Probleme mit int_klasse

[Zarathustra23]

Hallo!
ich bin dabei eine Klasse zu entwerfen, welche es ermoeglicht, positive Zahlen zu speichern, die groesser (2^32)-1 sind.
Das Problem ist jedoch, dass mein Programm Fehler erzeugt sobald die Zahlen groesser (2^96)-1 werden.
Ich scheine leider unfaehig zu sein den Fehler zu finden, ich hoffe hier kann mir geholfen werden..

#include<iostream>
#include<iomanip>
#include<cmath>
using namespace std;
class long_unsigned
	{
	public:
	long_unsigned();
	long_unsigned(unsigned);
	inline unsigned giveGR(){return GR;};
	inline unsigned give(unsigned op){if(op<GR)return Z[op];else {add(1);return Z[GR-1];}};
	void add(unsigned);
	void set(int,unsigned);
	void operator+=(long_unsigned);
	void operator*=(int);
	void kuerze();
	void print();
	private:
	unsigned *Z;
	unsigned GR;	
	};
long_unsigned::long_unsigned()
	{
	Z=new unsigned ;
	GR=1;
	Z[0]=0;	
	}
long_unsigned::long_unsigned(unsigned op)
	{
	Z=new unsigned;
	GR=1;
	Z[0]=op;	
	}
void long_unsigned::kuerze()
	{
	if(GR>1)
		{
		if(Z[GR-1]==0)
			{
			unsigned *new_Z=new unsigned [GR-1];
			for(int i=0;i<GR-2;i++)
				{
				new_Z[i]=Z[i];
				}
			delete [] Z;
			Z=new_Z;
			GR--;
			}
		}
	}

void long_unsigned::print()
	{
	for(int i=0;i<GR;i++)
	for(int j=0;j<32;j++)
	{
	if(j==0)cout<<"|";
	if(Z[i]&(unsigned)pow(2.,j))cout<<"1";
	else cout<<"0";
	}
	cout<<"|"<<endl;

	}
void long_unsigned::set(int nr, unsigned wert)
	{
	if(nr>=GR)add(nr-GR+1);
	Z[nr]=wert;
	}
void long_unsigned::add(unsigned anz)
	{
	unsigned *new_Z=new unsigned [GR+anz];
	for(int i=0;i<GR;i++)
	new_Z[i]=Z[i];
	delete [] Z;
	Z=new_Z;
	for(int i=GR;i<GR+anz;i++)
	Z[i]=0;
	GR=GR+anz;
	}		
void long_unsigned::operator+=(long_unsigned op)
	{
	bool merker=false;
	int nr=0;
	     do
		{

		unsigned act=0;
		for(int i=0;i<32;i++)
			{
			if(merker)
				{
				if(Z[nr]&(unsigned)pow(2.,i)&&op.give(nr)&(unsigned)pow(2.,i))
					{
					act=act|(unsigned)pow(2.,i);
					}
				else
					{
					if(Z[nr]&(unsigned)pow(2.,i)||op.give(nr)&(unsigned)pow(2.,i))
						{
						merker=true;
						}
					else
						{
						act=act|(unsigned)pow(2.,i);
						merker=false;
						}
					}
				}
			else
				{
				if(Z[nr]&(unsigned)pow(2.,i)&&op.give(nr)&(unsigned)pow(2.,i))
					{
					merker=true;
					}
				else
					{
					if(Z[nr]&(unsigned)pow(2.,i)||op.give(nr)&(unsigned)pow(2.,i))
						{
						act=act|(unsigned)pow(2.,i);
						}
					}
				}
			}
		Z[nr]=act;

		if(merker&&nr==GR-1){add(1);}
		nr++;
		}while(nr<GR);

	}	
void long_unsigned::operator*=(int op)
	{
	for(int i=1;i<op;i++)*this+=*this;

	}		
int main()
{
long_unsigned a(pow(2.,2));
for(int i=0;i<150;i++)
{
a+=a;

a.print();
}
return 0;
}

mfg Zarathustra23

[Simon2]

Ist ja interessant.

Zarathustra23 schrieb:

...mein Programm Fehler erzeugt sobald die Zahlen groesser (2^96)-1 werden....

Welche denn ?

Sorry, ein wenig mehr Information wäre hilfreich. Ansonsten kann ich nur zu einem Debugger raten (ist sowieso eine gute Idee).

Gruß,

Simon2.

