package Math::Int64;

use strict;
use warnings;

BEGIN {
our $VERSION = '0.54';
require XSLoader;
XSLoader::load('Math::Int64', $VERSION);
}

use warnings::register;

use constant MAX_INT64  => string_to_int64 ( '0x7fff_ffff_ffff_ffff');
use constant MIN_INT64  => string_to_int64 ('-0x8000_0000_0000_0000');
use constant MAX_UINT64 => string_to_uint64( '0xffff_ffff_ffff_ffff');

require Exporter;
our @ISA = qw(Exporter);
our @EXPORT_OK = qw(int64
                    int64_to_number
                    net_to_int64 int64_to_net
                    native_to_int64 int64_to_native
                    string_to_int64 hex_to_int64
                    BER_to_int64 int64_to_BER
                    int64_to_string int64_to_hex
                    int64_rand
                    int64_srand
                    uint64
                    uint64_to_number
                    net_to_uint64 uint64_to_net
                    native_to_uint64 uint64_to_native
                    string_to_uint64 hex_to_uint64
                    BER_to_uint64 uint64_to_BER
                    uint64_to_string uint64_to_hex
                    uint64_rand
                    BER_length
                    MAX_INT64 MIN_INT64 MAX_UINT64
                  );

my %available_pragmas = map { $_ => 1 } qw(native_if_available
                                           die_on_overflow);

sub import {
    my $pkg = shift;
    my (%pragmas, @subs, %native);
    for (@_) {
        if ($_ =~ /^:(.*)/ and $available_pragmas{$1}) {
            $pragmas{$1} = 1
        }
        else {
            push @subs, $_;
        }
    }

    if ($pragmas{die_on_overflow}) {
        require Math::Int64::die_on_overflow;
        Math::Int64::die_on_overflow->import;
    }

    if ($pragmas{native_if_available}) {
        require Math::Int64::native_if_available;
        Math::Int64::native_if_available->import;
    }

    Math::Int64->export_to_level(1, $pkg, @subs);
}

sub _check_pragma_compatibility {
    if ($^H{'Math::Int64::native_if_available'} and
        $^H{'Math::Int64::die_on_overflow'} and
        warnings::enabled()) {
        warnings::warn("Math::Int64::die_on_overflow pragma is useless when Math::Int64::native_if_available is also active");
    }
    1;
}

use overload ( '+' => \&_add,
               '+=' => \&_add,
               '-' => \&_sub,
               '-=' => \&_sub,
               '*' => \&_mul,
               '*=' => \&_mul,
               '**' => \&_pow,
               '**=' => \&_pow,
               '/' => \&_div,
               '/=' => \&_div,
               '%' => \&_rest,
               '%=' => \&_rest,
               'neg' => \&_neg,
               '++' => \&_inc,
               '--' => \&_dec,
               '!' => \&_not,
               '~' => \&_bnot,
               '&' => \&_and,
               '|' => \&_or,
               '^' => \&_xor,
               '<<' => \&_left,
               '>>' => \&_right,
               '<=>' => \&_spaceship,
               '>' => \&_gtn,
               '<' => \&_ltn,
               '>=' => \&_gen,
               '<=' => \&_len,
               '==' => \&_eqn,
               '!=' => \&_nen,
               'bool' => \&_bool,
               '0+' => \&_number,
               '""' => \&_string,
               '=' => \&_clone,
               fallback => 1 );

package # hide from PAUSE since it also has its own .pm file
    Math::UInt64;
use overload ( '+' => \&_add,
               '+=' => \&_add,
               '-' => \&_sub,
               '-=' => \&_sub,
               '*' => \&_mul,
               '*=' => \&_mul,
               '**' => \&_pow,
               '**=' => \&_pow,
               '/' => \&_div,
               '/=' => \&_div,
               '%' => \&_rest,
               '%=' => \&_rest,
               'neg' => \&_neg,
               '++' => \&_inc,
               '--' => \&_dec,
               '!' => \&_not,
               '~' => \&_bnot,
               '&' => \&_and,
               '|' => \&_or,
               '^' => \&_xor,
               '<<' => \&_left,
               '>>' => \&_right,
               '<=>' => \&_spaceship,
               '>' => \&_gtn,
               '<' => \&_ltn,
               '>=' => \&_gen,
               '<=' => \&_len,
               '==' => \&_eqn,
               '!=' => \&_nen,
               'bool' => \&_bool,
               '0+' => \&_number,
               '""' => \&_string,
               '=' => \&_clone,
               fallback => 1 );

