archmbox - a simple email archiver

archmbox [ -h | --version ]

archmbox MODE [ OPTIONS ] -d date mailbox [ mailbox ... ]

archmbox MODE [ OPTIONS ] -o days mailbox [ mailbox ... ]

Archmbox is a simple email archiver written in perl; it parses one or more mailboxes, select some or all messages and then perform specific actions on the selected messages.

Four different MODES are available:

•

list mode, which is useful to list all selected messages before archmbox performs the real operations (archiving or deleting)

•

kill mode, if messages should be deleted from the mailbox(es) rather than archived

•

archive mode, to archive the selected messages in a different mailbox

•

copy mode, to copy selected messages from a source mailbox(es) without modifying it

Messages selection is based upon a date criteria; an absolute date or a days offset can be specified. It is also possible to refine the selection using regular expressions on the header fields of the message.

All archived messages are stored in a new mailbox with the same name of the original one + .archived as extension (this is the default, but can be changed); the archive mailbox can be saved in gz or bz2 compressed format as well.

Please note that the archive mailbox format is always mbox, regardless of original mailbox format. Moreover, mailboxes must be specified using the full path.

Messages are appended to the archive mailbox to allow multiple executions of the script against the same mailbox.

-a, --archive

Selected messages are archived in a different mailbox.

-k, --kill

Selected messages are deleted rather than archived.

-l, --list

List all selected messages.

-y, --copy

Selected messages are copied from the source mailbox.

-b, --backup

Creates a backup of the original mailbox before archmbox execution. The mailbox is called mailbox.backup

--bzip2

Use bzip2 to compress the archive mailbox (use with -c).

-c, --compress

Compress the archive mailbox after script execution.

-d, --date <date>

Specifies the threshold date for messages. The date must be supplied in the following format: yyyy-mm-dd

-D, --date-header

Force the use of the "Date:" header to age a message. If the header is somehow corrupt, the date/time informations are gathered for the beginning line of the message.

-e, --extension <extension>

Specifies the suffix for the archive mailbox; the default is archived. If none is specified, no suffix will be used (use carefully).

-f, --full-name

Prepends the path of the mailbox to the name of the archive mailbox.

--format

Specifies the format of the mailboxes to parse. Legal values are mbox and mbx. Defaults to "mbox".

-h, --help

Prints help.

-i, --ignore <regexp>

Any mailbox/directory matching <regexp> will be skipped while archiving.

--keep-flagged

Flagged messages will not be archived.

--keep-unread

Unread messages will not be archived.

-m, --minsize

Specifies the minimum size of the mailbox to be archived. Mailboxes smaller than <minsize> will not be parsed for archiving.

--nosymlink

Do not follow symbolic links when processing mailboxes.

--nowarnings

Suppress mailbox related warnings. Use only if you know what you're doing!

--omit-prefix <prefix>

Omit <prefix> from the name of the mailbox when full name (option -f) is required.

-o, --offset <days>

Specifies the offset (in days) from today for threshold date of a message. This option replaces -d. If you specify -1, archmbox will operate on all messages.

-p, --path <directory>

Specifies where to store the archive mailbox (default: "."). <directory> must be specified using full path.

-r, --reverse

Reverse the sense of offset or date value. It usually means older than but with this switch, it means newer than.

-R, --recursive

Act recursively on directories. If one or more directories are specified on the command line, all mailboxes stored in those directories will be parsed for archiving. Implies option -f.

-t, --tmpdir <directory>

Specify a temporary working directory. This value overrides the one specified at compilation time or the one hard coded in the script.

<directory> must be specified using full path.

--time <time>

Use <time> in conjunction with <date> (option -d) to refine the threshold age for archiving. <time> must be specified in the following format: hh:mm:ss.

--totals

Prints an overall summary of the archiving operations. The summary contains the number of parsed and skipped mailboxes, the total number of messages parsed and saved, the total space used and saved.

-v, --verbose

Verbosity level. Default is 1 (line per message) in --list output. So, if set to 1 it only lists msgid, sender and subject. With -v=2, it also prints date.

--version

Prints version number.

-x, --regexp <header=regexp>

It is specified in form -x field='regexp', where field can be any header. The header part is case sensitive, while the regexp part is not.

If message satisfies date range, but does not satisfy regexp match on specified field, it won't be archived.

The option can be specified more than once; in this case, the message is regexp matched against all the given rules, an if it satisfies any, it will be archived.

Archmbox is completely written in perl, but it uses some shell helpers to perform its job (fuser, rm, gzip/gunzip etc.).

The correct path for the helpers (both required and optional ones) is probed at installation time. If one required helper is missing the installation will not take place. If one optional helper is missing, the feature provided using that helper will be unavailable, but the script will be installed anyway.

