Function riddle (C/C++)

[DeFender1031]

Who can tell me what this function is designed to do?

Code: Select all

void calculate(unsigned int number, int cow, char bleh, double three, long double eight)
{
  printf("%u\n",number);
  calculate(number+1, cow+5, three+2, eight, bleh-1);
}

[byuu]

You mean other than cause a runtime stack overflow exception? :P

Let's cut through the 'fluff' ...

Code: Select all

void calculate(int x, int y, int z) {
  printf("%2d %2d %2d\n", x, y, z);

//x, y, z -> y, z, x = rotate variables left by 1 (3-step loop)

//x = y + 2
//y = z
//z = x - 1

  calculate(y + 2, z, x - 1);
}

If x were y + 1, I'd think it was a rotation of some sort. Because it's +2, it causes all three values to increment by one every three iterations. This also causes the sum of (x,y,z) to always increment by one each recursive call (eg 2+0-1=1). What the iterations mean in the interim two of three steps though, I'm not sure.

Log of the first few runs:

Code: Select all

 0  0  0= 0
 2  0 -1= 1
 2 -1  1= 2

 1  1  1= 3
 3  1  0= 4
 3  0  2= 5

 2  2  2= 6
 4  2  1= 7
 4  1  3= 8

 3  3  3= 9
 5  3  2=10
 5  2  4=11

 4  4  4=12
 6  4  3=13
 6  3  5=14

[funkyass]

an expensive way of counting.

[whicker]

A crappy IIR filter with most of the coefficients (and input value) set to 1?

[Nach]

Cute, I think you meant to have that first param be unsigned and change it to %u though.

[DeFender1031]

Blah Blah Blah Code Blah Blah

Seriously, why are you making up your own code instead of telling me what mine does? You removed some key variables, you're printing totally different things than mine is... I'll give you one thing, you DID pull an interesting pattern out of some variables there.

Cute, I think you meant to have that first param be unsigned and change it to %u though.

Oh shoot... gotta fix that... *Goes back and edits the original post*

[odditude]

Is it safe to assume that 'char' ranges from -128 to 127, unsigned int from 0 to 65535, and int from -32768 to 32767?

If so, life will get a little more interesting after the original value of "bleh" rolls over (assuming there isn't an overflow error instead of a simple rollover).

Regardless, it still seems that all the program will print is the numbers from 0 to 65535, one per line, repeating ad infinitum, starting from whatever the initial value of "number" is.

[Nach]

Is it safe to assume that 'char' ranges from -128 to 127, unsigned int from 0 to 65535, and int from -32768 to 32767?

No.

[funkyass]

for those that don't know:

unsigned int is 0 -> +4,294,967,295 , int is -2,147,483,648 -> +2,147,483,647 unsigned char is 0-255, signed char is -128-127

this is a C++ function, what the sizes of the data types are is Dependant on the compile target, but for the sake of sanity assume its x86.

[DeFender1031]

Regardless, it still seems that all the program will print is the numbers from 0 to 65535, one per line, repeating ad infinitum, starting from whatever the initial value of "number" is.

That is incorrect. And would still be incorrect if you had said 0 to 4,294,967,295.

[byuu]

Seriously, why are you making up your own code instead of telling me what mine does? You removed some key variables, you're printing totally different things than mine is... I'll give you one thing, you DID pull an interesting pattern out of some variables there.

1) Because I don't know what your code does.
2) Because you don't ever actually use either cow or number for anything other than a printf statement.
3) Because that's the way I analyze problems.
4) I have to guess what inputs to give the first calculate call, since you never specified start values, hence why we have different outputs. The last three vars should be the same for both of ours. I only changed their names.

But if you're going to get upset over it, I'll move on. I don't know what it does. It looks like gibberish to me with no context. Maybe someone else can solve it, then.

[funkyass]

assuming the function in question would actually compile...

stuffing a char into a long double might make the compiler question the sanity of the programmer. Putting a double into a char would make the compiler phone the police.

