Consolidate handling of infinitely thin primitives and aa bloat handing WRT batch bounds

src/gpu/batches/GrBatch.cpp

Index: src/gpu/batches/GrBatch.cpp
diff --git a/src/gpu/batches/GrBatch.cpp b/src/gpu/batches/GrBatch.cpp
index 900f889ff870f759ba09ab7dbf1d7c26d527ef42..50f94b3902fbbdbc134a0bf672359608f6283d21 100644
--- a/src/gpu/batches/GrBatch.cpp
+++ b/src/gpu/batches/GrBatch.cpp
@@ -49,7 +49,9 @@ void GrBatch::operator delete(void* target) {
 GrBatch::GrBatch(uint32_t classID)
     : fClassID(classID)
     , fUniqueID(kIllegalBatchID) {
+    SkASSERT(classID == SkToU32(fClassID));
     SkDEBUGCODE(fUsed = false;)
+    SkDEBUGCODE(fBoundsFlags = kUninitialized_BoundsFlag);
 }
 
 GrBatch::~GrBatch() {}