Attribute::Persistent - Really lazy persistence


    use Attribute::Persistent;

    my %hash :persistent;
    $hash{counter}++; # Value retained between calls to the program.

    my %hash2 :persistent(SessionTable); # Explicitly provide a filename.


This module provides a way of abstracting away persistence of array and hash variables.

It's useful for quick hacks when you don't care about pulling in the right DBM library and calling tie and so on. Its job is to reduce fuss for the lazy programmer at the cost of flexibility.

It uses MLDBM, so you can use complex data structures in your arrays and hashes. It uses AnyDBM_File, so if you really care about which DBM you get, you can modify AnyDBM_File::ISA in a BEGIN block after loading this module.

It works out which DBMs belong to it by taking an md5 sum of the source code. This means that if you change your code, you lose your data. If you like to keep your data while messing about with your code, you need to explicitly give Attribute::Persistent a key, like this:

    BEGIN { $Attribute::Persistent::KEY = "MyProgram"; }
    use Attribute::Persistent; # Order is important.

This uniquely identifies your program, meaning that the module doesn't have to grub around with $0 and md5 sums.

But hell, it's not supposed to be this complex. Just use the module and slap :persistent onto your lexicals where appropriate, and it just works. That's all most people need to care about.


Originally by Simon Cozens,

Maintained by Scott Penrose,


Artistic and GPL.