Add flags option to Regex::exec (#12546)

Author: bryancall

include/tsutil/Regex.h

@@ -36,6 +36,7 @@ enum REFlags {
   RE_CASE_INSENSITIVE = 0x00000008u, ///< Ignore case (default: case sensitive).
   RE_UNANCHORED       = 0x00000400u, ///< Unanchored (DFA defaults to anchored).
   RE_ANCHORED         = 0x80000000u, ///< Anchored (Regex defaults to unanchored).
+  RE_NOTEMPTY         = 0x00000004u  ///< Not empty (default: may match empty string).
 };
 
 /// @brief Wrapper for PCRE2 match data.
@@ -124,7 +125,7 @@ class Regex
    *
    * It is safe to call this method concurrently on the same instance of @a this.
    */
-  bool exec(std::string_view subject) const;
+  bool exec(std::string_view subject, uint32_t flags = 0) const;
 
   /** Execute the regular expression.
    *
@@ -137,7 +138,7 @@ class Regex
    * Each capture group takes 3 elements of @a ovector, therefore @a ovecsize must
    * be a multiple of 3 and at least three times the number of desired capture groups.
    */
-  int exec(std::string_view subject, RegexMatches &matches) const;
+  int exec(std::string_view subject, RegexMatches &matches, uint32_t flags = 0) const;
 
   /// @return The number of capture groups in the compiled pattern.
   int get_capture_count();

src/tsutil/Regex.cc

@@ -31,9 +31,10 @@
 #include <vector>
 #include <mutex>
 
-static_assert(RE_CASE_INSENSITIVE == PCRE2_CASELESS, "Update RE_CASE_INSERSITIVE for current PCRE2 version.");
-static_assert(RE_UNANCHORED == PCRE2_MULTILINE, "Update RE_MULTILINE for current PCRE2 version.");
+static_assert(RE_CASE_INSENSITIVE == PCRE2_CASELESS, "Update RE_CASE_INSENSITIVE for current PCRE2 version.");
+static_assert(RE_UNANCHORED == PCRE2_MULTILINE, "Update RE_UNANCHORED for current PCRE2 version.");
 static_assert(RE_ANCHORED == PCRE2_ANCHORED, "Update RE_ANCHORED for current PCRE2 version.");
+static_assert(RE_NOTEMPTY == PCRE2_NOTEMPTY, "Update RE_NOTEMPTY for current PCRE2 version.");
 
 //----------------------------------------------------------------------------
 namespace
@@ -296,28 +297,28 @@ Regex::compile(std::string_view pattern, std::string &error, int &erroroffset, u
 
 //----------------------------------------------------------------------------
 bool
-Regex::exec(std::string_view subject) const
+Regex::exec(std::string_view subject, uint32_t flags) const
 {
   if (_Code::get(_code) == nullptr) {
     return false;
   }
   RegexMatches matches;
 
-  int count = this->exec(subject, matches);
+  int count = this->exec(subject, matches, flags);
   return count > 0;
 }
 
 //----------------------------------------------------------------------------
 int32_t
-Regex::exec(std::string_view subject, RegexMatches &matches) const
+Regex::exec(std::string_view subject, RegexMatches &matches, uint32_t flags) const
 {
   auto code = _Code::get(_code);
 
   // check if there is a compiled regex
   if (code == nullptr) {
     return 0;
   }
-  int count = pcre2_match(code, reinterpret_cast<PCRE2_SPTR>(subject.data()), subject.size(), 0, 0,
+  int count = pcre2_match(code, reinterpret_cast<PCRE2_SPTR>(subject.data()), subject.size(), 0, flags,
                           RegexMatches::_MatchData::get(matches._match_data), RegexContext::get_instance()->get_match_context());
 
   matches._size = count;

src/tsutil/unit_tests/test_Regex.cc

@@ -190,3 +190,40 @@ TEST_CASE("Regex", "[libts][Regex]")
   }
 #endif
 }
+
+TEST_CASE("Regex RE_NOTEMPTY flag behavior", "[libts][Regex][flags][RE_NOTEMPTY]")
+{
+  // Pattern that only matches empty string
+  Regex r;
+  REQUIRE(r.compile("^$") == true);
+
+  SECTION("default exec matches empty subject")
+  {
+    // boolean overload
+    CHECK(r.exec(std::string_view("")) == true);
+
+    // matches overload should return 1 (one match - the whole subject)
+    RegexMatches matches;
+    CHECK(r.exec(std::string_view(""), matches) == 1);
+    CHECK(matches.size() == 1);
+    CHECK(matches[0] == std::string_view(""));
+  }
+
+  SECTION("RE_NOTEMPTY prevents empty matches")
+  {
+    // boolean overload with RE_NOTEMPTY should not match
+    CHECK(r.exec(std::string_view(""), RE_NOTEMPTY) == false);
+
+    // matches overload should return a negative value (PCRE2_ERROR_NOMATCH)
+    RegexMatches matches;
+    int          rc = r.exec(std::string_view(""), matches, RE_NOTEMPTY);
+    CHECK(rc < 0);
+  }
+
+  SECTION("non-empty subject unaffected by RE_NOTEMPTY for this pattern")
+  {
+    // '^$' should not match 'a' in any case
+    CHECK(r.exec(std::string_view("a")) == false);
+    CHECK(r.exec(std::string_view("a"), RE_NOTEMPTY) == false);
+  }
+}