMeta(unquoted);
  RE re(quoted, options);
  CHECK(!re.FullMatch(should_not_match));
}

// Tests that quoted meta characters match their original strings,
// and that a few things that shouldn't match indeed do not.
static void TestQuotaMetaSimple() {
  TestQuoteMeta("foo");
  TestQuoteMeta("foo.bar");
  TestQuoteMeta("foo\\.bar");
  TestQuoteMeta("[1-9]");
  TestQuoteMeta("1.5-2.0?");
  TestQuoteMeta("\\d");
  TestQuoteMeta("Who doesn't like ice cream?");
  TestQuoteMeta("((a|b)c?d*e+[f-h]i)");
  TestQuoteMeta("((?!)xxx).*yyy");
  TestQuoteMeta("([");
}

static void TestQuoteMetaSimpleNegative() {
  NegativeTestQuoteMeta("foo", "bar");
  NegativeTestQuoteMeta("...", "bar");
  NegativeTestQuoteMeta("\\.", ".");
  NegativeTestQuoteMeta("\\.", "..");
  NegativeTestQuoteMeta("(a)", "a");
  NegativeTestQuoteMeta("(a|b)", "a");
  NegativeTestQuoteMeta("(a|b)", "(a)");
  NegativeTestQuoteMeta("(a|b)", "a|b");
  NegativeTestQuoteMeta("[0-9]", "0");
  NegativeTestQuoteMeta("[0-9]", "0-9");
  NegativeTestQuoteMeta("[0-9]", "[9]");
  NegativeTestQuoteMeta("((?!)xxx)", "xxx");
}

static void TestQuoteMetaLatin1() {
  TestQuoteMeta("3\xb2 = 9");
}

static void TestQuoteMetaUtf8() {
#ifdef SUPPORT_UTF8
  TestQuoteMeta("Pl\xc3\xa1\x63ido Domingo", pcrecpp::UTF8());
  TestQuoteMeta("xyz", pcrecpp::UTF8());            // No fancy utf8
  TestQuoteMeta("\xc2\xb0", pcrecpp::UTF8());       // 2-byte utf8 (degree symbol)
  TestQuoteMeta("27\xc2\xb0 degrees", pcrecpp::UTF8());  // As a middle character
  TestQuoteMeta("\xe2\x80\xb3", pcrecpp::UTF8());   // 3-byte utf8 (double prime)
  TestQuoteMeta("\xf0\x9d\x85\x9f", pcrecpp::UTF8()); // 4-byte utf8 (music note)
  TestQuoteMeta("27\xc2\xb0"); // Interpreted as Latin-1, but should still work
  NegativeTestQuoteMeta("27\xc2\xb0",               // 2-byte utf (degree symbol)
                        "27\\\xc2\\\xb0",
                        pcrecpp::UTF8());
#endif
}

static void TestQuoteMetaAll() {
  printf("Testing QuoteMeta\n");
  TestQuotaMetaSimple();
  TestQuoteMetaSimpleNegative();
  TestQuoteMetaLatin1();
  TestQuoteMetaUtf8();
}

//
// Options tests contributed by
// Giuseppe Maxia, CTO, Stardata s.r.l.
// July 2005
//
static void GetOneOptionResult(
                const char *option_name,
                const char *regex,
                const char *str,
                RE_Options options,
                bool full,
                string expected) {

  printf("Testing Option <%s>\n", option_name);
  if(VERBOSE_TEST)
    printf("/%s/ finds \"%s\" within \"%s\" \n",
                    regex,
                    expected.c_str(),
                    str);
  string captured("");
  if (full)
    RE(regex,options).FullMatch(str, &captured);
  else
    RE(regex,options).PartialMatch(str, &captured);
  CHECK_EQ(captured, expected);
}

static void TestOneOption(
                const char *option_name,
                const char *regex,
                const char *str,
                RE_Options options,
                bool full,
                bool assertive = true) {

  printf("Testing Option <%s>\n", option_name);
  if (VERBOSE_TEST)
    printf("'%s' %s /%s/ \n",
                  str,
                  (assertive? "matches" : "doesn't match"),
                  regex);
  if (assertive) {
    if (full)
      CHECK(RE(regex,options).FullMatch(str));
    else
      CHECK(RE(regex,options).PartialMatch(str));
  } else {
    if (full)
      CHECK(!RE(regex,options).FullMatch(str));
    else
      CHECK(!RE(regex,options).PartialMatch(str));
  }
}

static void Test_CASELESS() {
  RE_Options options;
  RE_Options options2;

  options.set_caseless(true);
  TestOneOption("CASELESS (class)",  "HELLO",    "hello", options, false);
  TestOneOption("CASELESS (class2)", "HELLO",    "hello", options2.set_caseless(true), false);
  TestOneOption("CASELESS (class)",  "^[A-Z]+$", "Hello", options, false);

