No need for subtree recalc for :-internal-spatial-navigation-focus.

LayoutTests/fast/css/invalidation/spatial-navigation-focus-pseudo-expected.txt

Index: LayoutTests/fast/css/invalidation/spatial-navigation-focus-pseudo-expected.txt
diff --git a/LayoutTests/fast/css/invalidation/spatial-navigation-focus-pseudo-expected.txt b/LayoutTests/fast/css/invalidation/spatial-navigation-focus-pseudo-expected.txt
new file mode 100644
index 0000000000000000000000000000000000000000..08f1a377410af414599895e3c7eb11a778c262bb
--- /dev/null
+++ b/LayoutTests/fast/css/invalidation/spatial-navigation-focus-pseudo-expected.txt
@@ -0,0 +1,19 @@
+Use descendant invalidation set in the presence of an :-internal-spatial-navigation-focus pseudo class.
+
+On success, you will see a series of "PASS" messages, followed by "TEST COMPLETE".
+
+
+This test relies on the bug (crbug.com/432436) that internal pseudo classes are being
+exposed to the extent that they are not dropped, but in the case of :-internal-spatial-navigation-focus,
+never matched. If the mentioned bug is fixed this test will start failing. When that happens,
+we need to either add the possibility to use internal css features in author style for tests,
+support injecting UA style for tests, or drop this test.
+
+PASS getComputedStyle(b).backgroundColor is transparent
+PASS internals.updateStyleAndReturnAffectedElementCount() is 2
+PASS getComputedStyle(b).backgroundColor is green
+PASS successfullyParsed is true
+
+TEST COMPLETE
+
+