[Zarathustra23]

|00010000000000000000000000000000|
|00001000000000000000000000000000|
|00000100000000000000000000000000|
|00000010000000000000000000000000|
|00000001000000000000000000000000|
|00000000100000000000000000000000|
|00000000010000000000000000000000|
|00000000001000000000000000000000|
|00000000000100000000000000000000|
|00000000000010000000000000000000|
|00000000000001000000000000000000|
|00000000000000100000000000000000|
|00000000000000010000000000000000|
|00000000000000001000000000000000|
|00000000000000000100000000000000|
|00000000000000000010000000000000|
|00000000000000000001000000000000|
|00000000000000000000100000000000|
|00000000000000000000010000000000|
|00000000000000000000001000000000|
|00000000000000000000000100000000|
|00000000000000000000000010000000|
|00000000000000000000000001000000|
|00000000000000000000000000100000|
|00000000000000000000000000010000|
|00000000000000000000000000001000|
|00000000000000000000000000000100|
|00000000000000000000000000000010|
|00000000000000000000000000000001|
|00000000000000000000000000000000|10000000000000000000000000000000|
|00000000000000000000000000000000|01000000000000000000000000000000|
|00000000000000000000000000000000|00100000000000000000000000000000|
|00000000000000000000000000000000|00010000000000000000000000000000|
|00000000000000000000000000000000|00001000000000000000000000000000|
|00000000000000000000000000000000|00000100000000000000000000000000|
|00000000000000000000000000000000|00000010000000000000000000000000|
|00000000000000000000000000000000|00000001000000000000000000000000|
|00000000000000000000000000000000|00000000100000000000000000000000|
|00000000000000000000000000000000|00000000010000000000000000000000|
|00000000000000000000000000000000|00000000001000000000000000000000|
|00000000000000000000000000000000|00000000000100000000000000000000|
|00000000000000000000000000000000|00000000000010000000000000000000|
|00000000000000000000000000000000|00000000000001000000000000000000|
|00000000000000000000000000000000|00000000000000100000000000000000|
|00000000000000000000000000000000|00000000000000010000000000000000|
|00000000000000000000000000000000|00000000000000001000000000000000|
|00000000000000000000000000000000|00000000000000000100000000000000|
|00000000000000000000000000000000|00000000000000000010000000000000|
|00000000000000000000000000000000|00000000000000000001000000000000|
|00000000000000000000000000000000|00000000000000000000100000000000|
|00000000000000000000000000000000|00000000000000000000010000000000|
|00000000000000000000000000000000|00000000000000000000001000000000|
|00000000000000000000000000000000|00000000000000000000000100000000|
|00000000000000000000000000000000|00000000000000000000000010000000|
|00000000000000000000000000000000|00000000000000000000000001000000|
|00000000000000000000000000000000|00000000000000000000000000100000|
|00000000000000000000000000000000|00000000000000000000000000010000|
|00000000000000000000000000000000|00000000000000000000000000001000|
|00000000000000000000000000000000|00000000000000000000000000000100|
|00000000000000000000000000000000|00000000000000000000000000000010|
|00000000000000000000000000000000|00000000000000000000000000000001|
|00000000000000000000000000000000|00000000000000000000000000000000|10000000000000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|01000000000000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00100000000000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00010000000000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00001000000000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000100000000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000010000000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000001000000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000100000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000010000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000001000000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000100000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000010000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000001000000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000100000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000010000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000001000000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000100000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000010000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000001000000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000100000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000010000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000001000000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000000100000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000000010000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000000001000000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000000000100000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000000000010000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000000000001000|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000000000000100|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000000000000010|
|00000000000000000000000000000000|00000000000000000000000000000000|00000000000000000000000000000001|
*** glibc detected *** ./a.out: double free or corruption (fasttop): 0x0804b008 ***
======= Backtrace: =========
/lib/tls/i686/cmov/libc.so.6[0xb7ce6a85]
/lib/tls/i686/cmov/libc.so.6(cfree+0x90)[0xb7cea4f0]
/usr/lib/libstdc++.so.6(_ZdlPv+0x21)[0xb7eb1b11]
/usr/lib/libstdc++.so.6(_ZdaPv+0x1d)[0xb7eb1b6d]
./a.out(__gxx_personality_v0+0x382)[0x8048af6]
./a.out[0x8049235]
./a.out[0x8048ca9]
./a.out[0x80490f8]
/lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe0)[0xb7c91450]
./a.out(__gxx_personality_v0+0x3d)[0x80487b1]
======= Memory map: ========
08048000-0804a000 r-xp 00000000 08:12 8831496    /mnt/no2/tino/QUELLTEXTE/AAA_neues/a.out
0804a000-0804b000 rw-p 00001000 08:12 8831496    /mnt/no2/tino/QUELLTEXTE/AAA_neues/a.out
0804b000-0806c000 rw-p 0804b000 00:00 0          [heap]
b7b00000-b7b21000 rw-p b7b00000 00:00 0 
b7b21000-b7c00000 ---p b7b21000 00:00 0 
b7c7a000-b7c7b000 rw-p b7c7a000 00:00 0 
b7c7b000-b7dc4000 r-xp 00000000 08:01 689466     /lib/tls/i686/cmov/libc-2.7.so
b7dc4000-b7dc5000 r--p 00149000 08:01 689466     /lib/tls/i686/cmov/libc-2.7.so
b7dc5000-b7dc7000 rw-p 0014a000 08:01 689466     /lib/tls/i686/cmov/libc-2.7.so
b7dc7000-b7dca000 rw-p b7dc7000 00:00 0 
b7dca000-b7dd4000 r-xp 00000000 08:01 671809     /lib/libgcc_s.so.1
b7dd4000-b7dd5000 rw-p 0000a000 08:01 671809     /lib/libgcc_s.so.1
b7dd5000-b7df8000 r-xp 00000000 08:01 689474     /lib/tls/i686/cmov/libm-2.7.so
b7df8000-b7dfa000 rw-p 00023000 08:01 689474     /lib/tls/i686/cmov/libm-2.7.so
b7dfa000-b7ee2000 r-xp 00000000 08:01 18925      /usr/lib/libstdc++.so.6.0.9
b7ee2000-b7ee5000 r--p 000e8000 08:01 18925      /usr/lib/libstdc++.so.6.0.9
b7ee5000-b7ee7000 rw-p 000eb000 08:01 18925      /usr/lib/libstdc++.so.6.0.9
b7ee7000-b7eee000 rw-p b7ee7000 00:00 0 
b7efd000-b7eff000 rw-p b7efd000 00:00 0 
b7eff000-b7f00000 r-xp b7eff000 00:00 0          [vdso]
b7f00000-b7f1a000 r-xp 00000000 08:01 671763     /lib/ld-2.7.so
b7f1a000-b7f1c000 rw-p 00019000 08:01 671763     /lib/ld-2.7.so
bfdd7000-bfdec000 rw-p bffeb000 00:00 0          [stack]
Aborted

