C++ debugging problems - GameCreators Forum

Author

Message

Neuro Fuzzy

17

Years of Service

User Offline

Joined: 11th Jun 2007

Location:

Posted: 25th May 2011 22:24

Link

I hope this'll turn into a forum game.

Straight to the point, what's wrong with these two programs?

(1)

+ Code Snippet

#include <iostream>
using namespace std;

class Test
{
    public:
    int* pdata;
    Test();
    Test(int data);
    ~Test();
};
Test::Test() : pdata(NULL){}
Test::Test(int data) : pdata(NULL)
{
    pdata=new int(data);
}
Test::~Test()
{
    delete pdata;
}
int main(int argc, char* argv)
{
    Test mytest=Test(4);
    cout<<*(mytest.pdata)<<endl;
    mytest=Test(3);
    return 0;
}

result of running that code: memory error

(2)

+ Code Snippet

#include <iostream>
using namespace std;

int div(int *divisor)
{
    int result=5;
    result=result/*divisor;    /*divide the numbers*/;
    *divisor=1;
    return result;
}

int main(int argc, char* argv)
{
    int num=5;
    cout<<"Division: 5/5= "<<div(&num)<<endl;
    return 0;
}

result: "Divison: 5/5= 5"

does anyone else have some good, confusing bugs?

Tell me if there's a broken link to images in a thread I post, and I'll fix 'em.

Back to top

Profile PM Email

Benjamin

22

Years of Service

User Offline

Joined: 24th Nov 2002

Location: France

Posted: 25th May 2011 22:29 Edited at: 25th May 2011 22:31

Link

It's because the temporary copy of the object has its destructor called the moment it is no longer of use (ie. after the assignment), which is freeing the memory. If you need to make a copy of an object that allocates memory (as you're doing) you should implement a proper copy constructor. See this link:

http://en.wikipedia.org/wiki/Rule_of_three_(C%2B%2B_programming)

An alternative is to declare the variable like this, which avoids the need to copy anything:

+ Code Snippet

Test mytest(4);

Support a charitable indie game project!

Back to top

Profile PM Email

budokaiman

FPSC Tool Maker

15

Years of Service

User Offline

Joined: 24th Jun 2009

Playing: Hard to get

Posted: 25th May 2011 22:37

Link

The second one you need a space between the / and * else it's a comment block.

Math is like a toilet. They both go in the other direction in Australia.

Back to top

Profile PM Email Website

David R

21

Years of Service

User Offline

Joined: 9th Sep 2003

Location: 3.14

Posted: 25th May 2011 22:54 Edited at: 25th May 2011 22:57

Link

This isn't a bug, but the following (annoying) question came up in my systems programming exam:

+ Code Snippet

int k = 23;
int* kp = &k;

//What does the line below do?
*kp++=0;

Hint:

+ Code Snippet

Look up the operator precedence table. It probably doesn't do 
what you first assume it does (unless you've seen this trick before)

09-f9-11-02-9d-74-e3-5b-d8-41-56-c5-63-56-88-c0

Back to top

Profile PM Email Website

budokaiman

FPSC Tool Maker

15

Years of Service

User Offline

Joined: 24th Jun 2009

Playing: Hard to get

Posted: 26th May 2011 00:15

Link

Wouldn't it set *kp to 0, then increase kp.

Math is like a toilet. They both go in the other direction in Australia.

Back to top

Profile PM Email Website

Neuro Fuzzy

17

Years of Service

User Offline

Joined: 11th Jun 2007

Location:

Posted: 26th May 2011 00:30

Link

it's the same as (*(kp++))=0 I think, sooo you get three bytes of data of *kp, and one byte of jibberish after that, and set all those bytes to 0

Tell me if there's a broken link to images in a thread I post, and I'll fix 'em.

Back to top

Profile PM Email

budokaiman

FPSC Tool Maker

15

Years of Service

User Offline

Joined: 24th Jun 2009

Playing: Hard to get

Posted: 26th May 2011 00:51 Edited at: 26th May 2011 00:52

Link

Quote: "three bytes of data of *kp, and one byte of jibberish"

I'm pretty sure that the compiler takes type size into consideration when increasing. if you had int* kp = 0x00000000 and did kp++, it wouldn't be 0x00000001, it would be 0x00000004.

So all of the data that you get would be garbage, not just some.

Math is like a toilet. They both go in the other direction in Australia.

Back to top

Profile PM Email Website

Interplanetary Funk

15

Years of Service

User Offline

Joined: 19th Apr 2010

Location: Ipswich, United Kingdom

Posted: 26th May 2011 01:18

Link

Has no one realised that in neuros second bug, he's actually trying to divide by the address of a variable, not the variable itself....?

