On ARM, many jump table reads are one-byte memory reads. Assuming one…

…-byte read yields a value in [0,255] would cause too many bogus edges. Disable this heuristics on ARM

parseAPI/src/IndirectAnalyzer.C

@@ -48,7 +48,7 @@ bool IndirectControlFlowAnalyzer::NewJumpTableAnalysis(std::vector<std::pair< Ad
     // After the slicing is done, we do one last check to 
     // see if we can resolve the indirect jump by assuming 
     // one byte read is in bound [0,255]
-    if (jumpTableOutEdges.empty() && jtp.jumpTableFormat) {
+    if (jumpTableOutEdges.empty() && jtp.jumpTableFormat && block->obj()->cs()->getArch() != Arch_aarch64) {
         GraphPtr g = jtp.BuildAnalysisGraph(s.visitedEdges);
 
 	BoundFactsCalculator bfc(func, g, func->entry() == block, rf, thunks, block->last(), true, jtp.expandCache);