Add regular expression support in json_schema component

components/json_schema/json_schema_validator.cc

Index: components/json_schema/json_schema_validator.cc
diff --git a/components/json_schema/json_schema_validator.cc b/components/json_schema/json_schema_validator.cc
index 18558d0e111213bd789ac96633b8bc43afaaa53f..2a3e60d93100ee6e7ca62b931076255f4a906b39 100644
--- a/components/json_schema/json_schema_validator.cc
+++ b/components/json_schema/json_schema_validator.cc
@@ -7,13 +7,17 @@
 #include <algorithm>
 #include <cfloat>
 #include <cmath>
+#include <vector>
 
 #include "base/json/json_reader.h"
+#include "base/logging.h"
+#include "base/memory/scoped_vector.h"
 #include "base/strings/string_number_conversions.h"
 #include "base/strings/string_util.h"
 #include "base/strings/stringprintf.h"
 #include "base/values.h"
 #include "components/json_schema/json_schema_constants.h"
+#include "third_party/re2/re2/re2.h"
 
 namespace schema = json_schema_constants;
 
@@ -86,6 +90,8 @@ bool IsValidSchema(const base::DictionaryValue* dict,
     { schema::kMinLength,               base::Value::TYPE_INTEGER     },
     { schema::kMinimum,                 base::Value::TYPE_DOUBLE      },
     { schema::kOptional,                base::Value::TYPE_BOOLEAN     },
+    { schema::kPattern,                 base::Value::TYPE_STRING      },
+    { schema::kPatternProperties,       base::Value::TYPE_DICTIONARY  },
     { schema::kProperties,              base::Value::TYPE_DICTIONARY  },
     { schema::kTitle,                   base::Value::TYPE_STRING      },
   };
@@ -186,10 +192,27 @@ bool IsValidSchema(const base::DictionaryValue* dict,
     // Validate the "properties" attribute. Each entry maps a key to a schema.
     if (it.key() == schema::kProperties) {
       it.value().GetAsDictionary(&dictionary_value);
-      for (base::DictionaryValue::Iterator it(*dictionary_value);
-           !it.IsAtEnd(); it.Advance()) {
-        if (!it.value().GetAsDictionary(&dictionary_value)) {
-          *error = "Invalid value for properties attribute";
+      for (base::DictionaryValue::Iterator iter(*dictionary_value);
+           !iter.IsAtEnd(); iter.Advance()) {
+        if (!iter.value().GetAsDictionary(&dictionary_value)) {
+          *error = "property must be a dictionary";

[Joao da Silva, 2014/03/13 15:10:52]

properties

[binjin, 2014/03/20 15:14:30]

Done.

+          return false;
+        }
+        if (!IsValidSchema(dictionary_value, options, error)) {
+          DCHECK(!error->empty());
+          return false;
+        }
+      }
+    }
+
+    // Validate the "patternProperties" attribute. Each entry maps a valid
+    // regular expression to a schema.
+    if (it.key() == schema::kPatternProperties) {
+      it.value().GetAsDictionary(&dictionary_value);
+      for (base::DictionaryValue::Iterator iter(*dictionary_value);
+           !iter.IsAtEnd(); iter.Advance()) {
+        if (!iter.value().GetAsDictionary(&dictionary_value)) {
+          *error = "patternProperty must be a dictionary";
           return false;
         }
         if (!IsValidSchema(dictionary_value, options, error)) {
@@ -551,8 +574,7 @@ void JSONSchemaValidator::ValidateObject(const base::DictionaryValue* instance,
                                          const base::DictionaryValue* schema,
                                          const std::string& path) {
   const base::DictionaryValue* properties = NULL;
-  schema->GetDictionary(schema::kProperties, &properties);
-  if (properties) {
+  if (schema->GetDictionary(schema::kProperties, &properties)) {
     for (base::DictionaryValue::Iterator it(*properties); !it.IsAtEnd();
          it.Advance()) {
       std::string prop_path = path.empty() ? it.key() : (path + "." + it.key());
@@ -574,17 +596,51 @@ void JSONSchemaValidator::ValidateObject(const base::DictionaryValue* instance,
     }
   }
 
