I think I've found a bug in the sprintf function:
D:\src\ruby>ruby -v
ruby 1.8.2 (2004-12-25) [i386-mswin32]
D:\src\ruby>irb
irb(main):001:0> sprintf("%.1f", 123456789012345678.to_f)
=> "123456789012345680.0"
irb(main):002:0> sprintf("%.1f", 123456789012345678.to_f).to_i
=> 123456789012345680
Notice that that sprintf changed the value by an order of magnitude...
Justin
jwesley wrote:
I think I've found a bug in the sprintf function:
D:\src\ruby>ruby -v
ruby 1.8.2 (2004-12-25) [i386-mswin32]D:\src\ruby>irb
irb(main):001:0> sprintf("%.1f", 123456789012345678.to_f)
=> "123456789012345680.0"
irb(main):002:0> sprintf("%.1f", 123456789012345678.to_f).to_i
=> 123456789012345680Notice that that sprintf changed the value by an order of magnitude...
Justin
FP imprecision changed the value by 2.0:
irb(main):008:0> orig = 123456789012345678
=> 123456789012345678
irb(main):009:0> sprintf("%.1f", orig.to_f).to_i - orig
=> 2
and:
irb(main):012:0> orig - 2.0 == orig
=> true
irb(main):013:0> orig + 2.0 == orig
=> true
irb(main):014:0> orig.to_f.to_i - orig
=> 2
It's not sprintf's fault.
--
vjoel : Joel VanderWerf : path berkeley edu : 510 665 3407
jwesley wrote:
I think I've found a bug in the sprintf function:
D:\src\ruby>ruby -v
ruby 1.8.2 (2004-12-25) [i386-mswin32]D:\src\ruby>irb
irb(main):001:0> sprintf("%.1f", 123456789012345678.to_f)
=> "123456789012345680.0"
irb(main):002:0> sprintf("%.1f", 123456789012345678.to_f).to_i
=> 123456789012345680Notice that that sprintf changed the value by an order of magnitude...
Justin
Nope, not a Ruby bug:
/* Solaris 10 */
#include <stdio.h>
int main(){
printf("%.1f\n", 123456789012345678.0);
return 0;
}
djberge@~/programming/C-528>gcc -Wall -o printftest printftest.c
djberge@~/programming/C-529>./printftest
123456789012345680.0
Regards,
Dan
Two more helpless kittens, struck down before their prime.
Please think of the kittens.
On 11/11/05, jwesley <justin.w.smith@gmail.com> wrote:
irb(main):001:0> sprintf("%.1f", 123456789012345678.to_f)
=> "123456789012345680.0"
irb(main):002:0> sprintf("%.1f", 123456789012345678.to_f).to_i
=> 123456789012345680
--
Rob
You're right, it change the value by 2...
Quoting Rob Rypka <rascal1182@gmail.com>:
> irb(main):001:0> sprintf("%.1f", 123456789012345678.to_f)
> => "123456789012345680.0"
> irb(main):002:0> sprintf("%.1f", 123456789012345678.to_f).to_i
> => 123456789012345680Two more helpless kittens, struck down before their prime.
Four, by my count.
Please think of the kittens.
I think God makes exceptions for test cases, though.
-mental
On 11/11/05, jwesley <justin.w.smith@gmail.com> wrote:
Is there a standard way to get a more precise floating point number?
Hi,
At Sat, 12 Nov 2005 07:32:13 +0900,
jwesley wrote in [ruby-talk:165361]:
Is there a standard way to get a more precise floating point number?
BigDecimal might help you.
--
Nobu Nakada
Short answer. BigDecimal: http://www.ruby-doc.org/stdlib/libdoc/bigdecimal/rdoc/index.html
Longer rambling answer and reason why it's not a bug in anything, just a design choice: IEEE 754 double-precision (64-bit) floating point numbers use 12 bits for sign and exponent, and 52 bits for the significand. Therefore, the smallest integer that can't be exactly represented as a float is 2^53+1
123456789012345678
9007199254740992 (2^53 + 1)
So the number used for the test is fairly obviously larger than that value, and there's going to be some sort of error if you flip it around and use it. Ferinstance:
irb(main):008:0> (2**53 + 1)
=> 9007199254740993
irb(main):009:0> (2**53 + 1).to_f.to_i
=> 9007199254740992
And look at *this* consequence of the inaccuracy:
irb(main):029:0> i = (2**53 - 1).to_f
=> 9.00719925474099e+15
irb(main):030:0> while i < (2**53 + 2)
irb(main):031:1> puts "#{i.to_i}"
irb(main):032:1> i += 1.0
irb(main):033:1> end
9007199254740991
9007199254740992
...
Fun stuff, eh?
It's also worth noting that depending on the hardware instruction, you can get different results for the same calculation. And since we don't program in assembly, you can get different results depending on the compiler, or potentially compiler options, while still being entirely consistent with the spec.
matthew smillie
On Nov 11, 2005, at 22:32, jwesley wrote:
Is there a standard way to get a more precise floating point number?
I'd recommended the following document if you really want to know about floating point.
And if you haven't read this document (or something like it) then you probably shouldn't
be writing programs (well, any *serious* programs) using floating point.
What Every Computer Scientist Should Know About Floating-Point Arithmetic
http://docs.sun.com/source/806-3568/ncg_goldberg.html
Gary Wright
On Nov 11, 2005, at 6:58 PM, Matthew Smillie wrote:
On Nov 11, 2005, at 22:32, jwesley wrote:
Is there a standard way to get a more precise floating point number?
Short answer. BigDecimal: http://www.ruby-doc.org/stdlib/libdoc/bigdecimal/rdoc/index.html