Comments on: Fun with printf http://boredzo.org/blog/archives/2006-11-17/fun-with-printf The personal weblog of Peter Hosey. Thu, 20 Nov 2008 16:48:51 +0000 http://wordpress.org/?v=2.5.1 By: Kevin Ballard http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-4758 Kevin Ballard Mon, 25 Dec 2006 09:00:34 +0000 http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-4758 "random memory address" Nothing random about it. It's printing, in octal, the integer value of the stack location above the second argument. According to the OS X ppc32 ABI (and I believe the i686 and ppc64, but I'm too lazy to check), this would correspond to the local variables area of the calling function (i.e. the function where this fprintf statement occurs). If there are no local variables, it will be in the saved registers area. So basically, it's printing the octal representation of the integer cast of sizeof(int) of a local variable, or if none, then one of the saved registers. I don't know what this stack location corresponds to in other ABIs, but it'll probably be something similar. But the point here is it's definitely not random. "random memory address"

Nothing random about it. It's printing, in octal, the integer value of the stack location above the second argument. According to the OS X ppc32 ABI (and I believe the i686 and ppc64, but I'm too lazy to check), this would correspond to the local variables area of the calling function (i.e. the function where this fprintf statement occurs). If there are no local variables, it will be in the saved registers area.

So basically, it's printing the octal representation of the integer cast of sizeof(int) of a local variable, or if none, then one of the saved registers.

I don't know what this stack location corresponds to in other ABIs, but it'll probably be something similar.

But the point here is it's definitely not random.

]]> By: Peter Hosey http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-870 Peter Hosey Fri, 17 Nov 2006 20:39:24 +0000 http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-870 Yes, that's the solution to the bug. But the point of this post was the diagnosis, not the fix. :) Thanks, though. Yes, that's the solution to the bug. But the point of this post was the diagnosis, not the fix. :)

Thanks, though.

]]> By: Hub http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-869 Hub Fri, 17 Nov 2006 20:27:21 +0000 http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-869 Just use '%%' to escape the '%' in a format string ;-) Just use '%%' to escape the '%' in a format string ;-)

]]> By: Peter Hosey http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-868 Peter Hosey Fri, 17 Nov 2006 20:14:52 +0000 http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-868 Correct! Correct!

]]> By: Augie Fackler http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-867 Augie Fackler Fri, 17 Nov 2006 20:13:55 +0000 http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-867 Oh, right, I'm a retard. It's printing a random memory address as an octal value. The f just randomly threw me. Oh, right, I'm a retard. It's printing a random memory address as an octal value. The f just randomly threw me.

]]> By: Peter Hosey http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-866 Peter Hosey Fri, 17 Nov 2006 20:09:11 +0000 http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-866 Close, but not quite. Close, but not quite.

]]> By: Augie Fackler http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-865 Augie Fackler Fri, 17 Nov 2006 20:04:12 +0000 http://boredzo.org/blog/archives/2006-11-17/fun-with-printf#comment-865 "% of" will be interpreted as a format string for a floating point value, thus printing an arbitrary point in memory as a float. "% of" will be interpreted as a format string for a floating point value, thus printing an arbitrary point in memory as a float.

]]>