Bug #112826 for Math-BigFloat: Precision does not behave as documented (doc confusion or bug)

Hi, The pod for Math::BigFLoat has a section "ACCURACY and PRECISION". It contains a table giving examples of different Precision values and how they affect rounded results. By my reading, it shows completely wrong values, essentially opposite to actual behavior. Of course, I could be interpreting the table backwards somehow; if so, please correct me and consider adding something to the docs so others won't have the same confusion as me. Here's the problem I see: The pod says this: Initial value P A Result String ------------------------------------------------------------ 1234.01 -3 1000 1000 1234 -2 1200 1200 1234.5 -1 1230 1230 1234.001 1 1234 1234.0 1234.01 0 1234 1234 1234.01 2 1234.01 1234.01 1234.01 5 1234.01 1234.01000 But actual behavior is this: Initial val P String 1234.01 -3 1234.010 1234 -2 1234.00 1234.5 -1 1234.5 1234.001 1 1234 1234.01 2 1230 1234.01 5 0 And here is the program used to generate the "actual" results: #!/usr/bin/perl use strict; use warnings; use Math::BigFloat; printf "%-11s %2s %s\n", "Initial val", "P", "String"; foreach( [ "1234.01", -3 ], [ "1234", -2 ], [ "1234.5", -1 ], [ "1234.001", 1 ], [ "1234.01", 2 ], [ "1234.01", 5 ], ) { my ($init_str, $P) = @$_; my $x = Math::BigFloat->new($init_str); $x->round(undef,$P); printf "%-11s %2d %s\n", $init_str, $P, scalar($x); }