| 1 | package Tie::SubstrHash; |
| 2 | |
| 3 | our $VERSION = '1.00'; |
| 4 | |
| 5 | =head1 NAME |
| 6 | |
| 7 | Tie::SubstrHash - Fixed-table-size, fixed-key-length hashing |
| 8 | |
| 9 | =head1 SYNOPSIS |
| 10 | |
| 11 | require Tie::SubstrHash; |
| 12 | |
| 13 | tie %myhash, 'Tie::SubstrHash', $key_len, $value_len, $table_size; |
| 14 | |
| 15 | =head1 DESCRIPTION |
| 16 | |
| 17 | The B<Tie::SubstrHash> package provides a hash-table-like interface to |
| 18 | an array of determinate size, with constant key size and record size. |
| 19 | |
| 20 | Upon tying a new hash to this package, the developer must specify the |
| 21 | size of the keys that will be used, the size of the value fields that the |
| 22 | keys will index, and the size of the overall table (in terms of key-value |
| 23 | pairs, not size in hard memory). I<These values will not change for the |
| 24 | duration of the tied hash>. The newly-allocated hash table may now have |
| 25 | data stored and retrieved. Efforts to store more than C<$table_size> |
| 26 | elements will result in a fatal error, as will efforts to store a value |
| 27 | not exactly C<$value_len> characters in length, or reference through a |
| 28 | key not exactly C<$key_len> characters in length. While these constraints |
| 29 | may seem excessive, the result is a hash table using much less internal |
| 30 | memory than an equivalent freely-allocated hash table. |
| 31 | |
| 32 | =head1 CAVEATS |
| 33 | |
| 34 | Because the current implementation uses the table and key sizes for the |
| 35 | hashing algorithm, there is no means by which to dynamically change the |
| 36 | value of any of the initialization parameters. |
| 37 | |
| 38 | The hash does not support exists(). |
| 39 | |
| 40 | =cut |
| 41 | |
| 42 | use Carp; |
| 43 | |
| 44 | sub TIEHASH { |
| 45 | my $pack = shift; |
| 46 | my ($klen, $vlen, $tsize) = @_; |
| 47 | my $rlen = 1 + $klen + $vlen; |
| 48 | $tsize = [$tsize, |
| 49 | findgteprime($tsize * 1.1)]; # Allow 10% empty. |
| 50 | local $self = bless ["\0", $klen, $vlen, $tsize, $rlen, 0, -1]; |
| 51 | $$self[0] x= $rlen * $tsize->[1]; |
| 52 | $self; |
| 53 | } |
| 54 | |
| 55 | sub CLEAR { |
| 56 | local($self) = @_; |
| 57 | $$self[0] = "\0" x ($$self[4] * $$self[3]->[1]); |
| 58 | $$self[5] = 0; |
| 59 | $$self[6] = -1; |
| 60 | } |
| 61 | |
| 62 | sub FETCH { |
| 63 | local($self,$key) = @_; |
| 64 | local($klen, $vlen, $tsize, $rlen) = @$self[1..4]; |
| 65 | &hashkey; |
| 66 | for (;;) { |
| 67 | $offset = $hash * $rlen; |
| 68 | $record = substr($$self[0], $offset, $rlen); |
| 69 | if (ord($record) == 0) { |
| 70 | return undef; |
| 71 | } |
| 72 | elsif (ord($record) == 1) { |
| 73 | } |
| 74 | elsif (substr($record, 1, $klen) eq $key) { |
| 75 | return substr($record, 1+$klen, $vlen); |
| 76 | } |
| 77 | &rehash; |
| 78 | } |
| 79 | } |
| 80 | |
| 81 | sub STORE { |
| 82 | local($self,$key,$val) = @_; |
| 83 | local($klen, $vlen, $tsize, $rlen) = @$self[1..4]; |
| 84 | croak("Table is full ($tsize->[0] elements)") if $$self[5] > $tsize->[0]; |
| 85 | croak(qq/Value "$val" is not $vlen characters long/) |
| 86 | if length($val) != $vlen; |
| 87 | my $writeoffset; |
| 88 | |
| 89 | &hashkey; |
| 90 | for (;;) { |
| 91 | $offset = $hash * $rlen; |
| 92 | $record = substr($$self[0], $offset, $rlen); |
| 93 | if (ord($record) == 0) { |
| 94 | $record = "\2". $key . $val; |
| 95 | die "panic" unless length($record) == $rlen; |
| 96 | $writeoffset = $offset unless defined $writeoffset; |