  TestOneOption("CASELESS (function)", "HELLO",    "hello", pcrecpp::CASELESS(), false);
  TestOneOption("CASELESS (function)", "^[A-Z]+$", "Hello", pcrecpp::CASELESS(), false);
  options.set_caseless(false);
  TestOneOption("no CASELESS", "HELLO",    "hello", options, false, false);
}

static void Test_MULTILINE() {
  RE_Options options;
  RE_Options options2;
  const char *str = "HELLO\n" "cruel\n" "world\n";

  options.set_multiline(true);
  TestOneOption("MULTILINE (class)",    "^cruel$", str, options, false);
  TestOneOption("MULTILINE (class2)",   "^cruel$", str, options2.set_multiline(true), false);
  TestOneOption("MULTILINE (function)", "^cruel$", str, pcrecpp::MULTILINE(), false);
  options.set_multiline(false);
  TestOneOption("no MULTILINE", "^cruel$", str, options, false, false);
}

static void Test_DOTALL() {
  RE_Options options;
  RE_Options options2;
  const char *str = "HELLO\n" "cruel\n" "world";

  options.set_dotall(true);
  TestOneOption("DOTALL (class)",    "HELLO.*world", str, options, true);
  TestOneOption("DOTALL (class2)",   "HELLO.*world", str, options2.set_dotall(true), true);
  TestOneOption("DOTALL (function)",    "HELLO.*world", str, pcrecpp::DOTALL(), true);
  options.set_dotall(false);
  TestOneOption("no DOTALL", "HELLO.*world", str, options, true, false);
}

static void Test_DOLLAR_ENDONLY() {
  RE_Options options;
  RE_Options options2;
  const char *str = "HELLO world\n";

  TestOneOption("no DOLLAR_ENDONLY", "world$", str, options, false);
  options.set_dollar_endonly(true);
  TestOneOption("DOLLAR_ENDONLY 1",    "world$", str, options, false, false);
  TestOneOption("DOLLAR_ENDONLY 2",    "world$", str, options2.set_dollar_endonly(true), false, false);
}

static void Test_EXTRA() {
  RE_Options options;
  const char *str = "HELLO";

  options.set_extra(true);
  TestOneOption("EXTRA 1", "\\HELL\\O", str, options, true, false );
  TestOneOption("EXTRA 2", "\\HELL\\O", str, RE_Options().set_extra(true), true, false );
  options.set_extra(false);
  TestOneOption("no EXTRA", "\\HELL\\O", str, options, true );
}

static void Test_EXTENDED() {
  RE_Options options;
  RE_Options options2;
  const char *str = "HELLO world";

  options.set_extended(true);
  TestOneOption("EXTENDED (class)",    "HELLO world", str, options, false, false);
  TestOneOption("EXTENDED (class2)",   "HELLO world", str, options2.set_extended(true), false, false);
  TestOneOption("EXTENDED (class)",
                    "^ HE L{2} O "
                    "\\s+        "
                    "\\w+ $      ",
                    str,
                    options,
                    false);

  TestOneOption("EXTENDED (function)",    "HELLO world", str, pcrecpp::EXTENDED(), false, false);
  TestOneOption("EXTENDED (function)",
                    "^ HE L{2} O "
                    "\\s+        "
                    "\\w+ $      ",
                    str,
                    pcrecpp::EXTENDED(),
                    false);

  options.set_extended(false);
  TestOneOption("no EXTENDED", "HELLO world", str, options, false);
}

static void Test_NO_AUTO_CAPTURE() {
  RE_Options options;
  const char *str = "HELLO world";
  string captured;

  printf("Testing Option <no NO_AUTO_CAPTURE>\n");
  if (VERBOSE_TEST)
    printf("parentheses capture text\n");
  RE re("(world|universe)$", options);
  CHECK(re.Extract("\\1", str , &captured));
  CHECK_EQ(captured, "world");
  options.set_no_auto_capture(true);
  printf("testing Option <NO_AUTO_CAPTURE>\n");
  if (VERBOSE_TEST)
    printf("parentheses do not capture text\n");
  re.Extract("\\1",str, &captured );
  CHECK_EQ(captured, "world");
}

static void Test_UNGREEDY() {
  RE_Options options;
  const char *str = "HELLO, 'this' is the 'world'";

