File Coverage

blib/lib/Test/Mock/Redis.pm

Criterion	Covered	Total	%
statement	129	129	100.0
branch	30	30	100.0
condition	11	15	73.3
subroutine	45	45	100.0
pod	1	4	25.0
total	216	223	96.8

package Test::Mock::Redis;

use warnings;
use strict;

use Carp;
use Config;
use Scalar::Util qw/blessed/;
use Class::Method::Modifiers;
use Package::Stash;
use Try::Tiny;
use namespace::clean;   # important: clean all subs imported above this line

=head1 NAME

Test::Mock::Redis - use in place of Redis for unit testing

=head1 VERSION

Version 0.21

=cut

our $VERSION = '0.21';

=head1 SYNOPSIS

Test::Mock::Redis can be used in place of Redis for running
tests without needing a running redis instance.

    use Test::Mock::Redis;

    my $redis = Test::Mock::Redis->new(server => 'whatever');

    $redis->set($key, 'some value');

    $redis->get($key);

    ...

This module is designed to function as a drop in replacement for
Redis.pm for testing purposes.

See perldoc Redis and the redis documentation at L

=head1 PERSISTENCE

The "connection" to the mocked server (and its stored data) will persist beyond
the object instance, just like a real Redis server. This means that you do not
need to save the instance to this object in order to preserve your data; simply
call C with the same server parameter and the same instance will be
returned, with all data preserved.

=head1 SUBROUTINES/METHODS

=head2 new

    Create a new Test::Mock::Redis object.

    It can be used in place of a Redis object for unit testing.

    It accepts the "server" argument, just like Redis.pm's new.

=head2 num_databases

Redis ships with a default of 16 databases, and that's what this module
handles by default. If you need to change that, do

    use Test::Mock::Redis num_databases => 21;

or at run-time

    Test::Mock::Redis::change_num_databases(21);

=cut

my $NUM_DBS = 16;

sub import {
    my ($class, %args) = @_;

    if ($args{num_databases}){
        change_num_databases($args{num_databases});
    }
}

sub change_num_databases {
    $NUM_DBS = shift;
}


sub _new_db {
    tie my %hash, 'Test::Mock::Redis::PossiblyVolatile';
    return \%hash;
}


sub _defaults {
    my @hex = (0..9, 'a'..'f');
    return (
        _quit      => 0,
        _shutdown  => 0,
        _stash     => [ map { _new_db } (1..$NUM_DBS) ],
        _num_dbs   => $NUM_DBS,
        _db_index  => 0,
        _up_since  => time,
        _last_save => time,
        _run_id    => (join '', map { $hex[rand @hex] } 1..40), # E.G. '0e7e19fc45139fdb26ff3dd35ca6725d9882f1b7',
    );
}


my $instances;

sub new {
    my $class = shift;
    my %args = @_;

    my $server = defined $args{server}
               ? $args{'server'}
               : 'localhost:6379';

    if( $instances->{$server} ){
        confess "Could not connect to Redis server at $server" if $instances->{$server}->{_shutdown};
        $instances->{$server}->{_quit} = 0;
        return $instances->{$server};
    }

    my $self = bless {$class->_defaults, server => $server}, $class;

    $instances->{$server} = $self;

    return $self;
}

sub ping {
    my $self = shift;

    return !$self->{_shutdown}
        && !$self->{_quit};
}

sub auth {
    my $self = shift;

    confess '[auth] ERR wrong number of arguments for \'auth\' command' unless @_;

    return 'OK';
}

sub quit {
    my $self = shift;

    my $return = !$self->{_quit};

    $self->{_quit} = 1;
    return $return;
}

sub shutdown {
    my $self = shift;

    $self->{_shutdown} = 1;
}

sub set {
    my ( $self, $key, $value, @args ) = @_;
    my $expires = 0;
    while (my $option = shift @args) {
        $option = lc $option;
        # Only set if key exists
        if ($option eq 'xx') {
            return unless $self->exists($key);

        # Only set if key doesn't exist
        } elsif ($option eq 'nx') {
            return if $self->exists($key);

        # Set expire time (in seconds)
        } elsif ($option eq 'ex') {
            my $new_expires = shift @args;
            if ($new_expires > $expires) {
                $expires = $new_expires;
            }

        # Set expire time (in milliseconds)
        } elsif ($option eq 'px') {
            my $new_expires = shift @args;
            $new_expires /= 1000; # To seconds
            if ($new_expires > $expires) {
                $expires = $new_expires;
            }
        } else {
            confess '[error] ERR syntax error';
        }
    }
    $self->_stash->{$key} = "$value";
    if ($expires) {
        $self->expire($key, $expires);
    }

    return 'OK';
}

sub setnx {
    my ( $self, $key, $value ) = @_;

    return 0 if $self->exists($key);

