1 files changed, 40 insertions, 0 deletions
diff --git a/js/src/jit-test/tests/saved-stacks/caching-and-frame-count.js b/js/src/jit-test/tests/saved-stacks/caching-and-frame-count.js
new file mode 100644
index 0000000000..de40e564ae
--- /dev/null
+++ b/js/src/jit-test/tests/saved-stacks/caching-and-frame-count.js
@@ -0,0 +1,40 @@
+// |jit-test| skip-if: getBuildConfiguration()['wasi']
+//
+// Test that the SavedFrame caching doesn't mess up counts. Specifically, that
+// if we capture only the first n frames of a stack, we don't cache that stack
+// and return it for when someone else captures another stack and asks for more
+// frames than that last time.
+
+function stackLength(stack) {
+  return stack === null
+    ? 0
+    : 1 + stackLength(stack.parent);
+}
+
+(function f0() {
+  (function f1() {
+    (function f2() {
+      (function f3() {
+        (function f4() {
+          (function f5() {
+            (function f6() {
+              (function f7() {
+                (function f8() {
+                  (function f9() {
+                    const s1 = saveStack(3);
+                    const s2 = saveStack(5);
+                    const s3 = saveStack(0 /* no limit */);
+
+                    assertEq(stackLength(s1), 3);
+                    assertEq(stackLength(s2), 5);
+                    assertEq(stackLength(s3), 11);
+                  }());
+                }());
+              }());
+            }());
+          }());
+        }());
+      }());
+    }());
+  }());
+}());