1;

# ABSTRACT: Manipulate 64 bits integers in Perl

__END__

=encoding UTF-8

=head1 NAME

Math::Int64 - Manipulate 64 bits integers in Perl

=head1 SYNOPSIS

  use Math::Int64 qw(int64 uint64);

  my $i = int64(1);
  my $j = $i << 40;
  print($i + $j * 1000000);

  my $k = uint64("12345678901234567890");

=head1 DESCRIPTION

This module adds support for 64 bit integers, signed and unsigned, to
Perl.

=head2 Exportable functions

=over 4

=item int64()

=item int64($value)

Creates a new int64 value and initializes it to C<$value>, where
$value can be a Perl number or a string containing a number.

For instance:

  $i = int64(34);
  $j = int64("-123454321234543212345");

  $k = int64(1234567698478483938988988); # wrong!!!
                                         #  the unquoted number would
                                         #  be converted first to a
                                         #  real number causing it to
                                         #  loose some precision.

Once the int64 number is created it can be manipulated as any other
Perl value supporting all the standard operations (addition, negation,
multiplication, postincrement, etc.).


=item net_to_int64($str)

Converts an 8 bytes string containing an int64 in network order to the
internal representation used by this module.

=item int64_to_net($int64)

Returns an 8 bytes string with the representation of the int64 value
in network order.

=item native_to_int64($str)

=item int64_to_native($int64)

similar to net_to_int64 and int64_to_net, but using the native CPU
order.

=item int64_to_number($int64)

returns the optimum representation of the int64 value using Perl
internal types (IV, UV or NV). Precision may be lost.

For instance:

  for my $l (10, 20, 30, 40, 50, 60) {
    my $i = int64(1) << $l;
    my $n = int64_to_number($i);
    print "int64:$i => perl:$n\n";
  }

=item string_to_int64($str, $base)

Converts the string to a int64 value. The conversion is done according
to the given base, which must be a number between 2 and 36 inclusive
or the special value 0. C<$base> defaults to 0.

The string may begin with an arbitrary amount of white space followed
by a single optional C<+> or C<-> sign. If base is zero or 16, the
string may then include a "0x" prefix, and the number will be read in
base 16; otherwise, a zero base is taken as 10 (decimal) unless the
next character is '0', in which case it is taken as 8 (octal).

Underscore characters (C<_>) between the digits are ignored.

No overflow checks are performed by this function unless the
C<die_on_overflow> pragma is used (see L</Die on overflow> below).

See also L<strtoll(3)>.

=item hex_to_int64($i64)

Shortcut for string_to_int64($str, 16)

=item int64_to_string($i64, $base)

Converts the int64 value to its string representation in the given
base (defaults to 10).

=item int64_to_hex($i64)

Shortcut for C<int64_to_string($i64, 16)>.

=item int64_to_BER($i64)

Converts the int64 value to its BER representation (see
L<perlfunc/pack> for a description of the BER format).

In the case of signed numbers, they are transformed into unsigned
numbers before encoding them in the BER format with the following
rule:

  $neg = ($i64 < 0 ? 1 : 0);
  $u64 = (($neg ? ~$i64 : $i64) << 1) | $neg;

That way, positive and negative integers are interleaved as 0, -1, 1,
2, -2, .... The format is similar to that used by Google protocol
buffers to encode signed variants but with the most significant groups
first (protocol buffers uses the least significant groups first
variant).

If you don't want that preprocessing for signed numbers, just use the
C<uint64_to_BER> function instead.

