structs :shock:

[bugzpodder]

code:

struct BigInteger{                 int sz,start,maxsize, *arr;         BigInteger(){start=0;maxsize=1;sz=1; arr=new int[1];}              BigInteger(int val){BigInteger(); arr[0]=val;} }

I dont know what the heck the problem is with the code above, but whenever I call BigInteger using the second constructor, sz is always 0!!

this can be seen from:

code:

#include<iostream> using namespace std; struct BigInteger{                 int sz,start,maxsize, *arr;         BigInteger(){start=0;maxsize=1;sz=1; arr=new int[1]; cout<<sz<<endl;}              BigInteger(int val){BigInteger(); cout<<sz<<endl; arr[0]=val;} }; int main(){         BigInteger u(20);         return 0; }

in which you'll see a 1 and 0 outputted. why on earth is that happening?


And also, for something like:

code:

pair<BigInteger,int> div(BigInteger &v1, int v2){              .......         cout<<ret.first.arr[0]<<endl;         return ret; }             ....         p=div(v,1e9);  //although 1e9 is a double rather than an int, it is not causing the problmes         cout<<p.first.arr[0]<<endl;

and this produce totally different outputs!!!
I am completely lost as to whats going on...

[rizzix]

try:

c++:

class BigInteger {    protected:    int sz,start,maxsize, *arr;         public:     BigInteger(){          start=0;          maxsize=1;          sz=1;           arr=new int[1];      }         BigInteger(int val): BigInteger() {           arr[0]=val;      } };

[wtd]

First thing's first. You should be using proper constructor syntax, and since you're dynamically allocating memory, you should have a destructor.

c++:

struct BigInteger {    int sz,start,maxsize, *arr;    BigInteger()    : start(0)    , maxsize(1)    , sz(1)    , arr(new int[1])    { }    BigInteger(int val)    : start(0)    , maxsize(1)    , sz(1)    , arr(new int[1])    {       arr[0] = val;    }    ~BigInteger()    {       delete [] arr;    } };

[rizzix]

hmmm is that syntax C friendly?

[bugzpodder]

i do have a destructor, but didnt bother to post it. my full code is 200 lines long. Okay, so a class probably works, but what the heck is wrong with struct?

[wtd]

None of this is valid C. It is, however, perfectly valid C++ code.

[bugzpodder]

hmm interesting, never seen code like that before wtd.

and the second problem is... well here is the full code (first problem was fixed): funny things are happening in toString (outputs dont match), which calls div

no clue why it is happening

code:

#include<iostream> #include<fstream> #include<string> #include<algorithm> using namespace std; #define pb push_back            //pushes an element in the back #define pf push_front           //pushes an element in the front #define fe first                //first element of a pair container #define se second               //second element of a pair container #define len length()            //length of a string using namespace std; typedef unsigned int uint;      //32-bit unsigned int typedef long long ll;           //64-bit signed long long typedef unsigned long long ull; //64-bit unsigned long long struct BigInteger{                 int sz,start,maxsize;         uint *arr;         BigInteger(int val=0){start=0;maxsize=1;sz=1; arr=new uint[1];arr[0]=val;}         ~BigInteger(){delete [] arr;}                 int operator[](int i) {return arr[(start+i)%maxsize];}                 void resize(int size){                 if (sz>=size) return;                 uint *oldarr=arr;                 arr=new uint[size];                 for (int i=0;i<sz;i++) arr[i]=oldarr[(start+i)%maxsize];                 delete [] oldarr;                 start=0; maxsize=size;                         }         void pb(uint v){                 if (sz==maxsize) resize(maxsize*2);                 arr[(start+(sz++))%maxsize]=v;         }         void pop_back(){                 if (sz>0) sz--;         }         void pf(uint v){                 if (sz==maxsize) resize(maxsize*2);                 sz++; start=(start+maxsize-1)%maxsize;                 arr[start]=v;         } }; typedef pair<bool,BigInteger> bigint;   //signed BigInt typedef pair<BigInteger,uint> divtype;  //type returned by division BigInteger add(BigInteger v1, BigInteger v2){   bool carry=false;   int i; ull v;BigInteger ret;   for (i=0;i<v1.sz || i<v2.sz;i++){     v=(ull)(i<v1.sz?v1[i]:0)+(i<v2.sz?v2[i]:0)+carry;     carry=v>=(1ULL<<32);     ret.pb((uint)v);      }   if (carry) ret.pb(1);   return ret; } BigInteger sub(BigInteger v1, BigInteger v2){   bool carry=false;   int i; ull v;BigInteger ret;   for (i=0;i<v1.sz || i<v2.sz;i++){     v=(ull)(i<v1.sz?v1[i]:0)-(i<v2.sz?v2[i]:0)-carry;     carry=v>=(1ULL<<32);         ret.pb((uint)v);      }   if (carry) ret.pb(1);   return ret; } divtype div(BigInteger &v1, uint v2){         int i; ull v=0;         divtype ret; ret.fe.sz=0;         for (i=v1.sz-1;i>=0;i--){                 v=v*(1ULL<<32)+v1[i];                       ret.fe.pf(v/v2);                 v%=v2;         }         if (ret.fe.sz>1 && ret.fe[ret.fe.sz-1]==0) ret.fe.pop_back();         ret.se=v;         cout<<ret.fe.start<<' '<<ret.fe[0]<<' '<<ret.fe.sz<<' '<<ret.se<<endl;         return ret; } BigInteger convert(uint val){         BigInteger ret;ret.pb(val); return ret; } int compareTo(BigInteger &v1, BigInteger &v2){         if (v1.sz!=v2.sz) return v1.sz<v2.sz?1:-1;         for (int i=v1.sz-1;i>=0;i--)         if (v1[i]!=v2[i]) return v1[i]<v2[i]?1:-1;         return 0; } string toString(BigInteger v){   if (v.sz==1 && v[0]==0) return "0";   divtype p; string ret;   BigInteger zero(0);   int i;     //while(compareTo(v,zero)!=0){         p=div(v,1e9);         v=p.fe;         cout<<v.start<<' '<<v[0]<<' '<<v.sz<<' '<<p.se<<endl;         for (i=0;i<9;i++){ret+=p.se%10+'0';p.se/=10;}   //}   reverse(ret.begin(),ret.end());   for (i=0;i<ret.len-1 && ret[i]=='0';i++);   return ret.substr(i); } BigInteger num[5001]; int main(){         BigInteger u(50);         cout<<u.sz<<endl;         u.resize(20);         cout<<toString(u)<<endl;         cout<<u[0]<<endl; return 0; }

