Delta calculation for matching of delta in CTPX should be done with context operators

[Leo-Nogismus]

In CTPX::reduce(), we calculate the delta between the consequent and target value at the beginning of the function. This lets us create a binding_map of this delta value and match patterns against this. For this, we are currently still relying on simple minus calculation of Atom::Float() values. This should be updated to the same operator usage as in the find_guard_builder() function.

See

AERA/r_exec/pattern_extractor.cpp

Lines 975 to 987 in f774c52

	bool have_delta = false;
	BindingMap bm_with_delta;
	if (target_->get_reference(0)->code(0).asOpcode() == Opcodes::MkVal) {
	// This is the same as the searched_for delta in find_guard_builder.
	float32 delta = consequent->get_reference(0)->code(MK_VAL_VALUE).asFloat() -
	target_->get_reference(0)->code(MK_VAL_VALUE).asFloat();
	if (delta != 0) {
	have_delta = true;
	P<Code> delta_code(new LObject());
	delta_code->code(0) = Atom::Float(delta);
	bm_with_delta.init(delta_code, 0);
	}
	}