    $self->_stash->{$key} = "$value";

    return 1;
}

sub setex {
    my ( $self, $key, $ttl, $value ) = @_;
    $self->set($key, $value);
    $self->expire($key, $ttl);
    return 'OK';
}

sub expire {
    my ( $self, $key, $ttl ) = @_;

    return $self->expireat($key, time + $ttl);
}

sub expireat {
    my ( $self, $key, $when ) = @_;

    return 0 unless exists $self->_stash->{$key};

    my $slot = $self->_stash;
    my $tied = tied(%$slot);

    $tied->expire($key, $when);

    return 1;
}

sub persist {
    my ( $self, $key, $ttl ) = @_;

    return 0 unless exists $self->_stash->{$key};

    my $slot = $self->_stash;
    my $tied = tied(%$slot);

    $tied->persist($key);

    return 1;
}

sub ttl {
    my ( $self, $key, $ttl ) = @_;

    return -1 unless exists $self->_stash->{$key};

    my $slot = $self->_stash;
    my $tied = tied(%$slot);

    return $tied->ttl($key);
}

sub exists :method {
    my ( $self, $key ) = @_;
    return exists $self->_stash->{$key} ? 1 : 0;
}

sub get {
    my ( $self, $key  ) = @_;

    return $self->_stash->{$key};
}

sub append {
    my ( $self, $key, $value ) = @_;

    $self->_stash->{$key} .= $value;

    return $self->strlen($key);
}

sub strlen {
    my ( $self, $key ) = @_;
    return do { use bytes; length $self->_stash->{$key}; };
}

sub getset {
    my ( $self, $key, $value ) = @_;

    #TODO: should return error when original value isn't a string
    my $old_value = $self->_stash->{$key};

    $self->set($key, $value);

    return $old_value;
}

sub incr {
    my ( $self, $key  ) = @_;

    $self->_stash->{$key} ||= 0;

    return ++$self->_stash->{$key};
}

sub incrby {
    my ( $self, $key, $incr ) = @_;

    $self->_stash->{$key} ||= 0;

    return $self->_stash->{$key} += $incr;
}

sub decr {
    my ( $self, $key ) = @_;

    return --$self->_stash->{$key};
}

sub decrby {
    my ( $self, $key, $decr ) = @_;

    $self->_stash->{$key} ||= 0;

    return $self->_stash->{$key} -= $decr;
}

sub mget {
    my ( $self, @keys ) = @_;

    return map { $self->_stash->{$_} } @keys;
}

sub mset {
    my ( $self, %things ) = @_;

    @{ $self->_stash }{keys %things} = (values %things);

    return 'OK';
}

sub msetnx {
    my ( $self, %things ) = @_;

    $self->exists($_) && return 0 for keys %things;

    @{ $self->_stash }{keys %things} = (values %things);

    return 1;
}

sub del {
    my ( $self, @keys ) = @_;

    my $ret = 0;

    for my $key (@keys) {
        $ret++ if $self->exists($key);
        delete $self->_stash->{$key};
    }

    return $ret;
}

sub type {
    my ( $self, $key ) = @_;
    # types are string, list, set, zset and hash

    return 'none' unless $self->exists($key);

    my $type = ref $self->_stash->{$key};

    return !$type
         ? 'string'
         : $type eq 'Test::Mock::Redis::Hash'
           ? 'hash'
           : $type eq 'Test::Mock::Redis::Set'
             ? 'set'
             : $type eq 'Test::Mock::Redis::ZSet'
               ? 'zset'
                 : $type eq 'Test::Mock::Redis::List'
                   ? 'list'
                   : 'unknown'
    ;
}

sub keys :method {
    my ( $self, $match ) = @_;

    confess q{[KEYS] ERR wrong number of arguments for 'keys' command} unless defined $match;

    # TODO: we're not escaping other meta-characters
    $match =~ s/(?
    $match =~ s/(?

    return @{[ sort { $a cmp $b }
        grep { exists $self->_stash->{$_} }
        grep { /^$match/ }
        keys %{ $self->_stash }]};
}

sub randomkey {
    my $self = shift;

    return ( keys %{ $self->_stash } )[
                int(rand( scalar keys %{ $self->_stash } ))
            ]
    ;
}

sub rename {
    my ( $self, $from, $to, $whine ) = @_;

    confess '[rename] ERR source and destination objects are the same' if $from eq $to;
    confess '[rename] ERR no such key' unless $self->exists($from);
    confess 'rename to existing key' if $whine && $self->_stash->{$to};

    $self->_stash->{$to} = $self->_stash->{$from};
    delete $self->_stash->{$from};
    return 'OK';
}

sub renamenx {
    my ( $self, $from, $to ) = @_;

    return 0 if $self->exists($to);
    return $self->rename($from, $to);
}

sub dbsize {
    my $self = shift;

    return scalar keys %{ $self->_stash };
}

sub rpush {
    my ( $self, $key, $value ) = @_;

    $self->_make_list($key);

    push @{ $self->_stash->{$key} }, "$value";
    return scalar @{ $self->_stash->{$key} };
}

sub lpush {
    my ( $self, $key, $value ) = @_;

    confess "[lpush] ERR Operation against a key holding the wrong kind of value"
        unless !$self->exists($key) or $self->_is_list($key);