  options.set_ungreedy(true);
  GetOneOptionResult("UNGREEDY 1", "('.*')", str, options, false, "'this'" );
  GetOneOptionResult("UNGREEDY 2", "('.*')", str, RE_Options().set_ungreedy(true), false, "'this'" );
  GetOneOptionResult("UNGREEDY", "('.*?')", str, options, false, "'this' is the 'world'" );

  options.set_ungreedy(false);
  GetOneOptionResult("no UNGREEDY", "('.*')", str, options, false, "'this' is the 'world'" );
  GetOneOptionResult("no UNGREEDY", "('.*?')", str, options, false, "'this'" );
}

static void Test_all_options() {
  const char *str = "HELLO\n" "cruel\n" "world";
  RE_Options options;
  options.set_all_options(PCRE_CASELESS | PCRE_DOTALL);

  TestOneOption("all_options (CASELESS|DOTALL)", "^hello.*WORLD", str , options, false);
  options.set_all_options(0);
  TestOneOption("all_options (0)", "^hello.*WORLD", str , options, false, false);
  options.set_all_options(PCRE_MULTILINE | PCRE_EXTENDED);

  TestOneOption("all_options (MULTILINE|EXTENDED)", " ^ c r u e l $ ", str, options, false);
  TestOneOption("all_options (MULTILINE|EXTENDED) with constructor",
                  " ^ c r u e l $ ",
                  str,
                  RE_Options(PCRE_MULTILINE | PCRE_EXTENDED),
                  false);

  TestOneOption("all_options (MULTILINE|EXTENDED) with concatenation",
                  " ^ c r u e l $ ",
                  str,
                  RE_Options()
                       .set_multiline(true)
                       .set_extended(true),
                  false);

  options.set_all_options(0);
  TestOneOption("all_options (0)", "^ c r u e l $", str, options, false, false);

}

static void TestOptions() {
  printf("Testing Options\n");
  Test_CASELESS();
  Test_MULTILINE();
  Test_DOTALL();
  Test_DOLLAR_ENDONLY();
  Test_EXTENDED();
  Test_NO_AUTO_CAPTURE();
  Test_UNGREEDY();
  Test_EXTRA();
  Test_all_options();
}

static void TestConstructors() {
  printf("Testing constructors\n");

  RE_Options options;
  options.set_dotall(true);
  const char *str = "HELLO\n" "cruel\n" "world";

  RE orig("HELLO.*world", options);
  CHECK(orig.FullMatch(str));

  RE copy1(orig);
  CHECK(copy1.FullMatch(str));

  RE copy2("not a match");
  CHECK(!copy2.FullMatch(str));
  copy2 = copy1;
  CHECK(copy2.FullMatch(str));
  copy2 = orig;
  CHECK(copy2.FullMatch(str));

  // Make sure when we assign to ourselves, nothing bad happens
  orig = orig;
  copy1 = copy1;
  copy2 = copy2;
  CHECK(orig.FullMatch(str));
  CHECK(copy1.FullMatch(str));
  CHECK(copy2.FullMatch(str));
}

int main(int argc, char** argv) {
  // Treat any flag as --help
  if (argc > 1 && argv[1][0] == '-') {
    printf("Usage: %s [timing1|timing2|timing3 num-iters]\n"
           "       If 'timingX ###' is specified, run the given timing test\n"
           "       with the given number of iterations, rather than running\n"
           "       the default corectness test.\n", argv[0]);
    return 0;
  }

  if (argc > 1) {
    if ( argc == 2 || atoi(argv[2]) == 0) {
      printf("timing mode needs a num-iters argument\n");
      return 1;
    }
    if (!strcmp(argv[1], "timing1"))
      Timing1(atoi(argv[2]));
    else if (!strcmp(argv[1], "timing2"))
      Timing2(atoi(argv[2]));
    else if (!strcmp(argv[1], "timing3"))
      Timing3(atoi(argv[2]));
    else
      printf("Unknown argument '%s'\n", argv[1]);
    return 0;
  }

  printf("Testing FullMatch\n");

  int i;
  string s;

  /***** FullMatch with no args *****/

  CHECK(RE("h.*o").FullMatch("hello"));
  CHECK(!RE("h.*o").FullMatch("othello"));
  CHECK(!RE("h.*o").FullMatch("hello!"));

  /***** FullMatch with args *****/

  // Zero-arg
  CHECK(RE("\\d+").FullMatch("1001"));

  // Single-arg
  CHECK(RE("(\\d+)").FullMatch("1001",   &i));
  CHECK_EQ(i, 1001);
  CHECK(RE("(-?\\d+)").FullMatch("-123", &i));