All batches do their own pipeline compare

src/gpu/GrAALinearizingConvexPathRenderer.cpp

Index: src/gpu/GrAALinearizingConvexPathRenderer.cpp
diff --git a/src/gpu/GrAALinearizingConvexPathRenderer.cpp b/src/gpu/GrAALinearizingConvexPathRenderer.cpp
index 31d61a42574b5ba5c4c5003dbd50aa081a640e57..d899e7d0841425dd598ebe0f4c48b0aa47551bf7 100644
--- a/src/gpu/GrAALinearizingConvexPathRenderer.cpp
+++ b/src/gpu/GrAALinearizingConvexPathRenderer.cpp
@@ -261,6 +261,10 @@ private:
     }
 
     bool onCombineIfPossible(GrBatch* t) override {
+        if (!this->pipeline()->isEqual(*t->pipeline())) {
+            return false;
+        }
+
         AAFlatteningConvexPathBatch* that = t->cast<AAFlatteningConvexPathBatch>();
 
         SkASSERT(this->usesLocalCoords() == that->usesLocalCoords());