Initialising unsigned integer pointer location in memory

[sigur47]

Hi i am trying to initialise my bitcontents in my structure to zero while making a new structure.My current method is not working .Any suggestions

/ a bit vector struct
    struct bitset{
        int size;
        unsigned int *bitcontents;
        char buffer[27];


    };
 /*initialisng a bitset struct to the size of the elements
 */
  struct bitset * bitset_new(int size){
      struct bitset * p=  ( struct  bitset *)  malloc(sizeof(*p));
      p->size=size;
          p->bitcontents=  (unsigned int*) malloc( (sizeof(int) *  size/8)  + (size % 8)? 1:0) ;

   int i=0;
          while(i<31){
          p->bitcontents[i]=0;
          i++;

          }

          return p;
  }

[Tclausex]

Not sure what errors you're getting, but...

What is sizeof(*p) when not only has p not been declared yet, but even if it did, what would p be pointing to to dereference? Rethink Line 12.

Line 14 ? So you are either allocating 1 byte or 0 bytes to assign to p->bitcontents ?

[MK27]

There are standard functions to do this kind of thing, qv. memset().

FYI, though it would be bad practice to count on it, modern operating systems always zero out heap memory anyway (partially for security and because of virtual addressing).

But your method should work. Why do you believe it has not?

You should determine the size of bitcontents and use it consistently -- you have a formula in the malloc, then in the while loop you just hardcode 31. Tch.

I notice you are using casts with malloc(). Are you still using a C++ compiler, as per bit abstract data type Why?

[iMalc]

What is sizeof(*p) when not only has p not been declared yet, but even if it did, what would p be pointing to to dereference? Rethink Line 12.

Line 14 ? So you are either allocating 1 byte or 0 bytes to assign to p->bitcontents ?

There's nothing wrong with sizeof(*p) In fact it's the best method for getting he size correct and writing maintainable code that there is. sizeof never cares when it's written as it doesn't exceute at runtime anyway, it executes at compile-time.

Line 14, not he's allocating many bytes, and adding 1 or not just to round up. Now both of those lines do have issues, but they are not what you commented on.

OP: You do need to explain what "not working" means. Not compiling? Not running? Not giving the right result?
You should not be casting the return value of malloc.
The easier way to round up is (size+7)/8
Why is there a 31 in that code? You don't know that the buffer allocated is even this big.
What is buffer for, it is unused?

[Tclausex]

There's nothing wrong with sizeof(*p) In fact it's the best method for getting he size correct and writing maintainable code that there is. sizeof never cares when it's written as it doesn't exceute at runtime anyway, it executes at compile-time.

I was not aware of that. Thanks. And yeah, I misread the ternary.

[sigur47]

There's nothing wrong with sizeof(*p) In fact it's the best method for getting he size correct and writing maintainable code that there is. sizeof never cares when it's written as it doesn't exceute at runtime anyway, it executes at compile-time.

Line 14, not he's allocating many bytes, and adding 1 or not just to round up. Now both of those lines do have issues, but they are not what you commented on.

OP: You do need to explain what "not working" means. Not compiling? Not running? Not giving the right result?
You should not be casting the return value of malloc.
The easier way to round up is (size+7)/8
Why is there a 31 in that code? You don't know that the buffer allocated is even this big.
What is buffer for, it is unused?

when i try to initialise my bitcontents with 0.it does not work thats what i meant by the code not working.I am sorry for not beeing clear

[sigur47]

There are standard functions to do this kind of thing, qv. memset().

FYI, though it would be bad practice to count on it, modern operating systems always zero out heap memory anyway (partially for security and because of virtual addressing).

But your method should work. Why do you believe it has not?

You should determine the size of bitcontents and use it consistently -- you have a formula in the malloc, then in the while loop you just hardcode 31. Tch.

I notice you are using casts with malloc(). Are you still using a C++ compiler, as per bit abstract data type Why?

ye i think its from the compiler cause it gave me a warning..When i put in zero to my bitcontents for 32 bits.. it initialises some index with zero but some are still left with junk in it. e.g this is my output when i print it out
0000260108816100000026000000000.. it should be alll zeros