+  // Allowing any additional items will ignore pattern properties as well.
   const base::DictionaryValue* additional_properties_schema = NULL;
-  if (SchemaAllowsAnyAdditionalItems(schema, &additional_properties_schema))
-    return;
+  bool allow_any_additional_properties =
+      SchemaAllowsAnyAdditionalItems(schema, &additional_properties_schema);
 
-  // Validate additional properties.
+  const base::DictionaryValue* pattern_properties = NULL;
+  ScopedVector<re2::RE2> pattern_properties_pattern;
+  std::vector<const base::DictionaryValue*> pattern_properties_schema;
+
+  if (schema->GetDictionary(schema::kPatternProperties, &pattern_properties)) {
+    for (base::DictionaryValue::Iterator it(*pattern_properties); !it.IsAtEnd();
+         it.Advance()) {
+      re2::RE2* prop_pattern = new re2::RE2(it.key());
+      if (!prop_pattern->ok()) {
+        LOG(WARNING) << "Regular expression in schema is not valid: /"
+                     << it.key() << "/.";

[Joao da Silva, 2014/03/13 15:10:52]

LOG prop_pattern.error() and push it as an error too

[binjin, 2014/03/20 15:14:30]

Done.

+      }
+      const base::DictionaryValue* prop_schema = NULL;
+      CHECK(it.value().GetAsDictionary(&prop_schema));
+      pattern_properties_pattern.push_back(prop_pattern);
+      pattern_properties_schema.push_back(prop_schema);
+    }
+  }
+
+  // Validate pattern properties and additional properties.
   for (base::DictionaryValue::Iterator it(*instance); !it.IsAtEnd();
        it.Advance()) {
-    if (properties && properties->HasKey(it.key()))
-      continue;
 

[Joao da Silva, 2014/03/13 15:10:52]

nit: remove newline

[binjin, 2014/03/20 15:14:30]

Done.

     std::string prop_path = path.empty() ? it.key() : path + "." + it.key();
+
+    bool found_matching_pattern = false;
+    for (size_t index = 0; index < pattern_properties_pattern.size(); ++index) {
+      if (pattern_properties_pattern[index]->ok() &&
+          re2::RE2::PartialMatch(it.key(),
+                                 *pattern_properties_pattern[index])) {
+        found_matching_pattern = true;
+        Validate(&it.value(), pattern_properties_schema[index], prop_path);
+        break;
+      }
+    }
+
+    if (found_matching_pattern || allow_any_additional_properties ||
+        (properties && properties->HasKey(it.key())))
+      continue;
+
     if (!additional_properties_schema) {
       errors_.push_back(Error(prop_path, kUnexpectedProperty));
     } else {
@@ -707,7 +763,17 @@ void JSONSchemaValidator::ValidateString(const base::Value* instance,
     }
   }
 
-  CHECK(!schema->HasKey(schema::kPattern)) << "Pattern is not supported.";
+  std::string pattern;
+  if (schema->GetString(schema::kPattern, &pattern)) {
+    re2::RE2 compiled_regex(pattern);
+    if (!compiled_regex.ok()) {
+      LOG(WARNING) << "Regular expression in schema is not valid: /" << pattern
+                   << "/.";

[Joao da Silva, 2014/03/13 15:10:52]

This should push an error too. Otherwise a typo will just silently accept any
string, which is worse than rejecting all strings.

[Joao da Silva, 2014/03/13 15:10:52]

Include compiled_regex.error() in the LOG and the error pushed too

[binjin, 2014/03/20 15:14:30]

Done.

+    } else if (!re2::RE2::PartialMatch(value, compiled_regex)) {
+      errors_.push_back(
+          Error(path, FormatErrorMessage(kStringPattern, pattern)));
+    }
+  }
 }
 
 void JSONSchemaValidator::ValidateNumber(const base::Value* instance,