| Subject: | BigFloat bpi() NaN when global precision() set |
| Date: | Wed, 01 May 2013 06:00:24 +1000 |
| To: | bug-Math-BigInt [...] rt.cpan.org |
| From: | Kevin Ryde <user42 [...] zip.com.au> |
In Math::BigFloat 1.997 and recent debian i386 perl 5.14.3, if a global
Math::BigFloat->precision() has been set then bpi() returns NaN.
Eg. the program foo.pl below prints "NaN".
foo.pl asks for precision -100 to have 100 decimal places and I hoped
bpi(5) asking for pi to 5 places still would work in that case.
Nosing around the code I see bpi() calls bdiv() with a specified
accuracy, and it seems the usual _find_round_parameters() rejects a
simultaneous accuracy and precision in force.
I suppose if the precision() is less than the bpi number of digits then
take the lesser of the two, the same as results when an accuracy() is in
force and you ask bpi for more digits than that accuracy.
use strict;
use Math::BigFloat;
Math::BigFloat->precision(-100);
my $pi = Math::BigFloat->bpi(5);
print "$pi\n";