    $self->_make_list($key);

    unshift @{ $self->_stash->{$key} }, "$value";
    return scalar @{ $self->_stash->{$key} };
}

sub rpushx {
    my ( $self, $key, $value ) = @_;

    return unless $self->_is_list($key);

    push @{ $self->_stash->{$key} }, "$value";
    return scalar @{ $self->_stash->{$key} };
}

sub lpushx {
    my ( $self, $key, $value ) = @_;

    return unless $self->_is_list($key);

    unshift @{ $self->_stash->{$key} }, "$value";
    return scalar @{ $self->_stash->{$key} };
}

sub rpoplpush {
    my ( $self, $source_key, $destination_key ) = @_;

    my $popped_element = $self->rpop( $source_key ) or return;

    $self->lpush( $destination_key, $popped_element );

    return $popped_element;
}

sub llen {
    my ( $self, $key ) = @_;

    return 0 unless $self->exists($key);

    return scalar @{ $self->_stash->{$key} };
}

sub lrange {
    my ( $self, $key, $start, $end ) = @_;

    my $array = $self->_stash->{$key};
    ($start,$end) = _normalize_range(scalar(@$array),$start,$end);
    return @{ $array }[$start..$end];
}

sub ltrim {
    my ( $self, $key, $start, $end ) = @_;

    my $array = $self->_stash->{$key};
    ($start,$end) = _normalize_range(scalar(@$array),$start,$end);
    $self->_stash->{$key} = [ @{ $array }[$start..$end] ];
    return 'OK';
}

sub lindex {
    my ( $self, $key, $index ) = @_;

    my $array = $self->_stash->{$key};
    $index = _normalize_index(scalar(@$array),$index);
    return $array->[$index];
}

sub lset {
    my ( $self, $key, $index, $value ) = @_;

    my $array = $self->_stash->{$key};
    $index = _normalize_index(scalar(@$array),$index);
    $array->[$index] = "$value";
    return 'OK';
}

sub lrem {
    my ( $self, $key, $count, $value ) = @_;
    my $removed;
    my @indicies = (0..$#{ $self->_stash->{$key} });
    @indicies = reverse @indicies if $count < 0;
    $count = abs $count;

    my @to_remove;
    for my $index (@indicies){
        if($self->_stash->{$key}->[$index] eq $value){
            push @to_remove, $index;
            $removed++;
            last if $count && $removed >= $count;
        }
    }

    # reverse sort so that the higher indecies are removed first
    for my $rm_idx (sort { $b <=> $a } @to_remove){
        splice @{ $self->_stash->{$key} }, $rm_idx, 1;
    }

    return $removed;
}

sub lpop {
    my ( $self, $key ) = @_;

    return undef unless $self->exists($key);

    return shift @{ $self->_stash->{$key} };
}

sub rpop {
    my ( $self, $key ) = @_;

    return undef unless $self->exists($key);

    return pop @{ $self->_stash->{$key} };
}

sub select {
    my ( $self, $index ) = @_;

    my $max_index = $#{ $self->{_stash} };
    if ($index > $max_index ){
        die "You called select($index), but max allowed is $max_index unless you configure more databases";
    }

    $self->{_db_index} = $index;
    return 'OK';
}

sub _stash {
    my ( $self, $index ) = @_;
    $index = $self->{_db_index} unless defined $index;

    return $self->{_stash}->[$index];
}

sub sadd {
    my ( $self, $key, $value ) = @_;

    $self->_make_set($key);
    my $return = exists $self->_stash->{$key}->{$value}
               ? 0
               : 1;
    $self->_stash->{$key}->{$value} = 1;
    return $return;
}

sub scard {
    my ( $self, $key ) = @_;

    return $self->_is_set($key)
         ? scalar $self->smembers($key)
         : 0;
}

sub sismember {
    my ( $self, $key, $value ) = @_;

    return exists $self->_stash->{$key}->{$value}
            ? 1
            : 0;
}

sub srem {
    my ( $self, $key, $value ) = @_;

    return 0 unless exists $self->_stash->{$key}
                 && exists $self->_stash->{$key}->{$value};

    delete $self->_stash->{$key}->{$value};
    return 1;
}

sub spop {
    my ( $self, $key ) = @_;

    return undef unless $self->_is_set($key);

    my $value = $self->srandmember($key);
    delete $self->_stash->{$key}->{$value};
    return $value;
}

sub smove {
    my ( $self, $source, $dest, $value ) = @_;

    confess "[smove] ERR Operation against a key holding the wrong kind of value"
        if ( $self->exists($source) and not $self->_is_set($source) )
        or ( $self->exists($dest)   and not $self->_is_set($dest)   );

    if( (delete $self->_stash->{$source}->{$value}) ){
        $self->_make_set($dest) unless $self->_is_set($dest);
        $self->_stash->{$dest}->{$value} = 1;
        return 1;
    }
    return 0;  # guess it wasn't in there
}

sub srandmember {
    my ( $self, $key ) = @_;

    return undef unless $self->_is_set($key);

    return ($self->smembers($key))[rand int $self->scard($key)];
}

sub smembers {
    my ( $self, $key ) = @_;

    return keys %{ $self->_stash->{$key} };
}

sub sinter {
    my ( $self, @keys ) = @_;

    my $r = {};

    foreach my $key (@keys){
        $r->{$_}++ for keys %{ $self->_stash->{$key} };
    }

    return grep { $r->{$_} >= @keys } keys %$r;
}

sub sinterstore {
    my ( $self, $dest, @keys ) = @_;

