LDP/LDP/howto/linuxdoc/Outlook-to-Unix-Mailbox.sgml

<!doctype linuxdoc system>

<article>

<title>MS Outlook to Unix Mailbox Conversion mini HOWTO

<author>Greg Lindahl, <tt/lindahl@pbm.com/

<date>v1.4, 2004-01-08

<abstract>
This MiniHowto covers conversion of old email in Microsoft 
Outlook (not Outlook Express!) to typical Unix file formats.
</abstract>


<!-- Table of contents -->
<toc>

<!-- Begin the document -->

<sect>Introduction

<p>
While several programs exist to convert some formats such as Microsoft
Outlook Express to Unix formats, Outlook users have a bit more of a
challenge. One way to convert uses Mozilla Mail under Windows; another
involves a more complicated method. Both are explained in this
miniHOWTO.

<p>
The database format that Outlook uses for .PST files, called Jet, is
documented at:

<url url="http://msdn.microsoft.com/library/techart/olexcoutlk.htm"
     name="http://msdn.microsoft.com/library/techart/olexcoutlk.htm">

<sect1>Copyright
<p>

Copyright (c) 2001-2004 by Greg Lindahl

<p>

Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.1 or
any later version published by the Free Software Foundation; with no
Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
Texts. A copy of the license may be found at:

<url url="http://www.gnu.org/copyleft/fdl.html"
     name="http://www.gnu.org/copyleft/fdl.html">

<p>

I request that corrections and/or comments be forwarded to the
document maintainer. If you're considering making a derived work other
than a translation, I request that you discuss your plans with
the current maintainer.

<sect1>Disclaimer
<p>
Use the information in this document at your own risk. I disavow any
potential liability for the contents of this document. Use of the
concepts, examples, and/or other content of this document is entirely
at your own risk.

All copyrights are owned by their owners, unless specifically noted
otherwise.  Use of a term in this document should not be regarded as
affecting the validity of any trademark or service mark.

Naming of particular products or brands should not be seen as endorsements.

You are strongly recommended to take a backup of your system before
major installations, and backup your system at regular intervals.

Do not place your cat in a running microwave oven.

<sect1>News
<p>
1.01: Fixed minor typo in inetd/xinetd startup instructions.

1.1: Added information about Kmailcvt, Mozilla, and how to get
Outlook to talk to IMAP servers if it's in Exchange mode.

1.2: Added details about using Mozilla to do this same task

1.2.1: Fixed formatting typo.

1.2.2: Relicensed under the GFDL, more minor typo fixes.

1.3: Yet more minor fixes.

1.4: Information about TNEF stuff from Scott Phelps (thanks!)

<sect1>Other ways of doing this
<p>
A list of conversion utilities, many commercial, may be found at:

<url url="http://www.emailman.com/conversion/index.html"
     name="http://www.emailman.com/conversion/index.html">

I've had a claim that the program Kmailcvt now converts Outlook
mailboxes. However, I'm not 100% sure that this claim is true, since
Kmailcvt definitely converts Outlook Express mailboxes, which are
completely different from Outlook mailboxes. So, if you figure this
out, please let me know.

<sect>Converting using Mozilla Mail
<p>

I've also heard that Mozilla Mail under Windows can convert Outlook
mailboxes -- but the only solid report I've gotten was that
attachments were not converted, so if they matter to you, don't use
this method.

There is some documentation on the Ximian Evolution website. I haven't
tried this since I don't have a Windows machine anymore. But, here's
what they say to do:

Run Mozilla Mail

Go to "File > Import" and select that you wish to import
mail from Outlook. When that's done, you're almost there.

I believe that Mozilla stores all its mailboxes as Unix mbox format
files, even under Windows. So, all you have to do is transfer those
files to your Linux box. You can find them in:

C:/windows/Application Data/Mozilla/Profiles/default/XXX/Mail/imported.mail/, where XXX will
be some collection of digits. If there are multiple users on your
Windows machine, "default" will instead be your username.

As I mentioned earlier, I've never tried this method, so I don't know
how well it works. In particular, the issues raised in section 3.4
probably also apply to this method of conversion.

By the way, the reason that Mozilla can only read Outlook mailboxes
under Windows is because it calls a Microsoft DLL to do it.

