NAME

Net::FTP::Recursive - Recursive FTP Client class

SYNOPSIS

use Net::FTP::Recursive;

$ftp = Net::FTP::Recursive->new("some.host.name", Debug => 0);
$ftp->login("anonymous",'me@here.there');
$ftp->cwd('/pub');
$ftp->rget( ParseSub => \&yoursub );
$ftp->quit;

DESCRIPTION

Net::FTP::Recursive is a class built on top of the Net::FTP package that implements recursive get and put methods for the retrieval and sending of entire directory structures.

This module's default behavior is such that the remote ftp server should understand the "dir" command and return UNIX-style directory listings. If you'd like to provide your own function for parsing the data retrieved from this command (in case the ftp server does not understand the "dir" command), all you need do is provide a function to one of the Recursive method calls. This function will take the output from the "dir" command (as a list of lines) and should return a list of Net::FTP::Recursive::File objects. This module is described below.

All of the methods also take an optional KeepFirstLine argument which is passed on to the default parsing routine. This argument supresses the discarding of the first line of output from the dir command. wuftpd servers provide a total line, the default behavior is to throw that total line away. If yours does not provide the total line, KeepFirstLine is for you. This argument is used like the others, you provide the argument as the key in a key value pair where the value is true (ie, KeepFirstLine => 1).

When the Debug flag is used with the Net::FTP object, the Recursive package will print some messages to STDERR.

CONSTRUCTOR

new (HOST [,OPTIONS])

A call to the new method to create a new Net::FTP::Recursive object just calls the Net::FTP new method. Please refer to the Net::FTP documentation for more information.

METHODS

rget ( [ParseSub =>\&yoursub] [,FlattenTree => 1] [,RemoveRemoteFiles => 1] )

The recursive get method call. This will recursively retrieve the ftp object's current working directory and its contents into the local current working directory.

This will also take an optional argument that will control what happens when a symbolic link is encountered on the ftp server. The default is to ignore the symlink, but you can control the behavior by passing one of these arguments to the rget call (ie, $ftp->rget(SymlinkIgnore => 1)):

SymlinkIgnore - disregards symlinks (default)

SymlinkCopy - copies the link target from the server to the client (if accessible). Works on files other than a directory. For directories, see the SymlinkFollow option.

SymlinkLink - creates the link on the client.

SymlinkFollow - will recurse into a symlink if it points to a directory. This option may be given along with one of the others above.

The SymlinkFollow option, as of v1.6, does more sophisticated handling of symlinks. It will detect and avoid cycles, on all client platforms. Also, if on a UNIX (tm) platform, if it detects a cycle, it will create a symlink to the location where it downloaded the directory (or will download it subsequently, if it is in the subtree under where the recursing started). On Windows, it will call symlink just as on UNIX (tm), but that's probably not gonna do much for you. :)

The FlattenTree optional argument will retrieve all of the files from the remote directory structure and place them in the current local directory. This option will resolve filename conflicts by retrieving files with the same name and renaming them in a "$filename.$i" fashion, where $i is the number of times it has retrieved a file with that name.

The optional RemoveRemoteFiles argument to the function will allow the client to delete files from the server after it retrieves them. The default behavior is to leave all files and directories intact.

For the v1.6 release, I have also added some additional functionality that will allow the client to be more specific in choosing those files that are retrieved. All of these options take a regex object (made using the qr operator) as their value. You may choose to use one or more of these options, they are applied in the order that they are listed. They are:

MatchAll - Will process file that matches this regex, regardless of whether it is a plainish file, directory, or a symlink. This behavior can be overridden with the previous options.

OmitAll - Do not process file that matches this regex. Also may be overridden with the previous options.

MatchFiles - Only transfer plainish (not a directory or a symlink) files that match this pattern.

OmitFiles - Omit those plainish files that match this pattern.

MatchDirs - Only recurse into directories that match this pattern.

OmitDirs - Do not recurse into directories that match this pattern.

MatchLinks - Only deal with those links that match this pattern (based on your symlink option, above).

OmitLinks - Do not deal with links that match this pattern.

Currently, the added functionality given to the rget method is not implemented for the rput method.

rput ( [ParseSub => \&yoursub] [,DirCommand => $cmd] [,FlattenTree => 1] [,RemoveLocalFiles => 1])

The recursive put method call. This will recursively send the local current working directory and its contents to the ftp object's current working directory.