Visit my blog for any and all info on what I do coding wise.

Back to top

Profile PM

budokaiman

FPSC Tool Maker

15

Years of Service

User Offline

Joined: 24th Jun 2009

Playing: Hard to get

Posted: 26th May 2011 01:35

Link

Quote: "Has no one realised that in neuros second bug, he's actually trying to divide by the address of a variable, not the variable itself....?"

No he isn't...

Math is like a toilet. They both go in the other direction in Australia.

Back to top

Profile PM Email Website

Interplanetary Funk

15

Years of Service

User Offline

Joined: 19th Apr 2010

Location: Ipswich, United Kingdom

Posted: 26th May 2011 01:45

Link

Okay, now I'm confused then :S whenever I have a function with a pointer as an argument, you never put the * in front when you want to use it :S

Oh well, back to the books then.

Visit my blog for any and all info on what I do coding wise.

Back to top

Profile PM

Neuro Fuzzy

17

Years of Service

User Offline

Joined: 11th Jun 2007

Location:

Posted: 26th May 2011 02:23 Edited at: 26th May 2011 02:41

Link

if that were a + operator instead of a / operator that would be fine!

try putting the code into an IDE with syntax highlighting

[edit]
O_O
I just tried what I thought was correct for DarkCoder's post...

Apparently the ++ operator increases a value, but then returns the value before it was increased.
(so in: n=5; bool s=((n++)==5); t=(n==5) s is true, t is false)
that's tricky...

and you're right about pointers relying on the size of the datatype... apparently my pointer arithmetic is bad

Tell me if there's a broken link to images in a thread I post, and I'll fix 'em.

Back to top

Profile PM Email

Interplanetary Funk

15

Years of Service

User Offline

Joined: 19th Apr 2010

Location: Ipswich, United Kingdom

Posted: 26th May 2011 02:48

Link

n++ will return then increment the value. ++n will increment then return the value.

so in the following for loop, it would actually start counting from 1.

+ Code Snippet

for(int i=0; i<5; ++i)
    Function();

Visit my blog for any and all info on what I do coding wise.

Back to top

Profile PM

budokaiman

FPSC Tool Maker

15

Years of Service

User Offline

Joined: 24th Jun 2009

Playing: Hard to get

Posted: 26th May 2011 03:12

Link

Quote: "o in the following for loop, it would actually start counting from 1."

No, in for loops it will still always start from the initializing value.

Math is like a toilet. They both go in the other direction in Australia.

Back to top

Profile PM Email Website

Interplanetary Funk

15

Years of Service

User Offline

Joined: 19th Apr 2010

Location: Ipswich, United Kingdom

Posted: 26th May 2011 03:27 Edited at: 26th May 2011 03:31

Link

It didn't when I was using borland in college. mind you it was the ancient one.

It seems my understanding of C++ is being completely rebuilt tonight.

edit: revised using while loops:

+ Code Snippet

#include <iostream>
using namespace std;

int main(int argc, char* argv)
{
	cout<<"i:"<<endl;
	int i=0;
	while(i<=5)
	{
		cout<<i++<<endl;		
	}
	int j=0;
	cout<<"j:"<<endl;
	while(j<=5)
	{
		cout<<++j<<endl;		
	}
    return 0;
}

output:

+ Code Snippet

in j it starts at 1 because it increments it before outputting it.

Visit my blog for any and all info on what I do coding wise.

Back to top

Profile PM

AutoBot

15

Years of Service

User Offline

Joined: 25th Sep 2009

Location: Everywhere

Posted: 26th May 2011 04:25 Edited at: 26th May 2011 04:53

Link

Yeah, it's cool seeing how all the operators work.

The 2nd one that Neuro posted was pretty creative, didn't spot that comment mark. I would've seen it in an IDE, though.

EDIT: I actually did have a bug, but it involves code thats extremely big, lol. I was trying to port an OpenGL program to an object model, like this:

+ Code Snippet

//This is inside the singleton class startup function
InputManager.Start();
WindowManager.Start();
GraphicsManager.Start();

//do some coding

//Shutdown function
GraphicsManager.End();
WindowManager.End();
InputManager.End();

From the book I learned this technique from, it spoke of how this applied to games, but this could be a valid object model for any program.

But the bug was just a runtime error that I got at startup (access violation). My problem was that my singleton class (represents the program as a whole) was initializing the OpenGL subsystems before WinMain (I used a constructor), and that screwed things up. I changed it to a Start() member function and put it inside my entrypoint, and with tweaking, it worked.

Back to top

Profile PM Email

Sorry your browser is not supported!

Geek Culture / C++ debugging problems