&effects) {
  effects.reserve(2 * getArgs().size() + 1);
  effects.emplace_back(MemoryEffects::Write::get(),
                       ResourceEffects::_XlaRun::get());

  // Conservatively mark resource handles as read and write, as without
  // analyzing _XlaCompile, there is not sufficient information to determine
  // effects on resources.
  for (Value value : getArgs()) {
    MarkResourceAsReadAndWrite(value, effects);

    loader = it->second.get();
  }

  // Cannot pass `op.getArgs().getTypes()` to `loader->RefineDynamicShapes`
  // because `op` and `loader` are using different MLIR contexts. See comments
  // on `xla_call_module_context_` for details.
  std::vector<xla::Shape> input_shapes;
  input_shapes.reserve(op.getArgs().size());
  for (mlir::Type type : op.getArgs().getTypes()) {
    input_shapes.push_back(xla::TypeToShape(type));

            }
        }

        /**
         * Generates the init method with deprecation logging disabled.
         *
         * This is necessary because the initialization work invokes property getters on the decorated instance.
         */
        private void generateInitMethod() {

            // private void $gradleInit() { DeprecationLogger.whileDisabled(this::$gradleInitWork) }

Method <org.gradle.process.ExecSpec.getArgs()> does not have raw return type assignable to org.gradle.api.provider.Property in (ExecSpec.java:0)
Method <org.gradle.process.ExecSpec.getArgumentProviders()> does not have raw return type assignable to org.gradle.api.provider.Provider in (ExecSpec.java:0)

org.codehaus.plexus.util.cli; public synchronized class Commandline$Argument { private String[] parts; public void Commandline$Argument(); public void setValue(String); public void setLine(String); public void setFile(java.io.File); public String[] getParts(); } org/codehaus/plexus/util/cli/Commandline$Marker.class package org.codehaus.plexus.util.cli; public synchronized class Commandline$Marker { private int position; private int realPos; void Commandline$Marker(Commandline, int); public int getPosition();...

           << func_name;
  }

  FunctionType func_ty = func.getFunctionType();
  int func_arg_count = func_ty.getNumInputs();
  int arg_count = getArgs().size();

  if (arg_count != func_arg_count) {
    return emitError() << "argument count mismatch: 'args' has " << arg_count
                       << " argument(s), but '" << func_name << "' expects "

      else if (auto stats_op =
                   llvm::dyn_cast_or_null<mlir::quantfork::StatisticsOp>(
                       operand.getDefiningOp()))
        operands.push_back(tensor_index_map.lookup(stats_op.getArg()));
      else
        operands.push_back(tensor_index_map.lookup(operand));
    }

    // CustomTfOp is just a wrapper around a TF op, we export the custom Op