All other relevant configuration options can be specified at installation time or at run time using the command line switches.

A complete example:

archmbox -a -b -c -e 01 -f -d 2002-01-01 -p ~/mail-archive ~/Mail/personal-stuff

This will archive all messages older than (received before...) Jan 1st 2002 from the personal-stuff mailbox in the Mail directory. Archive messages are saved in a mailbox called Mail-personal-stuff.01.gz in the ~/mail-archive directory. After execution, you'll find a mailbox called personal-stuff.backup in ~/Mail.

Some simpler examples:

archmbox -a -o 15 ~/Mail/personal-stuff

This will archive all messages older than 15 days in personal-stuff.archived (uncompressed mailbox).

archmbox -a -r -o 15 ~/Mail/personal-stuff

The same as above, but only messages newer than 15 days will be archived.

archmbox -k -o 15 ~/Mail/personal-stuff

This will delete all messages older than 15 days from Mail/personal-stuff

archmbox -a -o 15 ~/Mail/* -c

This will archive all messages older than 15 days in every mailbox found in ~/Mail. All the archive mailboxes will be compressed.

archmbox -l -r -c /tmp/mbox -o 20

List all messages in /tmp/mbox which are newer than 20 days. Option -c is meaningless (and so ignored...).

archmbox -l -r -c /tmp/mbox -o 20 -a --bzip2

Same as above, but archiving is forced (-a) and bzip2 is used for compression.

archmbox -a -x subject='archmbox' -o 7 ~/mbox

Select for archiving all messages older than 7 days whose subject field satisfies regexp match subject =~ /archmbox/ (it is case insensitive).

archmbox -l -x subject='archmbox' -x from='fritz' -o 7 ~/mbox

Select for archiving all messages older than 7 days whose subject field contains archmbox or the sender is fritz (matches are case insensitive).

archmbox -a -o 5 -R /tmp/mbox ~/Mail

archmbox will archive all messages older than five days in /tmp/mbox. It then start parsing all mailboxes stored in ~/Mail (recursion is active, and ~/Mail is a directory). If one or more directories will be found in ~/Mail, those directories will be explored as well.

archmbox -a -o -1 ~/Mail/my_mbx_mailbox --format mbx

archmbox archives all messages stored in my_mbx_mailbox and puts them into my_mbx_mailbox.archived. The source mailbox is a mbx mailbox (--format mbx is used). The archive mailbox will be a mbox mailbox.

When the script has to decide if a message needs to be selected from the mailbox, it looks for the header From generated by the mail server (this is the first line of the message) and doesn't care about the date specified by the sender's mail client. This is useful to avoid removing messages sent from misconfigured mail clients. This behavior can be changed by forcing the use of the "Date:" header (option -D).

Not all options are meaningfull in all modes, ie compression is meaningless in list or kill mode. If you specify a useless option for a particular mode, archmbox simply ignores it.

Archmbox uses a working directory to store temporary mailboxes. A default value for that directory is hard coded in the script, but can be changed during the configuration/installation process (see INSTALL for details). It might happen that your mailboxes are too big for the partition holding this temporary directory, or you might want to perform archiving on too much mailboxes at the same time. In other words, you may run out of space. Use the -t option to specify a suitable working directory at runtime.

If you see some differences in the mailbox's dimension (size/free space), keep in mind that your mailbox may contain a special message (512 bytes in size) with internal information related to the mailbox. This message is meaningless for you, though archmbox recognizes it and lets you be aware of it. That message is left untouched in your source mailbox.

A few words about locking. There has been a discussion about archmbox handles file locking. The answer is simple: no mailbox is ever locked. The reason behind this behavior is that I want archmbox to be as least invasive as possible, so other kind of checks are performed to ensure that no data is lost (mailbox has changed/mailbox is in use by another program). I will surely add some locking mechanism in the future.

You don't need to execute archmbox as root... just take care to have write permissions for the directories you use.

Archmbox can be downloaded from:

http://adc-archmbox.sourceforge.net

Archmbox is distributed under the terms of the GPL

Copyright (C) 2001-2005

Alessandro Dotti Contra <adotti@users.sourceforge.net>

Parts of the code were contributed by:

Alex Aminoff, Brian Medley, Buck Holsinger, Davor Ocelic, Fabrice Noilhan, Jayanth Varma, Juergen Edner, Laurent Cheylus, Nicolas Ecarnot, Paco Regodon, Scott Thompson.

The FreeBSD port is maintained by Talal Al-Dik.

The OpenDarwin port is maintained by Markus Weissman.

Please report bugs to <adotti@users.sourceforge.net>