=item BER_to_int64($str)

Decodes the int64 number in BER format from the given string.

There must not be any extra bytes on the string after the encoded number.

=item BER_length($str)

Given a string with a BER encoded number at the beginning, this
function returns the number of bytes it uses.

The right way to shift a BER encoded number from the beginning of some
string is as follows:

   $i64 = BER_to_int64(substr($str, 0, BER_length($str), ''));

=item int64_rand

Generates a 64 bit random number using ISAAC-64 algorithm.

=item int64_srand($seed)

=item int64_srand()

Sets the seed for the random number generator.

C<$seed>, if given, should be a 2KB long string.

=item uint64

=item uint64_to_number

=item net_to_uint64

=item uint64_to_net

=item native_to_uint64

=item uint64_to_native

=item string_to_uint64

=item hex_to_uint64

=item uint64_to_string

=item uint64_to_hex

These functions are similar to their int64 counterparts, but
manipulate 64 bit unsigned integers.

=item uint64_to_BER($u64)

Encodes the given unsigned integer in BER format (see L<perlfunc/pack>).

=item BER_to_uint64($str)

Decodes from the given string an unsigned number in BER format.

=item uint64_rand

Generates a 64 bit random unsigned number using ISAAC-64 algorithm.

=back

=head2 Die on overflow

The lexical pragma C<Math::Int64::die_on_overflow> configures the
module to throw an error when some operation results in integer
overflow.

For instance:

  use Math::Int64 qw(uint64);
  use Math::Int64::die_on_overflow;

  my $zero = uint64(0);
  say ($zero - 1);                 # dies as -1 falls outside
                                   # the uint64_t range

  no Math::Int64::die_on_overflow; # deactivates lexical pragma
  say ($zero - 1);                 # no error is detected here!

The pragma can also be activated as follows:

  use Math::Int64 ':die_on_overflow';

Once this pragma is used, several Math::Int64 operations may become
slower. Deactivating the pragma will not make them fast again.

On Perl 5.8.x, as lexical pragmas support is not available, the pragma
C<die_on_overflow> pragma is global and can not be deactivated.

=head2 Fallback to native 64bit support if available

If the lexical pragma C<Math::Int64::native_if_available> is used in
your program and the version of perl in use has native support for
64bit integers, the functions imported from the module that create
64bit integers (i.e. C<uint64>, C<int64>, C<string_to_int64>,
C<native_to_int64>, etc.) will return regular perl scalars.

For instance:

  use Math::Int64 qw(int64);

  $a = int64(34); # always returns an object of the class Math::Int64

  use Math::Int64::native_if_available;
  $a = int64(34); # returns a regular scalar on perls compiled with
                  # 64bit support

This feature is not enabled by default because the semantics for perl
scalars and for 64 bit integers as implemented in this module are not
identical.

Perl is prone to coerce integers into floats while this module keeps
then always as 64bit integers. Specifically, the division operation
and overflows are the most problematic cases. Also, when using native
integers, the signed/unsigned division blurs.

Besides that, in most situations it is safe to use the native fallback.

As happens with the C<die_on_overflow> pragma, on Perl 5.8.x it is
global.

The pragma can also be activated as follows:

  use Math::Int64 ':native_if_available';

=head2 Transparent conversion of objects to int64/uint64

When in some operation involving int64/uint64 numbers, a blessed
object is passed as an operand, the module would try to coerce the
object into an int64/uint64 number calling the methods
C<as_int64>/C<as_uint64> respectively.

If the corresponding method is not implemented, the object will be
stringified and then parsed as a base 10 number.

=head2 Storable integration

Objects of classes Math::Int64 and Math::UInt64 implement the
STORABLE_freeze and STORABLE_thaw methods for a transparent
integration with L<Storable>.

=head2 C API

This module provides a native C API that can be used to create and
read Math::Int64 int64 and uint64 SVs from your own XS modules.

