add impl limit for number of leaf-nodes in composecolorfilter

tests/ColorFilterTest.cpp

Index: tests/ColorFilterTest.cpp
diff --git a/tests/ColorFilterTest.cpp b/tests/ColorFilterTest.cpp
index f3f6a0a8e890f73d2752b6ef4b869276bf746e64..b2e37183b5e7b223cf1d545ee8138a5311e68b64 100644
--- a/tests/ColorFilterTest.cpp
+++ b/tests/ColorFilterTest.cpp
@@ -30,6 +30,26 @@ static SkColorFilter* reincarnate_colorfilter(SkFlattenable* obj) {
 
 ///////////////////////////////////////////////////////////////////////////////
 
+static SkColorFilter* make_filter() {
+    // pick a filter that cannot compose with itself via newComposed()
+    return SkColorFilter::CreateModeFilter(SK_ColorRED, SkXfermode::kColorBurn_Mode);
+}
+
+static void test_composecolorfilter_limit(skiatest::Reporter* reporter) {
+    // Test that CreateComposeFilter() has some finite limit (i.e. that the factory can return null)

[Stephen White, 2015/03/05 15:19:11]

Maybe we should add some more tests that reasonable numbers (less than the
limit) correctly return the number of leaf nodes?

[robertphillips, 2015/03/05 15:20:43]

kWayTooMany ?

+    const int way_too_many = 100;
+    SkAutoTUnref<SkColorFilter> parent(make_filter());
+    for (int i = 2; i < way_too_many; ++i) {
+        SkAutoTUnref<SkColorFilter> filter(make_filter());
+        parent.reset(SkColorFilter::CreateComposeFilter(parent, filter));
+        if (NULL == parent) {
+            REPORTER_ASSERT(reporter, i > 2); // we need to have succeeded at least once!
+            return;
+        }
+    }
+    REPORTER_ASSERT(reporter, false); // we never saw a NULL :(
+}
+
 #define ILLEGAL_MODE    ((SkXfermode::Mode)-1)
 
 DEF_TEST(ColorFilter, reporter) {
@@ -89,6 +109,8 @@ DEF_TEST(ColorFilter, reporter) {
             REPORTER_ASSERT(reporter, m2 == expectedMode);
         }
     }
+
+    test_composecolorfilter_limit(reporter);
 }
 
 ///////////////////////////////////////////////////////////////////////////////