| SPOPS documentation | Contained in the SPOPS distribution. |
SPOPS::HashFile - Implement as objects files containing perl hashrefs dumped to text
my $config = SPOPS::HashFile->new({ filename => '/home/httpd/myapp/server.perl',
perm => 'read' } );
print "My SMTP host is $config->{smtp_host}";
# Setting a different value is ok...
$config->{smtp_host} = 'smtp.microsoft.com';
# ...but this will throw an exception since you set the permission to
# read-only 'read' in the 'new' call
$config->save;
Implement a simple interface that allows you to use a perl data structure dumped to disk as an object. This is often used for configuration files, since the key/value, and the flexibility of the 'value' part of the equation, maps well to varied configuration directives.
new( { filename => $, [ perm => $ ] } )
Create a new SPOPS::HashFile object that uses the given filename
and, optionally, the given permission. The permission can be one of
three values: 'read', 'write' or 'new'. If you try to create a new
object without passing the 'new' permission, the action will die
because it cannot find a filename to open. Any value passed in that is
not 'read', 'write' or 'new' will get changed to 'read', and if no
value is passed in it will also be 'read'.
Note that the 'new' permission does not mean that a new file will overwrite an existing file automatically. It simply means that a new file will be created if one does not already exist; if one does exist, it will be used.
The 'read' permission only forbids you from saving the object or removing it entirely. You can still make modifications to the data in the object.
This overrides the new() method from SPOPS.
fetch( $filename, [ { perm => $ } ] )
Retrieve an existing config object (just a perl hashref data structure) from a file. The action will result in a 'die' if you do not pass a filename, if the file does not exist or for any reason you cannot open it.
save
Saves the object to the file you read it from.
remove
Deletes the file you read the object from, and blanks out all data in the object.
clone( { filename => $, [ perm => $ ] } )
Create a new object from the old, but you can change the filename and permission at the same time. Example:
my $config = SPOPS::HashFile->new( { filename => '~/myapp/spops.perl' } );
my $new_config = $config->clone( { filename => '~/otherapp/spops.perl',
perm => 'write' } );
$new_config->{base_dir} = '~/otherapp/spops.perl';
$new_config->save;
This overrides the clone() method from SPOPS.
No use of SPOPS::Tie
Use SPOPS::Tie
This is one of the few SPOPS implementations that will never use the
SPOPS::Tie class to implement its data holding. We still use a tied
hash, but it is much simpler -- no field checking, no ensuring that
the keys match in case, etc. This just stores some information about
the object (filename, permission, and data) and lets you go on your
merry way.
However, since we recently changed SPOPS::Tie to make field-checking optional we might be able to use it.
Copyright (c) 2001-2004 intes.net, inc.. All rights reserved.
This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.
Chris Winters <chris@cwinters.com>
| SPOPS documentation | Contained in the SPOPS distribution. |
package SPOPS::HashFile; # $Id: HashFile.pm,v 3.4 2004/06/02 00:48:21 lachoy Exp $ use strict; use base qw( SPOPS ); use Data::Dumper; $SPOPS::HashFile::VERSION = sprintf("%d.%02d", q$Revision: 3.4 $ =~ /(\d+)\.(\d+)/); # Just grab the tied hash from the SPOPS::TieFileHash sub new { my ( $pkg, $p ) = @_; my $class = ref $pkg || $pkg; my ( %data ); my $int = tie %data, 'SPOPS::TieFileHash', $p->{filename}, $p->{perm}; my $object = bless( \%data, $class ); $object->initialize( $p ); return $object; } # Subclasses can override sub initialize { return 1 } sub class_initialize { my $class = shift; return $class->_class_initialize( @_ ); } # Just pass on the parameters to 'new' sub fetch { my ( $class, $filename, $p ) = @_; $p ||= {}; return undef unless ( $class->pre_fetch_action( $filename, $p ) ); my $object = $class->new( { filename => $filename, %{ $p } } ); return undef unless( $object->post_fetch_action( $filename, $p ) ); return $object; } # Ensure we can write and that the filename is kosher, then # dump out the data to the file. sub save { my ( $self, $p ) = @_; my $obj = tied %{ $self }; unless ( $obj->{perm} eq 'write' ) { SPOPS::Exception->throw( "Cannot save [$obj->{filename}]; opened as read-only" ); } unless ( $obj->{filename} ) { SPOPS::Exception->throw( "Cannot save data: the filename has been " . "erased. Did you assign an empty hash to the object?" ); } my $temp_filename = "$obj->{filename}.tmp"; if ( -f $temp_filename ) { unlink( $temp_filename ); # just to be sure... } if ( -f $obj->{filename} ) { rename( $obj->{filename}, $temp_filename ) || SPOPS::Exception->throw( "Cannot rename old file to make room for new one. Error: $!" ); } return undef unless ( $self->pre_save_action( $p ) ); my %data = %{ $obj->{data} }; $p->{dumper_level} ||= 2; local $Data::Dumper::Indent = $p->{dumper_level}; eval { open( INFO, "> $obj->{filename}" ) || die $! }; if ( $@ ) { rename( $temp_filename, $obj->{filename} ) || SPOPS::Exception->throw( "Cannot open file for writing [$@] and " . "cannot move backup file to original place [$!]" ); SPOPS::Exception->throw( "Cannot open file for writing [$@]. Backup file restored ok." ); } print INFO Data::Dumper->Dump( [ \%data ], [ 'data' ] ); close( INFO ); if ( -f $temp_filename ) { unlink( $temp_filename ) || warn "Cannot remove the old data file. It still lingers in ($temp_filename)\n"; } return undef unless ( $self->post_save_action( $p ) ); return $self; } sub remove { my ( $self, $p ) = @_; my $obj = tied %{ $self }; unless ( $obj->{perm} eq 'write' ) { SPOPS::Exception->throw( "Cannot save [$obj->{filename}]; opened as read-only" ); } unless ( $obj->{filename} ) { SPOPS::Exception->throw( "Cannot remove data: the filename has been " . "erased. Did you assign an empty hash to the object?" ); } return undef unless ( $self->pre_remove_action( $p ) ); my $rv = %{ $self } = (); return undef unless ( $self->post_remove_action( $p ) ); return $rv; } # Create a new object from an old one, allowing any passed-in # values to override the ones from the old object sub clone { my ( $self, $p ) = @_; unless ( $p->{filename} ) { $p->{filename} = (tied %{ $self })->{filename}; } my $new = $self->new({ filename => $p->{filename}, perm => $p->{perm} }); while ( my ( $k, $v ) = each %{ $self } ) { $new->{ $k } = $p->{ $k } || $v; } return $new; } package SPOPS::TieFileHash; use strict; use File::Copy qw( cp ); $SPOPS::TieFileHash::VERSION = sprintf("%d.%02d", q$Revision: 3.4 $ =~ /(\d+)\.(\d+)/); # These are all very standard routines for a tied hash; more info: see # 'perldoc Tie::Hash' # Ensure that the file exists and can be read (unless they pass in # 'new' for the permissions, which means it's ok to start out with # blank data); store the meta info (permission and filename) in the # object, and the 'data' key holds the actual information sub TIEHASH { my ( $class, $filename, $perm ) = @_; $perm ||= 'read'; if ( $perm !~ /^(read|write|new|write\-new)$/ ) { SPOPS::Exception->throw( "Invalid permissions [$perm]; valid: [read|write|new|write-new]" ); } unless ( $filename ) { SPOPS::Exception->throw( "You must pass a filename to use for reading and writing." ); } my $file_exists = ( -f $filename ); unless ( $file_exists ) { if ( $perm eq 'write-new' or $perm eq 'new' ) { $perm = 'new'; } else { SPOPS::Exception->throw( "Cannot create object without existing file " . "or 'new' permission [$filename] [$perm]" ); } } $perm = 'write' if ( $perm eq 'write-new' ); my $data = undef; if ( $file_exists ) { # First create a backup... cp( $filename, "${filename}.backup" ); # Then open up the file open( PD, $filename ) || SPOPS::Exception->throw( "Cannot open [$filename]: $!" ); local $/ = undef; my $info = <PD>; close( PD ); # Note that we create the SIG{__WARN__} handler here to trap any # messages that might be sent to STDERR; we want to capture the # message and send it along in a 'die' instead { local $SIG{__WARN__} = sub { return undef }; no strict 'vars'; $data = eval $info; } if ( $@ ) { SPOPS::Exception->throw( "Error reading in perl code: $@" ); } } else { $data = {}; $perm = 'write'; } return bless({ data => $data, filename => $filename, perm => $perm }, $class ); } sub FETCH { my ( $self, $key ) = @_; return undef unless $key; return $self->{data}->{ $key }; } sub STORE { my ( $self, $key, $value ) = @_; return undef unless $key; return $self->{data}->{ $key } = $value; } sub EXISTS { my ( $self, $key ) = @_; return undef unless $key; return exists $self->{data}->{ $key }; } sub DELETE { my ( $self, $key ) = @_; return undef unless $key; return delete $self->{data}->{ $key }; } # This allows people to do '%{ $obj } = ();' and remove the object; is # this too easy to mistakenly do? I don't think so. sub CLEAR { my ( $self ) = @_; if ( $self->{perm} ne 'write' ) { SPOPS::Exception->throw( "Cannot remove [$self->{filename}]; " . "permission set to read-only" ); } unlink( $self->{filename} ) || SPOPS::Exception->throw( "Cannot remove [$self->{filename}]: $!" ); $self->{data} = undef; $self->{perm} = undef; } sub FIRSTKEY { my ( $self ) = @_; keys %{ $self->{data} }; my $first_key = each %{ $self->{data} }; return undef unless ( $first_key ); return $first_key; } sub NEXTKEY { my ( $self ) = @_; my $next_key = each %{ $self->{data} }; return undef unless ( $next_key ); return $next_key; } 1; __END__