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..0aefd5c29c2b0eb960416c1a107f63a1b66b7be1 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);

[not at google - send to devlin, 2014/03/21 15:27:20]

if it.value() isn't a DictionaryValue then |dictionary_value| will be untouched
and just be whatever the last value is (which may or may not be NULL), right?

i.e. you need to

if (!it.value().GetAsDictionary(&dictinoary_value))
  // fail

... and I guess, assuming I'm not misreading this, it was already a bug.

[binjin, 2014/03/24 14:35:09]

I think the type of it.value() have already been checked against corresponding
entry in kExpectedTypes. 

[not at google - send to devlin, 2014/03/25 14:38:05]

ah, right.

-      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 = "properties must be a dictionary";
+          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);

[not at google - send to devlin, 2014/03/21 15:27:20]

... though this would have the same bug.

[binjin, 2014/03/24 14:35:09]

as noted above

+      for (base::DictionaryValue::Iterator iter(*dictionary_value);
+           !iter.IsAtEnd(); iter.Advance()) {
+        if (!iter.value().GetAsDictionary(&dictionary_value)) {

[not at google - send to devlin, 2014/03/21 15:27:20]

add a comment why you're not validating iter.key() as a regex

[binjin, 2014/03/24 14:35:09]

Done.

+          *error = "patternProperties must be a dictionary";
           return false;
         }
         if (!IsValidSchema(dictionary_value, options, error)) {
@@ -308,6 +331,8 @@ const char JSONSchemaValidator::kInvalidType[] =
     "Expected '*' but got '*'.";
 const char JSONSchemaValidator::kInvalidTypeIntegerNumber[] =
     "Expected 'integer' but got 'number', consider using Math.round().";
+const char JSONSchemaValidator::kInvalidRegex[] =
+    "Regular expression /*/ is invalid: *";
 
 
 // static
@@ -551,8 +576,7 @@ 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 +598,54 @@ void JSONSchemaValidator::ValidateObject(const base::DictionaryValue* instance,
     }
   }
 
+  // Allowing any additional items will ignore pattern properties as well.

[not at google - send to devlin, 2014/03/21 15:27:20]

I don't understand this comment, you're not ignoring pattern properties with
|allow_any_additional_properties|?

[binjin, 2014/03/24 14:35:09]

Removed. In previous changeset there is a "if(allow_any_additional_properties)"
statement to stop processing any further properties, which includes pattern
properties. I forget to remove this comment after removing the if statement.

   const base::DictionaryValue* additional_properties_schema = NULL;
-  if (SchemaAllowsAnyAdditionalItems(schema, &additional_properties_schema))
-    return;
+  bool allow_any_additional_properties =
+      SchemaAllowsAnyAdditionalItems(schema, &additional_properties_schema);
+
+  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 /" << it.key()
+                     << "/ is invalid: " << prop_pattern->error() << ".";
+        errors_.push_back(
+            Error(path,
+                  FormatErrorMessage(
+                      kInvalidRegex, it.key(), prop_pattern->error())));

[not at google - send to devlin, 2014/03/21 15:27:20]

hm, why not continue after this? what is the point in adding an invalid pattern
to this list?

[binjin, 2014/03/24 14:35:09]

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 additional properties.
+  // Validate pattern properties and additional properties.
   for (base::DictionaryValue::Iterator it(*instance); !it.IsAtEnd();
        it.Advance()) {
-    if (properties && properties->HasKey(it.key()))
+    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() &&

[not at google - send to devlin, 2014/03/21 15:27:20]

(in which case you wouldn't need this check)

[binjin, 2014/03/24 14:35:09]

Done.

+          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;
 
-    std::string prop_path = path.empty() ? it.key() : path + "." + it.key();
     if (!additional_properties_schema) {
       errors_.push_back(Error(prop_path, kUnexpectedProperty));
     } else {
@@ -707,7 +768,20 @@ 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 /" << pattern
+                   << "/ is invalid: " << compiled_regex.error() << ".";
+      errors_.push_back(Error(
+          path,
+          FormatErrorMessage(kInvalidRegex, pattern, compiled_regex.error())));
+    } else if (!re2::RE2::PartialMatch(value, compiled_regex)) {
+      errors_.push_back(
+          Error(path, FormatErrorMessage(kStringPattern, pattern)));
+    }
+  }
 }
 
 void JSONSchemaValidator::ValidateNumber(const base::Value* instance,