This method will take an optional set of arguments to tell it what the local directory listing command will be. By default, this is "ls -al". If you change the behavior through this argument, you probably also need to provide a ParseSub, as described above.

This will take an optional argument that will control what happens when a symbolic link is encountered on the ftp server. The default is to ignore the symlink, but you can control the behavior by passing one of these arguments to the rput call (ie, $ftp->rput(SymlinkIgnore => 1)):

SymlinkIgnore - disregards symlinks

SymlinkCopy - will copy the link target from the client to the server.

SymLinkFollow - will recurse into a symlink if it points to a directory. This does not do cycle checking, use with EXTREME caution. This option may be given along with one of the others above.

The FlattenTree optional argument will send all of the files from the local directory structure and place them in the current remote directory. This option will resolve filename conflicts by sending files with the same name and renaming them in a "$filename.$i" fashion, where $i is the number of times it has retrieved a file with that name.

The optional RemoveLocalFiles argument to the function will allow the client to delete files from the client after it sends them. The default behavior is to leave all files and directories intact. This option will only attempt to delete files that were actually transferred, symlinks (unless you set SymlinkCopy and it is a plain file) will not be removed (and of course any non-empty directories).

rdir ( Filehandle => $fh [, FilenameOnly => 1] [, ParseSub => \&yoursub ] )

The recursive dir method call. This will recursively retrieve directory contents from the server in a breadth-first fashion.

The method needs to be passed a filehandle to print to. The method call just does a print $fh, so as long as this call can succeed with whatever you pass to this function, it'll work.

The second, optional argument, is to retrieve only the filenames (including path information). The default is to display all of the information returned from the $ftp-dir call.

This method does NOT follow symlinks, it will only print them out. This will be changed for the next version of the module

rls ( Filehandle => $fh [, ParseSub => \&yoursub] )

The recursive ls method call. This will recursively retrieve directory contents from the server in a breadth-firth fashion. This is equivalent to calling $ftp-rdir( Filehandle => $fh, FilenameOnly => 1 )>.

rdelete ( [ ParseSub => \&yoursub ] )

The recursive delete method call. This will recursively delete everything in the directory structure. This disregards the SymlinkFollow option and does not recurse into symlinks that refer to directories.

Net::FTP::Recursive::File

This is a helper class that encapsulates the data representing one file in a directory listing.

METHODS

new ( )

This method creates the File object. It should be passed several parameters. It should always be passed:

OriginalLine => $line

Fields => \@fields

And it should also be passed at least one (but only one a true value) of:

IsPlainFile => 1

IsDirectory => 1

IsSymlink => 1

OriginalLine should provide the original line from the output of a directory listing.

Fields should provide an 8 element list that supplies information about the file. The fields, in order, should be:

Permissions

Link Count

User Owner

Group Owner

Size

Last Modification Date/Time

Filename

Link Target

The IsPlainFile, IsDirectory, and IsSymlink fields need to be supplied so that for the output on your particular system, your code (in the ParseSub) can determine which type of file it is so that the Recursive calls can take the appropriate action for that file. Only one of these three fields should be set to a "true" value.

TODO LIST

Cycle checking with symlinks (mostly finished)

Allow for formats to be given for output on rdir/rls.

Allow for a regex of directories/files to omit

REPORTING BUGS

When reporting bugs, please provide as much information as possible. A script that exhibits the bug would also be helpful, as well as output with the "Debug => 1" flag turned on in the FTP object.

AUTHOR

Jeremiah Lee <texasjdl@yahoo.com>

SEE ALSO

Net::FTP

Net::Cmd

ftp(1), ftpd(8), RFC 959

CREDITS

(in Chronological order)

Andrew Winkler - for various input into the module. Raj Mudaliar - documentation fix. Brian Reischl - for rdelete code. Chris Smith - for RemoveRemoteFiles code. Zainul Charbiwala - bug report & code to fix. Brian McGraw - bug report & feature request.

COPYRIGHT

Copyright (c) 2001-2003 Jeremiah Lee.

This program is free software; you may redistribute it and/or modify it under the same terms as Perl itself.

cpanm Net::FTP::Recursive

perl -MCPAN -e shell
install Net::FTP::Recursive