    $self->_stash->{$dest} = { map { $_ => 1 } $self->sinter(@keys) };
    bless $self->_stash->{$dest}, 'Test::Mock::Redis::Set';
    return $self->scard($dest);
}

sub sunion {
    my ( $self, @keys ) = @_;

    my $r = {};

    foreach my $key (@keys){
        $r->{$_}++ for keys %{ $self->_stash->{$key} };
    }

    return grep { $r->{$_} >= 1 } keys %$r;
}

sub sunionstore {
    my ( $self, $dest, @keys ) = @_;

    $self->_stash->{$dest} = { map { $_ => 1 } $self->sunion(@keys) };
    bless $self->_stash->{$dest}, 'Test::Mock::Redis::Set';
    return $self->scard($dest);
}

sub sdiff {
    my ( $self, $start, @keys ) = @_;

    my $r = { map { $_ => 0 } keys %{ $self->_stash->{$start} } };

    foreach my $key (@keys){
        $r->{$_}++ for keys %{ $self->_stash->{$key} };
    }

    return grep { $r->{$_} == 0 } keys %$r;
}

sub sdiffstore {
    my ( $self, $dest, $start, @keys ) = @_;

    $self->_stash->{$dest} = { map { $_ => 1 } $self->sdiff($start, @keys) };
    bless $self->_stash->{$dest}, 'Test::Mock::Redis::Set';
    return $self->scard($dest);
}

sub hset {
    my ( $self, $key, $hkey, $value ) = @_;

    confess '[hset] ERR Operation against a key holding the wrong kind of value'
         if $self->exists($key) and !$self->_is_hash($key);


    $self->_make_hash($key);

    my $ret = exists $self->_stash->{$key}->{$hkey}
            ? 0
            : 1;
    $self->_stash->{$key}->{$hkey} = $value;
    return $ret;
}

sub hsetnx {
    my ( $self, $key, $hkey, $value ) = @_;

    return 0 if exists $self->_stash->{$key}->{$hkey};

    $self->_make_hash($key);

    $self->_stash->{$key}->{$hkey} = "$value";
    return 1;
}

sub hmset {
    my ( $self, $key, %hash ) = @_;

    $self->_make_hash($key);

    foreach my $hkey ( keys %hash ){
        $self->hset($key, $hkey, $hash{$hkey});
    }

    return 'OK';
}

sub hget {
    my ( $self, $key, $hkey ) = @_;

    return undef unless $self->_is_hash($key);

    return $self->_stash->{$key}->{$hkey};
}

sub hmget {
    my ( $self, $key, @hkeys ) = @_;

    return undef unless $self->_is_hash($key);

    return map { $self->_stash->{$key}->{$_} } @hkeys;
}

sub hexists {
    my ( $self, $key, $hkey ) = @_;

    confess '[hexists] ERR Operation against a key holding the wrong kind of value'
         if $self->exists($key) and !$self->_is_hash($key);

    return exists $self->_stash->{$key}->{$hkey} ? 1 : 0;
}

sub hdel {
    my ( $self, $key, $hkey ) = @_;

    return 0 unless $self->_is_hash($key);

    my $ret = $self->hexists($key, $hkey);
    delete $self->_stash->{$key}->{$hkey};
    return $ret;
}

sub hincrby {
    confess "[hincrby] ERR wrong number of arguments for 'hincrby' command"
        unless @_ == 4;

    my ( $self, $key, $hkey, $incr ) = @_;

    confess '[hexists] ERR Operation against a key holding the wrong kind of value'
         if $self->exists($key) and !$self->_is_hash($key);

    confess "[hincrby] ERR hash value is not an integer"
         if $self->hexists($key, $hkey)                   # it exists
             and $self->hget($key, $hkey) !~ /^-?\d+$|^$/ # and it doesn't look like an integer (and it isn't empty)
    ;

    $self->_make_hash($key) unless $self->_is_hash($key);

