From @masak

Rakudo r32568​:

$ ./perl6 -e 'my $result = log(0); say "Still alive after getting $result"'
Still alive after getting -inf

S29​:

=item log

our Num multi method log ( Num $x​: Num :$base = Num​::e ) is export

Logarithm of base C<$base>, default Natural. Calling with C<$x == 0> is an
error.

(Since log is periodic along the imaginary axis, an error should
perhaps be thrown for other complex values as well, provided that
enough precision can be achieved to make this a problem.)

From @pmichaud

On Tue, Nov 11, 2008 at 11​:57​:02PM -0800, Carl Mäsak wrote​:

Rakudo r32568​:

$ ./perl6 -e 'my $result = log(0); say "Still alive after getting $result"'
Still alive after getting -inf

Now fixed in r32574​:

  $ ./parrot perl6.pbc -e 'my $result = log(0); say $result;'
  Can't take log of 0
 
  $

Note that instead of throwing an exception immediately, log(0)
returns a Failure that then throws an exception if it is used.

Pm

The RT System itself - Status changed from 'new' to 'open'

@pmichaud - Status changed from 'open' to 'resolved'

From @zoffixznet

Looks like this bug re-appeared and Rakudo's function all come down to nqp​::log_n()​:

  <ZoffixW> m​: say log 0
  <camelia> rakudo-moar 2dd0dd​: OUTPUT«-Inf␤»
  <ZoffixW> m​: use nqp; say nqp​::log_n(0e0)
  <camelia> rakudo-moar 2dd0dd​: OUTPUT«-Inf␤»

2008 IEEE 754[^1], in section 9.2, says log(0) needs to give a divideByZero exception.

[1] http://www.csee.umbc.edu/~tsimo1/CMSC455/IEEE-754-2008.pdf

On Wed Nov 12 07​:53​:52 2008, pmichaud wrote​:

On Tue, Nov 11, 2008 at 11​:57​:02PM -0800, Carl Mäsak wrote​:

Rakudo r32568​:

$ ./perl6 -e 'my $result = log(0); say "Still alive after getting
$result"'
Still alive after getting -inf

Now fixed in r32574​:

$ ./parrot perl6.pbc -e 'my $result = log(0); say $result;'
Can't take log of 0

$

Note that instead of throwing an exception immediately, log(0)
returns a Failure that then throws an exception if it is used.

Pm

@zoffixznet - Status changed from 'resolved' to 'open'

From @zoffixznet

I should've read more​:

"Section 7.3 Division by zero", says 'For logB(0) when logBFormat is a floating-point format, the sign of the infinity is minus (−∞).'

And I'm guessing this case matches it, because JVM and C standard lib too give -Inf and nqp just maps the operations to VM's operations.

Reclosing :)

On Mon Oct 17 05​:58​:14 2016, cpan@​zoffix.com wrote​:

Looks like this bug re-appeared and Rakudo's function all come down to
nqp​::log_n()​:

<ZoffixW> m​: say log 0
<camelia> rakudo-moar 2dd0dd​: OUTPUT«-Inf␤»
<ZoffixW> m​: use nqp; say nqp​::log_n(0e0)
<camelia> rakudo-moar 2dd0dd​: OUTPUT«-Inf␤»

2008 IEEE 754[^1], in section 9.2, says log(0) needs to give a
divideByZero exception.

[1] http://www.csee.umbc.edu/~tsimo1/CMSC455/IEEE-754-2008.pdf

On Wed Nov 12 07​:53​:52 2008, pmichaud wrote​:

On Tue, Nov 11, 2008 at 11​:57​:02PM -0800, Carl Mäsak wrote​:

Rakudo r32568​:

$ ./perl6 -e 'my $result = log(0); say "Still alive after getting
$result"'
Still alive after getting -inf

Now fixed in r32574​:

$ ./parrot perl6.pbc -e 'my $result = log(0); say $result;'
Can't take log of 0

$

Note that instead of throwing an exception immediately, log(0)
returns a Failure that then throws an exception if it is used.

Pm

@zoffixznet - Status changed from 'open' to 'resolved'