| 97 | substr($$self[0], $writeoffset, $rlen) = $record; |
| 98 | ++$$self[5]; |
| 99 | return; |
| 100 | } |
| 101 | elsif (ord($record) == 1) { |
| 102 | $writeoffset = $offset unless defined $writeoffset; |
| 103 | } |
| 104 | elsif (substr($record, 1, $klen) eq $key) { |
| 105 | $record = "\2". $key . $val; |
| 106 | die "panic" unless length($record) == $rlen; |
| 107 | substr($$self[0], $offset, $rlen) = $record; |
| 108 | return; |
| 109 | } |
| 110 | &rehash; |
| 111 | } |
| 112 | } |
| 113 | |
| 114 | sub DELETE { |
| 115 | local($self,$key) = @_; |
| 116 | local($klen, $vlen, $tsize, $rlen) = @$self[1..4]; |
| 117 | &hashkey; |
| 118 | for (;;) { |
| 119 | $offset = $hash * $rlen; |
| 120 | $record = substr($$self[0], $offset, $rlen); |
| 121 | if (ord($record) == 0) { |
| 122 | return undef; |
| 123 | } |
| 124 | elsif (ord($record) == 1) { |
| 125 | } |
| 126 | elsif (substr($record, 1, $klen) eq $key) { |
| 127 | substr($$self[0], $offset, 1) = "\1"; |
| 128 | return substr($record, 1+$klen, $vlen); |
| 129 | --$$self[5]; |
| 130 | } |
| 131 | &rehash; |
| 132 | } |
| 133 | } |
| 134 | |
| 135 | sub FIRSTKEY { |
| 136 | local($self) = @_; |
| 137 | $$self[6] = -1; |
| 138 | &NEXTKEY; |
| 139 | } |
| 140 | |
| 141 | sub NEXTKEY { |
| 142 | local($self) = @_; |
| 143 | local($klen, $vlen, $tsize, $rlen, $entries, $iterix) = @$self[1..6]; |
| 144 | for (++$iterix; $iterix < $tsize->[1]; ++$iterix) { |
| 145 | next unless substr($$self[0], $iterix * $rlen, 1) eq "\2"; |
| 146 | $$self[6] = $iterix; |
| 147 | return substr($$self[0], $iterix * $rlen + 1, $klen); |
| 148 | } |
| 149 | $$self[6] = -1; |
| 150 | undef; |
| 151 | } |
| 152 | |
| 153 | sub EXISTS { |
| 154 | croak "Tie::SubstrHash does not support exists()"; |
| 155 | } |
| 156 | |
| 157 | sub hashkey { |
| 158 | croak(qq/Key "$key" is not $klen characters long/) |
| 159 | if length($key) != $klen; |
| 160 | $hash = 2; |
| 161 | for (unpack('C*', $key)) { |
| 162 | $hash = $hash * 33 + $_; |
| 163 | &_hashwrap if $hash >= 1e13; |
| 164 | } |
| 165 | &_hashwrap if $hash >= $tsize->[1]; |
| 166 | $hash = 1 unless $hash; |
| 167 | $hashbase = $hash; |
| 168 | } |
| 169 | |
| 170 | sub _hashwrap { |
| 171 | $hash -= int($hash / $tsize->[1]) * $tsize->[1]; |
| 172 | } |
| 173 | |
| 174 | sub rehash { |
| 175 | $hash += $hashbase; |
| 176 | $hash -= $tsize->[1] if $hash >= $tsize->[1]; |
| 177 | } |
| 178 | |
| 179 | # using POSIX::ceil() would be too heavy, and not all platforms have it. |
| 180 | sub ceil { |
| 181 | my $num = shift; |
| 182 | $num = int($num + 1) unless $num == int $num; |
| 183 | return $num; |
| 184 | } |
| 185 | |
| 186 | # See: |
| 187 | # |
| 188 | # http://www-groups.dcs.st-andrews.ac.uk/~history/HistTopics/Prime_numbers.html |
| 189 | # |
| 190 | |
| 191 | sub findgteprime { # find the smallest prime integer greater than or equal to |
| 192 | use integer; |
| 193 | |
| 194 | my $num = ceil(shift); |
| 195 | return 2 if $num <= 2; |
| 196 | |
| 197 | $num++ unless $num % 2; |
| 198 | my $i; |
| 199 | my $sqrtnum = int sqrt $num; |
| 200 | my $sqrtnumsquared = $sqrtnum * $sqrtnum; |
| 201 | |
| 202 | NUM: |
| 203 | for (;; $num += 2) { |
| 204 | if ($sqrtnumsquared < $num) { |
| 205 | $sqrtnum++; |
| 206 | $sqrtnumsquared = $sqrtnum * $sqrtnum; |
| 207 | } |
| 208 | for ($i = 3; $i <= $sqrtnum; $i += 2) { |
| 209 | next NUM unless $num % $i; |
| 210 | } |
| 211 | return $num; |
| 212 | } |
| 213 | } |
| 214 | |
| 215 | 1; |