NAME

Net::FTP::Common - Perl extension for simplifying common usages of Net::FTP.

SYNOPSIS

use Net::FTP::Common;
%net_ftp_config = (Debug => 1, Timeout => 120);

%common_cfg = 
  (
    User => 'username',           # overwrite anonymous USER default
    Host => 'ftp.perl.com',       # overwrite ftp.microsoft.com HOST default
    Pass => 'tbone@soap-pan.org', # overwrite list@rebol.com PASS default
    Dir  => 'pub',                # overwrite slash DIR default
    Type => 'A'                   # overwrite I (binary) TYPE default
   );

# NOTE WELL!!! one is passed by reference, the other by value.
# This is inconsistent, but still it is A Very, Very Good Thing.
# Believe me! I thought about this. And I have a good reason for it:
# This is to allow the least modification of legacy Net::FTP source code.
$ez = Net::FTP::Common->new(\%common_cfg, %net_ftp_config); 

# can we login to the machine?
# Note: it is NEVER necessary to first login before calling Net::FTP::Common API functions.                               
# This function is just for checking to see if a machine is up. It is published as part of
# The API because I have found it useful when writing FTP scripts which scan for the 
# first available FTP site to use for upload.
# The exact call-and-return semantics for this function are described and justified below.
$ez->login || die "cant login: $@";

# Get a listing of a remote directory (list the 'Dir' option supplied to the constructor)
@listing =	$ez->dir; 

# Let's list a different directory, over-riding and changing the
# default directory
 @listing =	$ez->dir( New => { Dir => '/pub/rfcs' } ); 

# Let's list the default dir on several hosts
@host_listings = map { $ez->dir( New => { Host => $_ } ) } @host_list

# Let's get the listings of several directories
@dir_listings  = map { $ez->dir( New => { Dir  => $_ } ) } @dir_list;

# Get a file from the remote machine
$ez->get(File => 'codex.txt', LocalFile => '/tmp/crypto.txt');

# Send a file to the remote machine
$ez->send(File => 'codex.txt');

# test for a file's existence on the remote machine (using =~)
@file = $ez->grep(File => '[A-M]*[.]txt');

# a synonym for grep is glob (no difference, just another name)
@file = $ez->glob(File => 'n.*-file.t?t');

# test for a file on the remote machine (using eq)
$ez->exists(File => 'needed-file.txt');
# note this is no more than you manually calling:
# (scalar grep { $_ eq 'needed-file.txt' } $ez->dir) > 0;

DESCRIPTION

This module is intended to make the common uses of Net::FTP a one-line affair. Also, it made the development of Net::FTP::Shell straightforward.

Besides the constructor, all functions allow the over-riding of the initially supplied Net::FTP::Common configuration hashref by supplying a hashref with an API call.

Usage of this module is intended to be straightforward and stereotyped. The general steps to be used are:

• use Net::FTP::Common

• Define FTP configuration information

  This can be inlined within the script but oftentimes this will be stored in a module for usage in many other scripts.

• Use a Net::FTP::Common API function

  Note well that you NEVER have to login first. All API functions automatically log you in and change to the configured or specified directory. However, sometimes it is useful to see if you can actually login before attempting to do something else on an FTP site. This is the only time you will need the login() API method.

  Note well: though Net::FTP works in the stateful way that the FTP protocol does, Net::FTP::Common works in a stateless "one-hit" fashion. That is, for each separate call to the API, a connection is established, the particular Net::FTP::Common functionality is performed and the connection is dropped. The disadvantage of this approach is the (usually irrelevant and insignificant) overhead of connection and disconnection. The advantage is that there is much less chance of incurring failure due to timeout.

METHODS

new ( $net_ftp_common_hashref, %net_ftp_hash )

This method takes initialization information for Net::FTP::Common as well as Net::FTP and returns a new Net::FTP::Common object. Though the calling convention may seem a bit inconsistent, it is actually the best API to support re-use of legacy Net::FTP constructor calls. For example if you had a Net::FTP script which looked like this:

use Net::FTP;

$ftp = Net::FTP->new("some.host.name", Debug => 0);
$ftp->login("anonymous",'me@here.there');
$ftp->cwd("/pub");
$ftp->get("that.file");
$ftp->quit;

Here is all you would have to do to convert it to the Net::FTP::Common API:

           use Net::FTP::Common;

           $net_ftp_common_cfg = { Host => 'some.host.name', 
				   User => 'anonymous',
				   Pass => 'me@here.there',
				   Dir  => '/pub'
				   }	

           $ftp = Net::FTP::Common->new($net_ftp_common_cfg, Debug => 0);
           $ftp->get("that.file");
           $ftp->quit;

The example shows all the arguments that may be supplied to the Net::FTP::Common config hashref, except one, 'Type' which takes an argument of 'A' for ascii transfers and 'I' for binary transfers.

dir

When given no arguments, dir() uses Common configuration information to login to the ftp site, change directory and transfer type and then return an array of directory contents. You may only call this routine in array context and unlike Net::FTP, it returns a list representing the contents of the remote directory and in the case of no files, returns an empty array instead of (like Net::FTP) returning a 1-element array containing the element undef.

You may give this function any number of configuration arguments to over-ride the predefined configuration options. For example:

my %dir;
my @dir =qw (/tmp /pub /gnu);
map { @{$dir{$_}} = $ftp->dir(New => { Dir => $_ }) } @dir;

TRAPS FOR THE UNWARY

• @file = $ez->grep(File => '[A-M]*[.]txt');

  is correct

  @file = $ez->grep('[A-M]*[.]txt');

  looks correct but is not because you did not name the argument as you are supposed to.

EXPORT

None by default.

AUTHOR

T. M. Brannon <metaperl@yahoo.com>

SEE ALSO

REBOL (www.rebol.com) is a language which supports 1-line internet processing for the schemes of mailto:, http:, daytime:, and ftp:.

A Perl implementation of REBOL is in the works at www.metaperl.com.