[Nach]

Okay, since people are getting upset over this, I'll give away the answer, and this should have been obvious (at least it was to me when I first saw it).

This function is recursive, and seems to run until it crashes, each time printing out the value of number which increments by one per itteration. The other parameters don't seem to do anything but take up stack space.

So assuming that the first call to the function is with number being equal to 0/1, it counts how many times a recursive function can call itself with 2 ints, a char, a double, and a long double.

You should be safe to have recursion go up to 100,000 times or so on modern PCs.

[DeFender1031]

1) Because I don't know what your code does.
2) Because you don't ever actually use either cow or number for anything other than a printf statement.
3) Because that's the way I analyze problems.
4) I have to guess what inputs to give the first calculate call, since you never specified start values, hence why we have different outputs. The last three vars should be the same for both of ours. I only changed their names.

But if you're going to get upset over it, I'll move on. I don't know what it does. It looks like gibberish to me with no context. Maybe someone else can solve it, then.

1) Looking at a fridge won't help you figure out how an internal combustion engine works.
2) Incorrect, every variable there is vital.
3) That's your prerogative
4) true

I'm not upset. I do not get upset. I was just wondering why you think a fridge will help.

assuming the function in question would actually compile...

stuffing a char into a long double might make the compiler question the sanity of the programmer. Putting a double into a char would make the compiler phone the police.

No, it compiles fine, no warnings even.

Okay, since people are getting upset over this, I'll give away the answer, and this should have been obvious (at least it was to me when I first saw it).

This function is recursive, and seems to run until it crashes, each time printing out the value of number which increments by one per itteration. The other parameters don't seem to do anything but take up stack space.

So assuming that the first call to the function is with number being equal to 0/1, it counts how many times a recursive function can call itself with 2 ints, a char, a double, and a long double.

You should be safe to have recursion go up to 100,000 times or so on modern PCs.

correct. It's to test how much stack space there is. You win the prize, i WON'T come over to your house and kill you. Thank you for playing.

[Deathlike2]

Bah to infinite recursion. Could've seen that coming w/o having really guessed the output/behavior of the function itself.

[Nightcrawler]

Lame riddle. The answer is a crashing program...

[byuu]

correct. It's to test how much stack space there is.

Easy enough to do that without the need for all the variables. Then you can extrapolate by testing sizeof() on each type you want. I did this when choosing recommended defaults for thread heap sizes.

For 32-bit platforms and IIRC, DJGPP gives 512KB, Visual C++ and MinGW give 1MB; unless of course you override the default stack sizes.

I used something like:

Code: Select all

void heapsize() {
static int size = 0;
  size += sizeof( void (*)() );
  printf("heap ~= %d bytes\n", size);
  heapsize();
}

If I'm still missing what you're doing, then I give up.

1) Looking at a fridge won't help you figure out how an internal combustion engine works.

Analogies aside, I think I typically do quite well reverse engineering stuff. Of course, I can't even compare to Nach, blargg or anomie. Which is actually kind of nice, I like having others to seek advice from and aspire to.

But who knows, I'm sure my technique could use a lot more work. And sorry, your first reply came off like you were perturbed by my attempt.

Lame riddle. The answer is a crashing program...

Bah, seconded. I was hoping it would hold a special mathematic property or something.

[grinvader]

I was hoping it would hold a special mathematic property or something.

Code: Select all

typedef enum { knf, sw1, sw2, ax1, ax2, ham, spr, stf, fll, bow, glv, NA} wname;

struct
{
  uint8_t inc[7];
  bool used;
} weap[11];

static void next_step(uint8_t table[11], const wname last)
{
  wname lnz = last;
  uint8_t tmp;

  while (!table[lnz] || !weap[lnz].used) { lnz--; }

  if (lnz == last)
  {
    tmp = table[lnz];
    table[lnz] = 0;

    do { lnz--; } while (!table[lnz] || !weap[lnz].used);
  }
  else { tmp = 0; }

  table[lnz]--;

  do { lnz++; } while (!weap[lnz].used);
  table[lnz] = tmp+1;
}

Have fun.