'use bignum' scoping question

si_lence has asked for the wisdom of the Perl Monks concerning the following question:

Dear Monks,

I'm not sure how I can understand the output I'm getting from the sample program below. Maybe (probably) something quite simple that I'm missing..
I need to calculate square roots with more than the standard precision so I'm using use bignum. This works fine unless it is in a loop. Anybody care to explain?

use strict;
use warnings;

use Data::Dumper;
use bignum(a=>200);

my $s = sqrt(2);;
print Dumper(\$s);

for my $num (2..2) {
    $s = sqrt($num);
    print Dumper(\$s);
}
[download]

The output is:

$VAR1 = \bless( {
                   '_es' => '-',
                   '_m' => [
                             605715,
                             1152782,
                             9995050,
                             5927557,
                             5927505,
                             1322266,
                             9358314,
                             1497099,
                             1264412,
                             8507372,
                             4836055,
                             9702492,
                             2309122,
                             5013846,
                             4157273,
                             5343276,
                             8850387,
                             6210703,
                             7324784,
                             9737990,
                             7317667,
                             3769480,
                             9671875,
                             9807856,
                             7242096,
                             8801688,
                             7309504,
                             2135623,
                             1414
                           ],
                   '_a' => 200,
                   '_e' => [
                             199
                           ],
                   'sign' => '+'
                 }, 'Math::BigFloat' );
$VAR1 = \'1.4142135623731';
[download]

It looks as if the blessing of $s is lost in the loop, but I don't understand why.
si_lence (confused)
Update: Thanks everybody for pointing out the reason and the suggestions. I did read the documentation, but obviously glossed over the important part.

Comment on 'use bignum' scoping question Select or Download Code

Replies are listed 'Best First'.
Re: 'use bignum' scoping question by lamprecht (Friar) on Jan 10, 2010 at 23:00 UTC
Hi, from perldoc bignum: `Please note the following does not work as expected (prints nothing), +since overloading of '..' is not yet possible in Perl (as of v5.8.0): perl -Mbignum -le 'for (1..2) { print ref($_); }'` [download] this works: `for my $num (2..2) { $s = sqrt($num * 1); print Dumper(\$s); }` [download] Cheers, Christoph	[reply] [d/l] [select]
Re^2: 'use bignum' scoping question by ikegami (Patriarch) on Jan 11, 2010 at 01:02 UTC
And so does `for my $num (2..2) { $s = sqrt(Math::BigFloat->new($num)); print Dumper(\$s); }` [download] That's what bignum does to constants.	[reply] [d/l]
Re: 'use bignum' scoping question by syphilis (Archbishop) on Jan 10, 2010 at 23:17 UTC
Looks like a bignum bug to me. It's certainly not the behaviour I would expect, though I haven't checked to see what (if anything) the bignum documentation has to say on the matter. Update:Aaah ... but lamprecht has :-) Safest workaround would be to explicitly use Math::BigFloat: `use strict; use warnings; use Math::BigFloat; Math::BigFloat->accuracy(200); my $s = sqrt(Math::BigFloat->new(2)); print $s, "\n"; for my $num (2..2) { $s = sqrt(Math::BigFloat->new($num)); print $s; }` [download] Cheers, Rob	[reply] [d/l]
Re: 'use bignum' scoping question by Anonymous Monk on Jan 10, 2010 at 23:02 UTC
From bignum ...loaded and any constant number will be converted to an object* (Math::BigFloat for floats like 3.1415 and Math::BigInt for integers like 1234).* $num is not a constant, 0 is a constant, and 0 + $num is a Math::BigInt	[reply]