[bugzpodder]

and i still dont get what the heck is wrong with what I did (in my first post) the second constructor calls the first, and it IS getting called, so it seems perfectly fine

[wtd]

Do NOT do I/O in a function like that.

[bugzpodder]

??? err?? why not? its just debugging output? my code is getting SIGFPE (division/mod by 0) and SIGSEV's... these outputs pinpointed where the error is... I was wondering why the hell my code is getting SIGFPE until i found out that the variable sz mysteriously changed its value to 0 when the second constructor is called (so i made everything into one constructor... still not knowing why I cant keep two constructors like what I did before) and then I am getting an infinite loop in toString, which it turns out that the return value from div is different than the value I am getting in toString?!?! I am completely lost as to why this is happening...

anyhow i think the problem is with structs since the code has being tested if you remove the struct and use typedef deque<uint> BigInteger;

unfortunately deque is much slower... and I dont get why structs are giving me these resuts.

[wtd]

Because I/O is not meat to go into functions designed purely to output a new value. It's just bad. First make your changes, then check to see if things are as you expected. If you feel the need to do this inside a function, it's a good sign that function is doing too much.

Anyway, a sane basis for such work:

c++:

#include <vector> template <typename T> class LimitedVector : public std::vector<T> {    protected:       size_t max_size;    public:       LimitedVector() : std::vector<T>() { }       LimitedVector(size_t s) : std::vector<T>(s) { }          LimitedVector(size_t s, const T& v) : std::vector<T>(s, v) { }       ~LimitedVector() { delete this; }       void resize(size_t new_size)       {          if (new_size <= max_size)             std::vector<T>::resize(new_size);       } }; class BigInteger : public LimitedVector<unsigned int> {    public:       BigInteger() : LimitedVector<unsigned int>(1, *new unsigned int) { }       BigInteger(unsigned int init_val)       : LimitedVector<unsigned int>(1, *new unsigned int)       {          at(0) = init_val;       }       ~BigInteger() { delete this; } };

[rizzix]

bugzpodder wrote:

i do have a destructor, but didnt bother to post it. my full code is 200 lines long. Okay, so a class probably works, but what the heck is wrong with struct?

its because u have to pay close attention to the Consturctor() : initialization {} statements

although it appears to be two ways to intialize members of a class/struct in c++. there really is only one recognized official way. i suggest u create get/set accessor/mutators or simply define variables protected, for this to work out well.

also note Constructor(some_args) : Consturct() {} is the only working way to go about with calling previously defined constructs including the parent class' constuctor in c++. that is you cannot call the constuctor within the {} block of another constrcutor. although this should work, its does not always work well with all compilers.

although in D, Java, and a hell a lot of other languages, you dont have such annoying complications.

[bugzpodder]

wtd, One of my original plans is to push performance, since there is just too much overhead in vectors.

rizzix, ic, well it seemed to work :S heh

[rizzix]

the fist most common way in most languages:

code:

class BigInteger {    protected:    int sz,start,maxsize, *arr;        public:    BigInteger(){start=0;maxsize=1;sz=1; arr=new int[1];} };

the second the more c++ specific way"

code:

class BigInteger {    protected:    int sz,start,maxsize, *arr;        public:    BigInteger() : start(0), maxsize(1), sz(1), arr(new int[1]) {} };

Note that if you want to initialzise members of class using a previously defined consructor always use the second method:

code:

BigInteger(int val) : BigInteger() {arr[0]=val;}

you may if you wish, initialize member of the class directly using the second way, although the first way will work just as well.

[wtd]

bugzpodder wrote:

wtd, One of my original plans is to push performance, since there is just too much overhead in vectors.

I can guarantee std::vector will be faster than what you can cook up. That's not to insult your skills. but a lot of people have invested a lot of time in making the STL fast.

Also, your use of macros, and lack of spacing in your code will not make it any faster.