Why are the values I am pulling from my ruby array to my c extension wrong?

Question

This method is just verifying that I'm able to see the elements of a ruby array correctly.

static VALUE 
print_cards(self) 
  VALUE self;
{
    VALUE cards;
    int i;

    cards = rb_ivar_get(self, rb_intern("@cards"));
    VALUE *ary_ptr = RARRAY_PTR(cards);
    int ary_length = RARRAY_LEN(cards);

    for(i=0; i< ary_length; i++)
        printf("%d\n", ary_ptr[i]);

  return Qnil;
}

void Init_ev() {
    rb_eval_string("require './lib/ev/pair_counter'");
    VALUE PairCounter = rb_path2class("EV::PairCounter");
    rb_define_method(PairCounter, "print_cards", print_cards, 0);
}

But when I put the method to use, the elements of the array are wrong. The strange thing is that it doesn't look like I'm getting some kind of address information, since the size of the number that is printed roughly corresponds with the size of the number in the ruby array. The numbers are also consistent each time I create a new object and run print_cards.

ruby-1.9.2-p180 :001 > p = EV::PairCounter.new   #=> #<EV::PairCounter:0x000001046a10f8 @pairs={}, @cards=[]>
ruby-1.9.2-p180 :002 > p.add_card(1)   #=> 1
ruby-1.9.2-p180 :003 > p.print_cards
3                                      #=> nil
ruby-1.9.2-p180 :004 > p.add_card(5)   #=> 2
ruby-1.9.2-p180 :005 > p.add_card(88)   #=> 3
ruby-1.9.2-p180 :006 > p
=> #<EV::PairCounter:0x000001046a10f8 @pairs={1=>1, 5=>1, 88=>1}, @cards=[1, 5, 88]>
ruby-1.9.2-p180 :007 > p.print_cards
3
11
177

score 6 · Answer 1 · answered Jun 22 '11 at 22:16

6

I needed to use

printf("%d\n", NUM2INT(ary_ptr[i]));

answered Jun 22 '11 at 22:16

Jeremy Smith

14,727
19
67
114

The entries in `ary_ptr` are actually `VALUE`s, right? Have you considered writing a C extension tutorial? – mu is too short Jun 22 '11 at 22:40
1

If you look, the numbers you were getting were all (2*x) + 1, where x is the number you were expecting, or equivalently left shift by one and set the low bit. This is because they are Fixnums, which are immediate values. Fixnums are stored as the 31 (or 63) bit value formed from shifting the bits left by one and setting the low bit to one. This means that you can test whether a `VALUE` is a Fixnum or not by simply checking the low bit. Since all memory references are aligned on 4 (or 8) bytes, the 2 low bits will always be 0 for a "normal" object, so they are used for immediate values... – matt Jun 23 '11 at 00:08
I could once I get a little further along mu. I'm still only a day into working with c extensions, and I don't have a blog up yet. And yeah, I assume the entries of ary_ptr are VALUES and they have to be cast once you access them. – Jeremy Smith Jun 23 '11 at 00:08
... such as Fixnums, true, false and nil. _Programming Ruby_ (the pickaxe book) has a pretty good section on this if you're going to be doing a lot of it. – matt Jun 23 '11 at 00:11
Thanks @matt, yeah I have the pickaxe chapter on this bookmarked and have gone over it a few times. – Jeremy Smith Jun 23 '11 at 00:22

score 0 · Answer 2 · edited May 23 '17 at 12:29

0

rb_ary_entry is the safe way to get the content from a Ruby array. They are not accessed like normal C arrays.

Seems to be related to this question: https://stackoverflow.com/a/9619163/486990

edited May 23 '17 at 12:29

Community

1
1

answered Mar 09 '12 at 15:19

thomthom

2,854
1
23
53

Why are the values I am pulling from my ruby array to my c extension wrong?

2 Answers2