    $self->_stash->{$key}->{$hkey} ||= 0;

    return $self->_stash->{$key}->{$hkey} += $incr;
}

sub hlen {
    my ( $self, $key ) = @_;

    return 0 unless $self->_is_hash($key);

    return scalar values %{ $self->_stash->{$key} };
}

sub hkeys {
    my ( $self, $key ) = @_;

    confess '[hkeys] ERR Operation against a key holding the wrong kind of value'
         if $self->exists($key) and !$self->_is_hash($key);

    return () unless $self->exists($key);

    return keys %{ $self->_stash->{$key} };
}

sub hvals {
    my ( $self, $key ) = @_;

    confess '[hvals] ERR Operation against a key holding the wrong kind of value'
         if $self->exists($key) and !$self->_is_hash($key);

    return values %{ $self->_stash->{$key} };
}

sub hgetall {
    my ( $self, $key ) = @_;

    confess "[hgetall] ERR Operation against a key holding the wrong kind of value"
         if $self->exists($key) and !$self->_is_hash($key);

    return $self->exists( $key )
         ? %{ $self->_stash->{$key} }
         : ();
}

sub move {
    my ( $self, $key, $to ) = @_;

    return 0 unless !exists $self->_stash($to)->{$key}
                 &&  exists $self->_stash->{$key}
    ;

    $self->_stash($to)->{$key} = $self->_stash->{$key};
    delete $self->_stash->{$key};
    return 1;
}

sub flushdb {
    my $self = shift;

    $self->{_stash}->[$self->{_db_index}] = _new_db;
}

sub flushall {
    my $self = shift;

    $self->{_stash} = [ map { _new_db }(1..$NUM_DBS) ];
}

sub sort {
    my ( $self, $key, $how ) = @_;

    my $cmp = do
    { no warnings 'uninitialized';
      $how =~ /\bALPHA\b/
      ? $how =~ /\bDESC\b/
        ? sub { $b cmp $a }
        : sub { $a cmp $b }
      : $how =~ /\bDESC\b/
        ? sub { $b <=> $a }
        : sub { $a <=> $b }
      ;
    };

    return sort $cmp @{ $self->_stash->{$key} };
}

sub save {
    my $self = shift;
    $self->{_last_save} = time;
    return 'OK';
}

sub bgsave {
    my $self = shift;
    return $self->save;
}

sub lastsave {
    my $self = shift;
    return $self->{_last_save};
}

sub info {
    my $self = shift;

    return {
        aof_current_rewrite_time_sec => '-1',
        aof_enabled => '0',
        aof_last_bgrewrite_status => 'ok',
        aof_last_rewrite_time_sec => '-1',
        aof_rewrite_in_progress => '0',
        aof_rewrite_scheduled => '0',
        arch_bits => $Config{use64bitint } ? '64' : '32',
        blocked_clients => '0',
        client_biggest_input_buf => '0',
        client_longest_output_list => '0',
        connected_clients => '1',
        connected_slaves => '0',
        evicted_keys => '0',
        expired_keys => '0',
        gcc_version => '4.2.1',
        instantaneous_ops_per_sec => '568',
        keyspace_hits => '272',
        keyspace_misses => '0',
        latest_fork_usec => '0',
        loading => '0',
        lru_clock => '1994309',
        mem_allocator => 'libc',
        mem_fragmentation_ratio => '1.61',
        multiplexing_api => 'kqueue',
        os => $Config{osname}.' '.$Config{osvers},  # should be like 'Darwin 12.2.1 x86_64', this is close
        process_id => $$,
        pubsub_channels => '0',
        pubsub_patterns => '0',
        rdb_bgsave_in_progress => '0',
        rdb_changes_since_last_save => '0',
        rdb_current_bgsave_time_sec => '-1',
        rdb_last_bgsave_status => 'ok',
        rdb_last_bgsave_time_sec => '-1',
        rdb_last_save_time => '1362120372',
        redis_git_dirty => '0',
        redis_git_sha1 => '34b420db',
        redis_mode => 'standalone',
        redis_version => '2.6.10',
        rejected_connections => '0',
        role => 'master',
        run_id => $self->{_run_id},
        tcp_port => '11084',
        total_commands_processed => '1401',
        total_connections_received => '1',
        uptime_in_days             => (time - $self->{_up_since}) / 60 / 60 / 24,
        uptime_in_seconds          => time - $self->{_up_since},
        used_cpu_sys => '0.04',
        used_cpu_sys_children => '0.00',
        used_cpu_user => '0.02',
        used_cpu_user_children => '0.00',
        used_memory => '1056288',
        used_memory_human => '1.01M',
        used_memory_lua => '31744',
        used_memory_peak => '1055728',
        used_memory_peak_human => '1.01M',
        used_memory_rss => '1699840',
        map { 'db'.$_ => sprintf('keys=%d,expires=%d',
                             scalar keys %{ $self->_stash($_) },
                             $self->_expires_count_for_db($_),
                         )
            } grep { scalar keys %{ $self->_stash($_) } > 0 }
                (0..15)
    };
}

sub _expires_count_for_db {
    my ( $self, $db_index ) = @_;

    my $slot = $self->_stash($db_index);
    my $tied = tied(%$slot);

    $tied->expire_count;
}

sub zadd {
    my ( $self, $key, $score, $value ) = @_;

    $self->_make_zset($key);

    my $ret = exists $self->_stash->{$key}->{$value}
            ? 0
            : 1;
    $self->_stash->{$key}->{$value} = $score;
    return $ret;
}


sub zscore {
    my ( $self, $key, $value ) = @_;
    return $self->_stash->{$key}->{$value};
}

sub zincrby {
    my ( $self, $key, $score, $value ) = @_;

