add taint step through array-union, array-uniq, and uniq

javascript/change-notes/2021-07-15-array-libs.md

@@ -5,4 +5,7 @@ lgtm,codescanning
     [array.prototype.find](https://npmjs.com/package/array.prototype.find),
     [array-find](https://npmjs.com/package/array-find),
     [arrify](https://npmjs.com/package/arrify),
-    [array-ify](https://npmjs.com/package/array-ify)
+    [array-ify](https://npmjs.com/package/array-ify),
+    [array-union](https://npmjs.com/package/array-union),
+    [array-uniq](https://npmjs.com/package/array-uniq),
+    [uniq](https://npmjs.com/package/uniq)

javascript/ql/src/semmle/javascript/Arrays.qll

@@ -323,6 +323,8 @@ private import ArrayLibraries
  * Classes and predicates modelling various libraries that work on arrays or array-like structures.
  */
 private module ArrayLibraries {
+  private import DataFlow::PseudoProperties
+
   /**
    * Gets a call to `Array.from` or a polyfill implementing the same functionality.
    */
@@ -353,4 +355,38 @@ private module ArrayLibraries {
       )
     }
   }
+
+  /**
+   * A call to a library that copies the elements of an array into another array.
+   * E.g. `array-union` that creates a union of multiple arrays, or `array-uniq` that creates an array with unique elements.
+   */
+  DataFlow::CallNode arrayCopyCall(DataFlow::Node array) {
+    result = API::moduleImport(["array-union", "array-uniq", "uniq"]).getACall() and
+    array = result.getAnArgument()
+  }
+
+  /**
+   * A taint step for a library that copies the elements of an array into another array.
+   */
+  private class ArrayCopyTaint extends TaintTracking::SharedTaintStep {
+    override predicate step(DataFlow::Node pred, DataFlow::Node succ) {
+      exists(DataFlow::CallNode call |
+        call = arrayCopyCall(pred) and
+        succ = call
+      )
+    }
+  }
+
+  /**
+   * A loadStoreStep for a library that copies the elements of an array into another array.
+   */
+  private class ArrayCopyLoadStore extends DataFlow::SharedFlowStep {
+    override predicate loadStoreStep(DataFlow::Node pred, DataFlow::Node succ, string prop) {
+      exists(DataFlow::CallNode call |
+        call = arrayCopyCall(pred) and
+        succ = call and
+        prop = arrayElement()
+      )
+    }
+  }
 }

javascript/ql/test/library-tests/Arrays/DataFlow.expected

@@ -10,6 +10,7 @@
 | arrays.js:2:16:2:23 | "source" | arrays.js:71:10:71:10 | x |
 | arrays.js:2:16:2:23 | "source" | arrays.js:74:8:74:29 | arr.fin ... llback) |
 | arrays.js:2:16:2:23 | "source" | arrays.js:77:8:77:35 | arrayFi ... llback) |
+| arrays.js:2:16:2:23 | "source" | arrays.js:81:10:81:10 | x |
 | arrays.js:18:22:18:29 | "source" | arrays.js:18:50:18:50 | e |
 | arrays.js:22:15:22:22 | "source" | arrays.js:23:8:23:17 | arr2.pop() |
 | arrays.js:25:15:25:22 | "source" | arrays.js:26:8:26:17 | arr3.pop() |

javascript/ql/test/library-tests/Arrays/arrays.js

@@ -75,4 +75,9 @@
 
   const arrayFind = require("array-find");
   sink(arrayFind(arr, someCallback)); // NOT OK
+
+  const uniq = require("uniq");
+  for (const x of uniq(arr)) {
+    sink(x); // NOT OK
+  }
 });

javascript/ql/test/library-tests/TaintTracking/BasicTaintTracking.expected

@@ -14,6 +14,7 @@ typeInferenceMismatch
 | array-mutation.js:39:17:39:24 | source() | array-mutation.js:40:8:40:8 | j |
 | arrays.js:2:15:2:22 | source() | arrays.js:5:10:5:20 | arrify(foo) |
 | arrays.js:2:15:2:22 | source() | arrays.js:8:10:8:22 | arrayIfy(foo) |
+| arrays.js:2:15:2:22 | source() | arrays.js:11:10:11:28 | union(["bla"], foo) |
 | booleanOps.js:2:11:2:18 | source() | booleanOps.js:4:8:4:8 | x |
 | booleanOps.js:2:11:2:18 | source() | booleanOps.js:13:10:13:10 | x |
 | booleanOps.js:2:11:2:18 | source() | booleanOps.js:19:10:19:10 | x |

javascript/ql/test/library-tests/TaintTracking/arrays.js

@@ -6,4 +6,7 @@ function test() {
 
     const arrayIfy = require("array-ify");
     sink(arrayIfy(foo)); // NOT OK
+
+    const union = require("array-union");
+    sink(union(["bla"], foo)); // NOT OK
 }