1. If we define "ModNum::_num" as unsigned, there may be problems in constructing ModNum from negative integer. For example, you will see --- mcalc> mset 7 mcalc> madd x 0 -2 x(2) = 0 + 2 which is NOT correct!! we should see --- mcalc> madd x 0 -2 x(5) = 0 + 5 The reference program "modCalc" in ccws* had this problem too (it's now fixed). An easy way to solve this is to change "ModNum::_num" to "int" type. Please change the "unsigned" in the following code to "int", or download hw3.tgz again... calc/calcModNum.h:19: ModNum(unsigned i = 0); calc/calcModNum.h:27: static void setModulus(unsigned m) { _modulus = m; } calc/calcModNum.h:28: static unsigned getModulus() { return _modulus; } calc/calcModNum.h:56: unsigned _num; calc/calcModNum.h:58: static unsigned _modulus; calc/calcCmd.cpp:27: if (ModNum::getModulus() != unsigned(m)) { Of course, if you have other way to fix this, that's fine too. Just make sure when the arguments are negative integers, you can handle the modular number operations right! 2. In "myStrNCmp(const string& s1, const string& s2, unsigned n)", the comparison will be wrong if the string s2 is a superset of s1. For example, s1 = "HELp", s2 = "HELpkk", n = 3 --- we will see the comparison result = 0 (i.e. equivalent), which is NOT correct. We should return a negative integer (i.e. s1 < s2). Please update the "myStrNCmp" as follows, or download the new hw3.tgz --- ============= int myStrNCmp(const string& s1, const string& s2, unsigned n) { assert(n > 0); if (s2.size() == 0) return -1; unsigned l = s1.size(); assert(l >= n); for (unsigned i = 0; i < l; ++i) { if (!s2[i]) return (i < n)? 1 : 0; char ch1 = (isupper(s1[i]))? tolower(s1[i]) : s1[i]; char ch2 = (isupper(s2[i]))? tolower(s2[i]) : s2[i]; if (ch1 != ch2) return (ch1 - ch2); } return (l - s2.size()); } ====== The red lines are modified. --

※ 发信站: 批踢踢实业坊(ptt.cc) ◆ From: 140.112.4.242