<sect>Converting using IMAP

<sect1>Preparation
<p>
First, be sure you want to do things this way. In particular, section
2 explains how to use Netscape Mail under Windows to do the
coversion. That's easier than doing it this way.

If you decide to do it this way, you need to make sure that your copy
of Outlook can talk to IMAP servers. When I first wrote this HOWTO, I
assumed that Outlook was Outlook was Outlook. Nope. If you are using
Outlook in Corporate/Workgroups mode (which talks to Exchange) instead
of Internet Mode (which talks to POP/IMAP servers), you'll have to
change modes.

To test to see if you're OK, look at the Outlook "Tools" menu, and see
if there is an "Accounts" item. If there is, you're OK, and you can go
on to section 3.2. If there isn't, go find your towel, and keep reading.

Changing modes is not trivial, but a helpful reader (Matt Huyck)
sent in the following instructions, which look dangerous, and
which I have not tested:

Select "Options..." from the "Tools" menu, and then go to the "Mail
Services" tab in the resulting Options dialog box.  At the bottom of
that tab there is a button labeled "Reconfigure Mail Support..."
Hold your breath and then click it.  A new window comes up with
"Outlook 2000 Startup" in the title bar and an "E-mail Service
Options" heading.  There are two relevant radio buttons: "Internet
Only" and "Corporate or Workgroup".  Change over to "Internet Only"
and click "Next >".  You are then prompted with a very long message
box which you should read carefully because you're about to make a
significant (but reversible) change to the Exchange client
configuration.  If you haven't passed out already, you can stop
holding your breath now.  Although it doesn't explicitly say so, you
will want to be sure you have a copy of the Microsoft Office install
CD before you click "Yes".  Click "Yes" and let Outlook do its thing
for a few seconds until it has quit completely.  Open up Outlook
again.  This is where you may be prompted for the install CD.  After
the re-configuration is complete you're ready to proceed with step
3.2 of the HOWTO.

To get back to your original Outlook configuration, follow the same
directions, but you'll obviously be clicking the *other* radio button.

One other thing is different if you've been using Exchange.  As you
point out at the end of section 3.4,  "the original 'From ' line" is not
preserved.  For Exchange users, however, the comment "Fortunately you
don't actually need that information" doesn't quite apply.  The
"From" header that's missing is the only one that contains the
identity of the sender in messages that were sent on an internal
Exchange server, i.e. messages that didn't pass through an Internet
gateway anywhere.  I have preserved my "From" headers by saving
copies of my mail folders as text files through the "Import and
Export..." command from the "File" menu.  I plan to hack out some
twisted Perl/Grep code that will re-insert those "From" headers into
my Linux mbox files.  If I get that to work I'll let you know.

<sect1>Install an IMAP server (temporarily!) on your Linux box
<p>
Installing things varies from Linux distribution to distribution, so I
will use RedHat 7.0 as an example.  First you need to install the
correct package, which generally is named "imap". or "uw-imap" or
something like that:
<tscreen><verb>
 cd /home/redhat-7.0-cd/RedHat/RPMS
 ls *imap*
 rpm -i imap*
</verb></tscreen>

Actually, since I had a "workstation" install, I also had to install
the xinetd package; rpm gave me an error which told me to do
this. And, of course, it was on the second CD of RedHat 7.0. Debian
users using "apt-get" don't have to worry about such issues.

<p>
Next, we need to enable the imap server. On my Gnome-basekd desktop
this is done using the graphical tool at Start -> System Settings ->
Server Settings -> Services, or you can do it by hand. This is usually
controlled by a line in the file <tt>/etc/inetd.conf</tt>:
<tscreen><verb>
 #imap    stream  tcp     nowait  root    /usr/sbin/tcpd	/usr/sbin/imapd
</verb></tscreen>

The above line is commented out; remove the leading # sign. Alternately,
if your system uses xinetd, instead edit <tt>/etc/xinetd.d/imap</tt>
and change "disable=yes" to "disable=no".

<p>
Then restart inetd or xinetd by doing:
<tscreen><verb>
 /etc/rc.d/init.d/inetd restart
</verb></tscreen>

or
<tscreen><verb>
 /etc/rc.d/init.d/xinetd restart
</verb></tscreen>

If all else fails, reboot.