Das Programm faengt bei 2^4 oder so an, und verdoppelt dann immer... wenn 32bit voll sind und bei der Addition noch ein uebertrag bleibt, werden einfach neue 4 byte angefordert.
Ab 2^96 kommt er dann nich mehr auf seinen Arbeitsspeicher fit, wie die ausgabe erahnen laesst...

mfg

[Badestrand]

Es liegt bestimmt nicht daran, aber eine Kleinigkeit: Du forderst im Konstruktor mit "new unsigned" Speicher an, gibst ihn aber mit "delete**[]**" frei, also am Besten überall mit "new[]" anfordern.

edit: Kann's sein, dass du in der Methode "operator+=" gar nicht den Speicher vergrößerst, wenn's erforderlich ist?

[Zarathustra23]

Erster Punkt bachte mich nicht weiter, haette mich auch gewundert wenn der g++ daran gescheitert waere^^

zweitne Punkt kann ich leider nur dementieren

Zeile 127:

if(merker&&nr==GR-1){add(1);}

(Wenn Uebertrag und kein Platz mehr, neuen Speicher anfordern)

mfg

[Badestrand]

Zarathustra23 schrieb:

Erster Punkt bachte mich nicht weiter, haette mich auch gewundert wenn der g++ daran gescheitert waere^^

Wenn die Klasse portabel sein soll (also auch auf anderen Plattformen/Compilern laufen soll), würde ich schon sauber programmieren.

Zarathustra23 schrieb:

zweitne Punkt kann ich leider nur dementieren

Oh stimmt, das "add" haste aber auch gut versteckt

[camper]

Deine Klasse besitzt einen Zeiger auf allokierten Speicher, aber keinen selbstgeschriebenen Copy-ctor oder Zuweisungsoperator. Der Copyctor wird zumindest für das Funktionsargument von += benötigt. Zwar sollte dies wohl ohnehin besser eine Referenz auf const sein - trotzdem ist die normale Kopierinfrastruktur wichtig. Generell würde ich empfehlen, diese bei jeder neuen Klasse als erstes zu implementieren - dann kann man das nicht vergessen. Es gibt noch eine ganze Reihe anderer Fehler, aber die sind nicht ursächlich für den Absturz.
Nebenbei: das bitweise addieren ist reichlich umständlich. Du solltest dir zunutze machen, dass vorzeichenlose Arithmetik nicht überlaufen kann.

c = a + b + carry;
carry = carry ? c <= a : c < a;

[Erhard Henkes]

Vielleicht kannst Du Dir hier etwas abschauen:
http://mattmccutchen.net/bigint/