Merge pull request #195 from paztis/master

Performances problems fixes on big files with a lot of references on same big object

Author: P0lip

lib/dereference.js

@@ -16,7 +16,7 @@ module.exports = dereference;
  */
 function dereference (parser, options) {
   // console.log('Dereferencing $ref pointers in %s', parser.$refs._root$Ref.path);
-  let dereferenced = crawl(parser.schema, parser.$refs._root$Ref.path, "#", [], parser.$refs, options);
+  let dereferenced = crawl(parser.schema, parser.$refs._root$Ref.path, "#", [], [], {}, parser.$refs, options);
   parser.$refs.circular = dereferenced.circular;
   parser.schema = dereferenced.value;
 }
@@ -28,60 +28,65 @@ function dereference (parser, options) {
  * @param {string} path - The full path of `obj`, possibly with a JSON Pointer in the hash
  * @param {string} pathFromRoot - The path of `obj` from the schema root
  * @param {object[]} parents - An array of the parent objects that have already been dereferenced
+ * @param {object[]} processedObjects - An array of all the objects that have already been processed
+ * @param {object} dereferencedCache - An map of all the dereferenced objects
  * @param {$Refs} $refs
  * @param {$RefParserOptions} options
  * @returns {{value: object, circular: boolean}}
  */
-function crawl (obj, path, pathFromRoot, parents, $refs, options) {
+function crawl (obj, path, pathFromRoot, parents, processedObjects, dereferencedCache, $refs, options) {
   let dereferenced;
   let result = {
     value: obj,
     circular: false
   };
 
-  if (obj && typeof obj === "object" && !ArrayBuffer.isView(obj)) {
-    parents.push(obj);
+  if (options.dereference.circular === "ignore" || processedObjects.indexOf(obj) === -1) {
+    if (obj && typeof obj === "object" && !ArrayBuffer.isView(obj)) {
+      parents.push(obj);
+      processedObjects.push(obj);
 
-    if ($Ref.isAllowed$Ref(obj, options)) {
-      dereferenced = dereference$Ref(obj, path, pathFromRoot, parents, $refs, options);
-      result.circular = dereferenced.circular;
-      result.value = dereferenced.value;
-    }
-    else {
-      for (let key of Object.keys(obj)) {
-        let keyPath = Pointer.join(path, key);
-        let keyPathFromRoot = Pointer.join(pathFromRoot, key);
-        let value = obj[key];
-        let circular = false;
-
-        if ($Ref.isAllowed$Ref(value, options)) {
-          dereferenced = dereference$Ref(value, keyPath, keyPathFromRoot, parents, $refs, options);
-          circular = dereferenced.circular;
-          // Avoid pointless mutations; breaks frozen objects to no profit
-          if (obj[key] !== dereferenced.value) {
-            obj[key] = dereferenced.value;
-          }
-        }
-        else {
-          if (parents.indexOf(value) === -1) {
-            dereferenced = crawl(value, keyPath, keyPathFromRoot, parents, $refs, options);
+      if ($Ref.isAllowed$Ref(obj, options)) {
+        dereferenced = dereference$Ref(obj, path, pathFromRoot, parents, processedObjects, dereferencedCache, $refs, options);
+        result.circular = dereferenced.circular;
+        result.value = dereferenced.value;
+      }
+      else {
+        for (let key of Object.keys(obj)) {
+          let keyPath = Pointer.join(path, key);
+          let keyPathFromRoot = Pointer.join(pathFromRoot, key);
+          let value = obj[key];
+          let circular = false;
+
+          if ($Ref.isAllowed$Ref(value, options)) {
+            dereferenced = dereference$Ref(value, keyPath, keyPathFromRoot, parents, processedObjects, dereferencedCache, $refs, options);
             circular = dereferenced.circular;
             // Avoid pointless mutations; breaks frozen objects to no profit
             if (obj[key] !== dereferenced.value) {
               obj[key] = dereferenced.value;
             }
           }
           else {
-            circular = foundCircularReference(keyPath, $refs, options);
+            if (parents.indexOf(value) === -1) {
+              dereferenced = crawl(value, keyPath, keyPathFromRoot, parents, processedObjects, dereferencedCache, $refs, options);
+              circular = dereferenced.circular;
+              // Avoid pointless mutations; breaks frozen objects to no profit
+              if (obj[key] !== dereferenced.value) {
+                obj[key] = dereferenced.value;
+              }
+            }
+            else {
+              circular = foundCircularReference(keyPath, $refs, options);
+            }
           }
-        }
 
-        // Set the "isCircular" flag if this or any other property is circular
-        result.circular = result.circular || circular;
+          // Set the "isCircular" flag if this or any other property is circular
+          result.circular = result.circular || circular;
+        }
       }
-    }
 
-    parents.pop();
+      parents.pop();
+    }
   }
 
   return result;
@@ -94,14 +99,38 @@ function crawl (obj, path, pathFromRoot, parents, $refs, options) {
  * @param {string} path - The full path of `$ref`, possibly with a JSON Pointer in the hash
  * @param {string} pathFromRoot - The path of `$ref` from the schema root
  * @param {object[]} parents - An array of the parent objects that have already been dereferenced
+ * @param {object[]} processedObjects - An array of all the objects that have already been dereferenced
+ * @param {object} dereferencedCache - An map of all the dereferenced objects 
  * @param {$Refs} $refs
  * @param {$RefParserOptions} options
  * @returns {{value: object, circular: boolean}}
  */
-function dereference$Ref ($ref, path, pathFromRoot, parents, $refs, options) {
+function dereference$Ref ($ref, path, pathFromRoot, parents, processedObjects, dereferencedCache, $refs, options) {
   // console.log('Dereferencing $ref pointer "%s" at %s', $ref.$ref, path);
 
   let $refPath = url.resolve(path, $ref.$ref);
+
+  if (dereferencedCache[$refPath]) {
+    const cache = dereferencedCache[$refPath];
+
+    const refKeys = Object.keys($ref);
+    if (refKeys.length > 1) {
+      const extraKeys = {};
+      for (let key of refKeys) {
+        if (key !== "$ref" && !(key in cache.value)) {
+          extraKeys[key] = $ref[key];
+        }
+      }
+      return {
+        circular: cache.circular,
+        value: Object.assign({}, cache.value, extraKeys),
+      };
+    }
+
+    return cache;
+  }
+
+
   let pointer = $refs._resolve($refPath, path, options);
 
   if (pointer === null) {
@@ -122,7 +151,7 @@ function dereference$Ref ($ref, path, pathFromRoot, parents, $refs, options) {
   // Crawl the dereferenced value (unless it's circular)
   if (!circular) {
     // Determine if the dereferenced value is circular
-    let dereferenced = crawl(dereferencedValue, pointer.path, pathFromRoot, parents, $refs, options);
+    let dereferenced = crawl(dereferencedValue, pointer.path, pathFromRoot, parents, processedObjects, dereferencedCache, $refs, options);
     circular = dereferenced.circular;
     dereferencedValue = dereferenced.value;
   }
@@ -138,10 +167,18 @@ function dereference$Ref ($ref, path, pathFromRoot, parents, $refs, options) {
     dereferencedValue.$ref = pathFromRoot;
   }
 
-  return {
+
+  const dereferencedObject = {
     circular,
     value: dereferencedValue
   };
+
+  // only cache if no extra properties than $ref
+  if (Object.keys($ref).length === 1) {
+    dereferencedCache[$refPath] = dereferencedObject;
+  }
+
+  return dereferencedObject;
 }
 
 /**