In order to use it you need to follow these steps:

=over 4

=item *

Import the files C<perl_math_int64.c>, C<perl_math_int64.h> and
optionally C<typemaps> from Math::Int64 C<c_api_client> directory into
your project directory.

=item *

Include the file C<perl_math_int64.h> in the C or XS source files
where you want to convert 64bit integers to/from Perl SVs.

Note that this header file requires the types int64_t and uint64_t to
be defined beforehand.

=item *

Add the file C<perl_math_int64.c> to your compilation targets (see the
sample Makefile.PL below).

=item *

Add a call to the macro C<PERL_MATH_INT64_LOAD_OR_CROAK> into the
C<BOOT> section of your XS file.

=back

For instance:

 --- Foo64.xs ---------

  #include "EXTERN.h"
  #include "perl.h"
  #include "XSUB.h"
  #include "ppport.h"

  /* #define MATH_INT64_NATIVE_IF_AVAILABLE */
  #include "math_int64.h"

  MODULE = Foo64		PACKAGE = Foo64
  BOOT:
      PERL_MATH_INT64_LOAD_OR_CROAK;

  int64_t
  some_int64()
  CODE:
      RETVAL = -42;
  OUTPUT:
      RETVAL


  --- Makefile.PL -----

  use ExtUtils::MakeMaker;
  WriteMakefile( NAME         => 'Foo64',
                 VERSION_FROM => 'lib/Foo64.pm',
                 OBJECT       => '$(O_FILES)' );


If the macro C<MATH_INT64_NATIVE_IF_AVAILABLE> is defined before
including C<perl_math_int64.h> and the perl interpreter is compiled
with native 64bit integer support, IVs will be used to represent 64bit
integers instead of the object representation provided by Math::Int64.

These are the C macros available from Math::Int64 C API:

=over 4

=item SV *newSVi64(int64_t i64)

Returns an SV representing the given int64_t value.

=item SV *newSVu64(uint64_t 64)

Returns an SV representing the given uint64_t value.

=item int64_t SvI64(SV *sv)

Extracts the int64_t value from the given SV.

=item uint64_t SvU64(SV *sv)

Extracts the uint64_t value from the given SV.

=item int SvI64OK(SV *sv)

Returns true is the given SV contains a valid int64_t value.

=item int SvU64OK(SV *sv)

Returns true is the given SV contains a valid uint64_t value.

=item uint64_t randU64(void)

Returns a random 64 bits unsigned integer.

=item SV sv_seti64(SV *sv, uint64_t i64)

Sets the value of the perl scalar to the given int64_t value.

=item SV sv_setu64(SV *sv, uint64_t i64)

Sets the value of the perl scalar to the given uint64_t value.

=back

If you require any other function available through the C API don't
hesitate to ask for it!

=head1 BUGS AND SUPPORT

The Storable integration feature is experimental.

The C API feature is experimental.

This module requires int64 support from the C compiler.

In order to report bugs you can send me and email to the address that
appears below or use the CPAN RT bug tracking system available at
L<http://rt.cpan.org>.

The source for the development version of the module is hosted at
GitHub: L<https://github.com/salva/p5-Math-Int64>.

=head2 My wishlist

If you like this module and you're feeling generous, take a look at my
Amazon Wish List: L<http://amzn.com/w/1WU1P6IR5QZ42>

=head1 SEE ALSO

The C API usage sample module L<Math::Int64::C_API::Sample>.

Other modules providing support for larger integers or numbers are
L<Math::BigInt>, L<Math::BigRat> and L<Math::Big>,
L<Math::BigInt::BitVect>, L<Math::BigInt::Pari> and
L<Math::BigInt::GMP>.

=head1 COPYRIGHT AND LICENSE

Copyright E<copy> 2007, 2009, 2011-2015 by Salvador Fandiño
(sfandino@yahoo.com)

Copyright E<copy> 2014-2015 by Dave Rolsky (autarch@urth.org)

This library is free software; you can redistribute it and/or modify
it under the same terms as Perl itself, either Perl version 5.8.8 or,
at your option, any later version of Perl 5 you may have available.

=cut