1Tie::Memoize(3pm)      Perl Programmers Reference Guide      Tie::Memoize(3pm)
2
3
4

NAME

6       Tie::Memoize - add data to hash when needed
7

SYNOPSIS

9         require Tie::Memoize;
10         tie %hash, 'Tie::Memoize',
11             \&fetch,                  # The rest is optional
12             $DATA, \&exists,
13             {%ini_value}, {%ini_existence};
14

DESCRIPTION

16       This package allows a tied hash to autoload its values on the first
17       access, and to use the cached value on the following accesses.
18
19       Only read-accesses (via fetching the value or "exists") result in calls
20       to the functions; the modify-accesses are performed as on a normal
21       hash.
22
23       The required arguments during "tie" are the hash, the package, and the
24       reference to the "FETCH"ing function.  The optional arguments are an
25       arbitrary scalar $data, the reference to the "EXISTS" function, and
26       initial values of the hash and of the existence cache.
27
28       Both the "FETCH"ing function and the "EXISTS" functions have the same
29       signature: the arguments are "$key, $data"; $data is the same value as
30       given as argument during tie()ing.  Both functions should return an
31       empty list if the value does not exist.  If "EXISTS" function is
32       different from the "FETCH"ing function, it should return a TRUE value
33       on success.  The "FETCH"ing function should return the intended value
34       if the key is valid.
35

Inheriting from Tie::Memoize

37       The structure of the tied() data is an array reference with elements
38
39         0:  cache of known values
40         1:  cache of known existence of keys
41         2:  FETCH  function
42         3:  EXISTS function
43         4:  $data
44
45       The rest is for internal usage of this package.  In particular, if
46       TIEHASH is overwritten, it should call SUPER::TIEHASH.
47

EXAMPLE

49         sub slurp {
50           my ($key, $dir) = shift;
51           open my $h, '<', "$dir/$key" or return;
52           local $/; <$h>                      # slurp it all
53         }
54         sub exists { my ($key, $dir) = shift; return -f "$dir/$key" }
55
56         tie %hash, 'Tie::Memoize', \&slurp, $directory, \&exists,
57             { fake_file1 => $content1, fake_file2 => $content2 },
58             { pretend_does_not_exists => 0, known_to_exist => 1 };
59
60       This example treats the slightly modified contents of $directory as a
61       hash.  The modifications are that the keys fake_file1 and fake_file2
62       fetch values $content1 and $content2, and pretend_does_not_exists will
63       never be accessed.  Additionally, the existence of known_to_exist is
64       never checked (so if it does not exists when its content is needed, the
65       user of %hash may be confused).
66

BUGS

68       FIRSTKEY and NEXTKEY methods go through the keys which were already
69       read, not all the possible keys of the hash.
70

AUTHOR

72       Ilya Zakharevich mailto:perl-module-hash-memoize@ilyaz.org
73       <mailto:perl-module-hash-memoize@ilyaz.org>.
74
75
76
77perl v5.16.3                      2013-02-26                 Tie::Memoize(3pm)
Impressum