    $self->_stash->{$key}->{$value} ||= 0;

    return $self->_stash->{$key}->{$value} += $score;
}

sub zrank {
    my ( $self, $key, $value ) = @_;
    my $rank = 0;
    foreach my $elem ( $self->zrange($key, 0, $self->zcard($key)) ){
        return $rank if $value eq $elem;
        $rank++;
    }
    return undef;
}

sub zrevrank {
    my ( $self, $key, $value ) = @_;
    my $rank = 0;
    foreach my $elem ( $self->zrevrange($key, 0, $self->zcard($key)) ){
        return $rank if $value eq $elem;
        $rank++;
    }
    return undef;
}

sub zrange {
    my ( $self, $key, $start, $stop, $withscores ) = @_;

    my $length = $self->zcard($key);
    ($start,$stop) = _normalize_range($length,$start,$stop);

    return map { $withscores ? ( $_, $self->zscore($key, $_) ) : $_ }
               ( map { $_->[0] }
                     sort { $a->[1] <=> $b->[1] || $a->[0] cmp $b->[0] }
                         map { [ $_, $self->_stash->{$key}->{$_} ] }
                             keys %{ $self->_stash->{$key} }
               )[$start..$stop]
    ;
}

sub zrevrange {
    my ( $self, $key, $start, $stop, $withscores ) = @_;

    my $length = $self->zcard($key);
    ($start,$stop) = _normalize_range($length,$start,$stop);

    return map { $withscores ? ( $_, $self->zscore($key, $_) ) : $_ }
               ( map { $_->[0] }
                     sort { $b->[1] <=> $a->[1] || $b->[0] cmp $a->[0] }
                         map { [ $_, $self->_stash->{$key}->{$_} ] }
                             keys %{ $self->_stash->{$key} }
               )[$start..$stop]
    ;
}

sub zrangebyscore {
    my ( $self, $key, $min, $max, $withscores ) = @_;

    my $min_inc = !( $min =~ s/^\(// );
    my $max_inc = !( $max =~ s/^\(// );

    my $cmp = !$min_inc && !$max_inc
            ? sub { $self->zscore($key, $_[0]) > $min && $self->zscore($key, $_[0]) < $max }
            : !$min_inc
              ? sub { $self->zscore($key, $_[0]) > $min && $self->zscore($key, $_[0]) <= $max }
              : !$max_inc
                ? sub { $self->zscore($key, $_[0]) >= $min && $self->zscore($key, $_[0]) <  $max }
                : sub { $self->zscore($key, $_[0]) >= $min && $self->zscore($key, $_[0]) <= $max }
    ;

    return map { $withscores ? ( $_, $self->zscore($key, $_) ) : $_ }
               grep { $cmp->($_) } $self->zrange($key, 0, $self->zcard($key)-1);
}

# note max and min are reversed from zrangebyscore
sub zrevrangebyscore {
    my ( $self, $key, $max, $min, $withscores ) = @_;

    my $not_with_scores = 0;

    return map { $withscores ? ( $_, $self->zscore($key, $_) ) : $_ }
               reverse $self->zrangebyscore($key, $min, $max, $not_with_scores);
}

sub zcount {
    my ( $self, $key, $min, $max ) = @_;
    return scalar $self->zrangebyscore($key, $min, $max);
}

sub zcard {
    my ( $self, $key ) = @_;
    return scalar values %{ $self->_stash->{$key} }
}

sub zremrangebyrank {
    my ( $self, $key, $start, $stop ) = @_;

    my @remove = $self->zrange($key, $start, $stop);
    delete $self->_stash->{$key}->{$_} for @remove;
    return scalar @remove;
}

sub zremrangebyscore {
    my ( $self, $key, $start, $stop ) = @_;

    my @remove = $self->zrangebyscore($key, $start, $stop);
    delete $self->_stash->{$key}->{$_} for @remove;
    return scalar @remove;
}

=head1 PIPELINING

See L -- most methods support the use of a callback sub as
the final argument. For this implementation, the callback sub will be called
immediately (before the result of the original method is returned), and
C does nothing.  Combining pipelining with C/C
is not supported.

=head1 TODO

Lots!

Not all Redis functionality is implemented.  The test files that output "TODO" are still to be done.

The top of all test files [except 01-basic.t] has the list of commands tested or to-be tested in the file.

Those marked with an "x" are pretty well-tested.
Those marked with an "o" need help.
Those that are unmarked have no tests, or are un-implemented.  For example:

x   AUTH          <--- has some tests

o   KEYS          <--- only partially tested and/or implemented

    ZINTERSTORE   <--- not tested (or maybe not implemented)



Beyond that, it would be neat to add methods to inspect how often keys were accessed and get other information that
allows the module user to confirm that their code interacted with redis (or Test::Mock::Redis) as they expected.