<p>
You don't actually want to leave the IMAP server enabled for that
long. This server runs as root and has had security bugs in the
past. For this reason, you shouldn't leave it enabled unless you wish
to use it permanently. We will disable this server in section 3.5.

<p>
In order to connect Outlook to this IMAP server, you will need to know
the name or IP address of the Linux box.

<sect1>Connect your Outlook client to the server
<p>
In order to copy over all our email to the server, we need to tell
your Outlook client about this new server. Select "Accounts..." from
the "Tools" menu, and then "Add" a new account "Mail...". The
important items are that the server uses IMAP to download email, that
the incoming mail server is the name or IP address of your Linux box
from section 3.2, and the username and password should be your username and
password on the Linux box. (As usual, it's a bad idea to use the root
account on Linux for this purpose.)

<p>
Once you've hit "Finish", set this new account to be the default by
highlighting it and clicking on "Set as Default". Outlook should
connect to your IMAP server, and the name of your IMAP server should
appear at the bottom of your folder list. Click on it; you should see
an Inbox folder. (Note that if /var/mail/yourusername doesn't exist on
your Linux box, you won't be able to drag-and-drop any messages into
your INBOX... and the error message will be confusing. However, that's
not what we're going to do.)

<sect1>Copy over all your email
<p>
At this point you can drag and drop entire folders of email from
Outlook onto the IMAP server name. This will copy the email, including
all attachments, to the Linux box. Unfortunately it also immediately
deletes it from Outlook. In order to copy items without deleting them,
right-click on the folder name and select the "Copy" option. For the
destination, pick your Linux server at the bottom of the list.

<p>

However, life isn't quite that simple. Outlook supports folders
containing folders which also contain messages. Linux IMAP servers do
not support that (at least those using Mailbox format); a folder is
either a regular file containing messages, or a directory containing
subdirectories and files. So if you have folders in Outlook with both
messages and subfolders, you can't copy the entire tree over to the
Linux IMAP server. Another incompatibility of the Linux IMAP server is
that you have to tell it in advance if a new folder will contain
subfolders or messages. You do this by appending a slash (<tt>/</tt>)
to the folder name when you create it. This slash will disappear when
the folder is created.

<p>
So, in order to copy a tree of folders to the Linux IMAP server, first
you need to create a replica of the structure of your existing folders
on the Linux IMAP server. While you're doing this, note which of the
existing folders contain both subfolders and messages. You will need
to move these messages elsewhere. Once you have the overall tree
created, then you can copy or move groups of folders to the Linux IMAP
server.

<p>
One final incompatibility to note is that the Linux IMAP server
doesn't allow folders with slashes (/) in their name. You'll need to
rename such folders before copying or moving them.

<p>
On the Linux box, folders appear as files and directories in your home
directory. The format of these files is the usual Unix mail format,
which most Unix/Linux mail tools either use directly or can convert
to/from. Files with attachments will have MIME attachments; there is
also one extra message per folder which is a (useless) header.

<p>
(One piece of data which doesn't get preserved is the original "From "
line, which contains the envelope address of the email. Fortunately
you don't actually need that information.)

<p>

An additional wrinkle applies regarding attachments. Microsoft
sometimes bundles together several attachments into a ms-tnef
attachment; TNEF stands for Transport Neutral Encapsulation
Format. This attachment contains several mime-encoded attachments.  I
was lucky that my mail folders didn't contain any of these -- they
seem to be created when people send you "Rich Text Format" email -- it
contains two alternatives, one in plain text and the other a
TNEF-encapsulated version of the message in rich text plus any
attachments.

<p>

Fortunately, there are some ways to unpack TNEF attachments.
One is the tnef project hosted on sourceforge, the other is
ktnef, which is a part of KDE.

<sect1>Deinstall IMAP from your Linux box
<p>
Once you've transferred all of your email, you will want to deinstall
the IMAP server from your Linux box, for the security reasons
mentioned earlier. This involves the same 2 steps you took to install
the server:
<enum>
<item> Remove the RPMs:
<tscreen><verb>
  rpm -e imap
</verb></tscreen>
</item>
<item> Remove the line in <tt>/etc/inetd.conf</tt> or <tt>/etc/xinetd.d/imap</tt></item>
<item> Restart inetd or xinetd, or reboot.</item>
</enum>

<p>
Voila! You have taken another step towards a Microsoft-free lifestyle.

</article>