drand48 returning 0's and nan

I'm having trouble using drand48 to feed an array, i tested the array using the loop counter and it stores that just fine.
When I use drand48 though, i get 7 0's in the first 7 spots, and the rest i get nan

Code:

/*
*  Program to write an array to a file
*  Array will be filled with random numbers using drand48 (0.0, 1.0)
*/

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

int main(int argc, char *argv[])
{
   FILE *ofp;
   int i, N;
   double *A;

  srand48((long)time(NULL));
   if (argc != 2)
   {
      printf("Error with arguments, must have 2, program name and a integer for array");
      exit(1);
   }
   N = atoi(argv[1]);

   A = malloc((N + 1) * sizeof(double));

   ofp = fopen("arrout.dat", "w");

   fprintf(ofp, "Array size: %d\n", N);

   for(i = 0; i < N; ++i)
   {
      A[i] = drand48();
      fprintf(ofp, "%lf\n", A[i]);
   }

   return 0;
}

I note that srand48 and drand48 are not part of the C standard library. Check that they really are declared in <stdlib.h>, otherwise you will need to find out what is the correct header to include (EDIT: or macro name to define, or just forward declare them yourself).

Could it be because you're using %lf in fprintf whereas you should be using %f ?

Originally Posted by oogabooga

Could it be because you're using %lf in fprintf whereas you should be using %f ?

%lf is a valid format specification (at least since C99), though it is redundant.

I note that srand48 and drand48 are not part of the C standard library.

O_o

I recall that they are POSIX where they will be in "stdlib.h" because the POSIX standard likes to pull double duty from C standard headers for historical reasons.

Soma

Originally Posted by laserlight

%lf is a valid format specification (at least since C99), though it is redundant.

Yeah, I should have looked that up first.

As for your point, it turns out that drand48 (and srand48) are declared in stdlib.h.
I guess it's okay to add things to a standard library header.

Originally Posted by laserlight

I note that srand48 and drand48 are not part of the C standard library. Check that they really are declared in <stdlib.h>, otherwise you will need to find out what is the correct header to include (EDIT: or macro name to define, or just forward declare them yourself).

Code:

NAME
       drand48,  erand48, lrand48, nrand48, mrand48, jrand48, srand48, seed48,
       lcong48 - generate uniformly distributed pseudo-random numbers

SYNOPSIS
       #include <stdlib.h>

yeah it's in there looks like. I tried seeding a different way and got the same results as well..

Originally Posted by phantomotap

I recall that they are POSIX where they will be in "stdlib.h" because the POSIX standard likes to pull double duty from C standard headers for historical reasons.

Indeed, hence I was surprised to discover that it was not so for me. It is plausible that it may require some macro or something (compiler option?) to "enable" these extensions.

Originally Posted by Destinps

yeah it's in there looks like. I tried seeding a different way and got the same results as well..

If you are showing me a manual page, don't bother: I already know, and that's why I actually tested your program because I couldn't see what was wrong (other than failure to call free(), which shouldn't matter for your problem). Did you compile at the highest warning level? What compiler are you using, on what OS?

Originally Posted by laserlight

Indeed, hence I was surprised to discover that it was not so for me. It is plausible that it may require some macro or something (compiler option?) to "enable" these extensions.

Compiling with full warning should ferret out that possibility.
With gcc, use:
gcc -W -Wall yourprog.c

using putty to connect to a school computer from home..
the school computer is running linux and i'm using gcc to compile
I'm new to C and Linux so I'm not sure what info is needed, but i compiled using gcc -ansi -g -o

Originally Posted by Destinps

using putty to connect to a school computer from home..
the school computer is running linux and i'm using gcc to compile
I'm new to C and Linux so I'm not sure what info is needed, but i compiled using gcc -ansi -g -o

Add the -W and -Wall flags to your gcc command.
Do you get any warnings doing that?

EDIT: I should mention that your program works fine for me.

Originally Posted by oogabooga

Add the -W and -Wall flags to your gcc command.
Do you get any warnings doing that?

EDIT: I should mention that your program works fine for me.

no warnings when compiling with those arguments, but if the program works for you that's good enough for me. I'm just coding it as practice for a test so it's not going to be turned in or anything.

Originally Posted by Destinps

the school computer is running linux and i'm using gcc to compile
I'm new to C and Linux so I'm not sure what info is needed

Right. Check your manual page again. On mine, I realised I missed:

Code:

   Feature Test Macro Requirements for glibc (see feature_test_macros(7)):

       All functions shown above: _SVID_SOURCE || _XOPEN_SOURCE

What this means is that if you define the _SVID_SOURCE or _XOPEN_SOURCE macros, then the functions specified will be declared in <stdlib.h>.

Originally Posted by Destinps

i compiled using gcc -ansi -g -o

You should add the flags oogabooga mentioned to increase the warning level. I usually also include -pedantic. So, including the macro that you should define, I would expect:

Code:

gcc -ansi -W -Wall -pedantic -D_XOPEN_SOURCE -g -o

(or -D_SVID_SOURCE, or whatever else you need to define)

Also, since you are going for -ansi without -std=c99, I think oogabooga's point about %f versus %lf may come into play, but I'm not that familiar with the ins and out of gcc options and the 1989/1990 version of the C standard to say for sure.

D_SVID_SOURCE worked, i read that in the man but had no idea how to comprehend it
thanks for the help people!