=head1 AUTHOR

Jeff Lavallee, C<<  >>

=head1 SEE ALSO

The real Redis.pm client whose interface this module mimics: L


=head1 BUGS

Please report any bugs or feature requests to C, or through
the web interface at L.  I will be notified, and then you'll
automatically be notified of progress on your bug as I make changes.



=head1 SUPPORT

You can find documentation for this module with the perldoc command.

    perldoc Test::Mock::Redis


You can also look for information at:

=over 4

=item * RT: CPAN's request tracker

L

=item * AnnoCPAN: Annotated CPAN documentation

L

=item * CPAN Ratings

L

=item * Search CPAN

L

=back


=head1 ACKNOWLEDGEMENTS

Salvatore Sanfilippo for redis, of course!

Dobrica Pavlinusic & Pedro Melo for Redis.pm

The following people have contributed to I:

=over

=item * Chris Reinhardt

=item * Ian Burrell

=item * Gianni Ceccarelli

=item * Karen Etheridge

=item * Keith Broughton

=item * Kevin Goess

=item * Neil Bowers

=item * Nigel Gregoire

=item * Thomas Bloor

=item * Yaakov Shaul

=back

=head1 LICENSE AND COPYRIGHT

Copyright 2018 Jeff Lavallee.

This program is free software; you can redistribute it and/or modify it
under the terms of either: the GNU General Public License as published
by the Free Software Foundation; or the Artistic License.

See L for more information.


=cut

sub _normalize_index {
    my ( $length, $index ) = @_;

    $index += $length if $index < 0;
    return $index;
}

sub _normalize_range {
    my ( $length, $start, $end ) = @_;

    $start = _normalize_index($length,$start);
    $end = _normalize_index($length,$end);
    $end = $length-1 if $end >= $length;

    return ($start,$end);
}

sub _is_list {
    my ( $self, $key ) = @_;

    return $self->exists($key)
        && blessed $self->_stash->{$key}
        && $self->_stash->{$key}->isa('Test::Mock::Redis::List') ;
}

sub _make_list {
    my ( $self, $key ) = @_;

    $self->_stash->{$key} = Test::Mock::Redis::List->new
        unless $self->_is_list($key);
}

sub _is_hash {
    my ( $self, $key ) = @_;

    return $self->exists($key)
        && blessed $self->_stash->{$key}
        && $self->_stash->{$key}->isa('Test::Mock::Redis::Hash') ;
}

sub _make_hash {
    my ( $self, $key ) = @_;

    $self->_stash->{$key} = Test::Mock::Redis::Hash->new
        unless $self->_is_hash($key);
}

sub _is_set {
    my ( $self, $key ) = @_;

    return $self->exists($key)
        && blessed $self->_stash->{$key}
        && $self->_stash->{$key}->isa('Test::Mock::Redis::Set') ;
}

sub _make_set {
    my ( $self, $key ) = @_;

    $self->_stash->{$key} = Test::Mock::Redis::Set->new
        unless $self->_is_set($key);
}

sub _is_zset {
    my ( $self, $key ) = @_;

    return $self->exists($key)
        && blessed $self->_stash->{$key}
        && $self->_stash->{$key}->isa('Test::Mock::Redis::ZSet') ;
}

sub _make_zset {
    my ( $self, $key ) = @_;

    $self->_stash->{$key} = Test::Mock::Redis::ZSet->new
        unless $self->_is_zset($key);
}


# MULTI/EXEC/DISCARD: http://redis.io/topics/transactions

sub multi {
    my ( $self ) = @_;

    confess '[multi] ERR MULTI calls can not be nested' if defined $self->{_multi_commands};

    # set up the list for storing commands sent between MULTI and EXEC/DISCARD
    $self->{_multi_commands} = [];

    return 'OK';
}

# methods that return a list, rather than a single value
my @want_list = qw(mget keys lrange smembers sinter sunion sdiff hmget hkeys hvals hgetall sort zrange zrevrange zrangebyscore);
my %want_list = map { $_ => 1 } @want_list;

sub exec {
    my ( $self ) = @_;

    # we are going to commit all the changes we saved up;
    # replay them now and return all their output

    confess '[exec] ERR EXEC without MULTI' if not defined $self->{_multi_commands};

    my @commands = @{$self->{_multi_commands}};
    delete $self->{_multi_commands};

    # replay all the queries that were queued up
    # the returned result is a nested array of the results of all the commands
    my @exceptions;
    my @results = map {
        my ($method, @args) = @$_;
        my @result =
            try { $self->$method(@args) }
            catch { push @exceptions, $_; (); };
        $want_list{$method} ? \@result : $result[0];
    } @commands;

    s/^\[\w+\] // for @exceptions;

    confess('[exec] ', join('; ', @exceptions)) if @exceptions;

    return @results;
}

sub discard {
    my ( $self ) = @_;

    confess '[discard] ERR DISCARD without MULTI' if not defined $self->{_multi_commands};

