Re: [C400-L] Re: Testing malloc results for null on the 400 (was: HTTP Authentication / base64 encoder?)

[James Rich <james@xxxxxxxxxxx>]

David Gibbs wrote:
> Well, if it fails because it could not malloc some memory, would there 
> be any indication of WHY it couldn't malloc the memory?
>
> However, since I want to be a good C programmer (not that I'm really a C 
> programmer at all, just hacking at it right now) ... what would be the 
> best way to indicate that the routine could not malloc memory?
>
> The prototype for the function is: char *str_base64_encode(char *str)

Here is an idea that might make things easier for you.  Try changing the 
function prototype to this:

int str_base64_encode (char *str, char *result);

Then in the function do something like this with the malloc (sizeneeded 
is a numeric var that says how much memory you want to malloc):

#include <stdio.h>
#include <stdlib.h>
#include <errno.h>

int str_base64_encode (char *str, char *result)
{
  /* never malloc memory in a function that is used by a caller - use 
 realloc */
  realloc (result, sizeof (char) * sizeneeded);
  if (result == NULL)
  {
    return errno;
  }
  do your processing here
  ...
  return 0;
}

Then have your RPG program check what str_base64_encode() returns.  If 
non-zero, check it against ENOMEM.

James Rich