Clarify the use of DOTALL in JSON Schema regular expressions (#1586)

Signed-off-by: Juan Cruz Viotti <[email protected]>

Author: jviotti

src/core/regex/include/sourcemeta/core/regex.h

@@ -27,7 +27,7 @@
 #include <variant>  // std::variant
 
 /// @defgroup regex Regex
-/// @brief A regex ECMA implementation for JSON Schema
+/// @brief An opinionated regex ECMA 262 implementation for JSON Schema
 ///
 /// This functionality is included as follows:
 ///
@@ -95,6 +95,10 @@ auto to_regex(const T &pattern) -> std::optional<Regex<T>> {
   if (pattern == ".*" || pattern == "^.*$" || pattern == "^(.*)$" ||
       pattern == "(.*)" || pattern == "[\\s\\S]*" || pattern == "^[\\s\\S]*$") {
     return RegexTypeNoop{};
+
+    // Note that the JSON Schema specification does not impose the use of any
+    // regular expression flag. Given popular adoption, we assume `.` matches
+    // new line characters (as in the `DOTALL`) option
   } else if (pattern == ".+" || pattern == "^.+$" || pattern == "^(.+)$" ||
              pattern == ".") {
     return RegexTypeNonEmpty{};
@@ -125,6 +129,9 @@ auto to_regex(const T &pattern) -> std::optional<Regex<T>> {
           // treated as non-marking sub-expressions (?:expr)
           boost::regex::nosubs |
 
+          // Make the `.` character class match new lines
+          boost::regex::mod_s |
+
           // Instructs the regular expression engine to make matching faster,
           // with the potential cost of making construction slower
           boost::regex::optimize};

test/regex/regex_matches_test.cc

@@ -93,6 +93,25 @@ TEST(Regex_matches, match_true_14) {
                                                      "@namespace/mypackage"));
 }
 
+TEST(Regex_matches, match_true_15) {
+  const auto regex{sourcemeta::core::to_regex<std::string>(".")};
+  EXPECT_TRUE(regex.has_value());
+  EXPECT_TRUE(sourcemeta::core::matches<std::string>(regex.value(), "\n"));
+}
+
+TEST(Regex_matches, match_true_16) {
+  const auto regex{sourcemeta::core::to_regex<std::string>(".")};
+  EXPECT_TRUE(regex.has_value());
+  EXPECT_TRUE(sourcemeta::core::matches<std::string>(regex.value(), "\r"));
+}
+
+TEST(Regex_matches, match_true_17) {
+  const auto regex{sourcemeta::core::to_regex<std::string>("^.+$")};
+  EXPECT_TRUE(regex.has_value());
+  EXPECT_TRUE(
+      sourcemeta::core::matches<std::string>(regex.value(), "foo\nbar\r"));
+}
+
 TEST(Regex_matches, match_false_1) {
   const auto regex{sourcemeta::core::to_regex<std::string>("^foo")};
   EXPECT_TRUE(regex.has_value());