    # discard all the accumulated commands, without executing them
    delete $self->{_multi_commands};

    return 'OK';
}

sub watch {
    my ($self) = shift;

    confess '[watch] ERR wrong number of arguments for \'watch\' command' unless @_;

    return 'OK';
}

sub unwatch {
    my ($self) = shift;

    confess '[error] ERR wrong number of arguments for \'unwatch\' command' if @_;

    return 'OK';
}

# now that we've defined all our subs, we need to wrap them all in logic that
# can check if we are in the middle of a MULTI, and if so, queue up the
# commands for later replaying.

my %no_transaction_wrap_methods = (
    new => 1,
    multi => 1,
    exec => 1,
    discard => 1,
    quit => 1,
    import => 1,
    change_num_databases => 1,
);

my @transaction_wrapped_methods =
    grep { !/^_/}
    grep { not $no_transaction_wrap_methods{$_} }
        Package::Stash->new(__PACKAGE__)->list_all_symbols('CODE');

foreach my $method (@transaction_wrapped_methods)
{
    around $method => sub {
        my $orig = shift;
        my $self = shift;

        # pass command through if we are not handling a MULTI
        return $self->$orig(@_) if not defined $self->{_multi_commands};

        push @{$self->{_multi_commands}}, [ $method, @_ ];
        return 'QUEUED';
    };
}


# PIPELINING SUPPORT

# these method modifications must be done after (over top of) the modification
# for transactions, as we need to check for/extract the $cb first.

my %no_pipeline_wrap_methods = (
    new => 1,
    multi => 1,
    discard => 1,
    quit => 1,
    ping => 1,
    subscribe => 1,
    unsubscribe => 1,
    psubscribe => 1,
    punsubscribe => 1,
    wait_all_responses => 1,
);

my @pipeline_wrapped_methods =
    grep { !/^_/}
    grep { not $no_pipeline_wrap_methods{$_} }
        Package::Stash->new(__PACKAGE__)->list_all_symbols('CODE');

# this is a bit messy, and the wantarray logic may not be quite right.
# Alternatively, we could implement all this by reusing the logic in the real
# Redis.pm -- subclass Redis, override new/multi/exec/discard (and probably
# some other special functions), and have __run_cmd use a dispatch table to
# call all our overridden implementations.

foreach my $method (@pipeline_wrapped_methods)
{
    around $method => sub {
        my $orig = shift;
        my $self = shift;
        my @args = @_;

        my $cb = @args && ref $args[-1] eq 'CODE' ? pop @args : undef;

        return $self->$orig(@args) if not $cb;

        # this may be officially supported eventually -- see
        # https://github.com/melo/perl-redis/issues/17
        # and "Pipeline management" in the Redis docs
        # To make this work, we just need to special-case exec, to collect all the
        # results and errors in tuples and send that to the $cb
        die 'cannot combine pipelining with MULTI' if $self->{_multi_commands};

        # We could also implement this with a queue, not bothering to process
        # the commands until wait_all_responses is called - but then we need to
        # make sure to call wait_all_responses explicitly as soon as a command
        # is issued without a $cb.

        my $error;
        my (@result) = try
        {
            $self->$orig(@args);
        }
        catch
        {
            $error = $_;
            ();
        };

        $cb->(
            # see notes above - this logic may not be quite right
            ( $want_list{$method} ? \@result : $result[0] ),
            $error,
        );
        return 1;
    };
}

# in a real Redis system, this will make all outstanding callbacks get called.
sub wait_all_responses {}


1; # End of Test::Mock::Redis

package Test::Mock::Redis::List;
sub new { return bless [], shift }
1;

package Test::Mock::Redis::Hash;
sub new { return bless {}, shift }
1;

package Test::Mock::Redis::ZSet;
sub new { return bless {}, shift }
1;

package Test::Mock::Redis::Set;
sub new { return bless {}, shift }
1;

package Test::Mock::Redis::PossiblyVolatile;

use strict; use warnings;
use Tie::Hash;
use base qw/Tie::StdHash/;

sub DELETE {
    my ( $self, $key ) = @_;

    delete $self->{$key};
}

my $expires;

sub FETCH {
    my ( $self, $key ) = @_;

    return $self->EXISTS($key)
         ? $self->{$key}
         : undef;
}

sub EXISTS {
    my ( $self, $key ) = @_;

    $self->_delete_if_expired($key);

    return exists $self->{$key};
}

sub _delete_if_expired {
    my ( $self, $key ) = @_;
    if(exists $expires->{$self}->{$key}
       && time >= $expires->{$self}->{$key}){
        delete $self->{$key};
        delete $expires->{$self}->{$key};
    }
}

sub expire {
    my ( $self, $key, $time ) = @_;

    $expires->{$self}->{$key} = $time;
}

sub expire_count {
    my ( $self ) = @_;

    # really, we should probably only count keys that haven't expired
    scalar keys %{ $expires->{$self} };
}

sub persist {
    my ( $self, $key, $time ) = @_;

    delete $expires->{$self}->{$key};
}

sub ttl {
    my ( $self, $key ) = @_;

    return -1 unless exists $expires->{$self}->{$key};
    return $expires->{$self}->{$key} - time;
}


1;