Courier-IMAP

For a general introduction and configuration settings for some popular IMAP clients, go and read imap/README(.html).

In this document:

Requirements

Now is the good time to read the FAQ, before you start. The FAQ is located in the file imap/FAQ(.html?).

UPGRADING

Upgrading from Courier-IMAP 1.3.0, and later versions, is a straightforward process. Follow the instructions in the INSTALLATION section, below, to install the new version. The "make install-configure" command automatically preserves the existing system configuration. However, note that new versions of Courier-IMAP will often introduce additional configuration options. After make install-configure a cursory inspection of configuration files in /usr/lib/courier-imap/etc (the default location of the configuration directory) is recommended, in order to identify any new configuration settings that might need adjustment.

Upgrading from Courier-IMAP 1.3.8.2 and earlier

The default configuration options have slightly changed. The default configuration script will now always build the authdaemon module, and build all real authentication modules inside authdaemond. This is true even with the authvchkpw module.

Upgrading from Courier-IMAP 1.2.3 and earlier

Courier-IMAP 1.3.0 introduced a new configuration file format that allows configuration files to be automatically upgraded. Additionally, several existing configuration files have been renamed in order for their names to be consistent with the Courier build:

Courier-IMAP < 1.3  Courier-IMAP 1.3.0
--------            ---------
imapd.config        imapd
imapd-ssl.config    imapd-ssl
pop3d.config        pop3d
pop3d-ssl.config    pop3d-ssl

The NEWS file has a detailed explanation of how configuration files are now installed. Basically, make install now installs configfilename.dist, and make install-configure copies configfilename.dist to configfilename, becoming the actual configuration file. If there is an existing configfilename, the old settings in configfilename which are still valid will be kept in the new configfilename.

This only works as long as both the old and the new configuration files are in the new format, so this will actually take effect with your next upgrade Courier-IMAP. If the previous installed version of Courier-IMAP did not use the new format for configuration files (1.2.3 and earlier), the old configuration file is backed up to configfilename.bak.

The recommended procedure for upgrading from versions 1.2.3 and earlier is as follows:

The recommended upgrade procedure is as follows:

All configuration files are kept in the configuration directory. Nothing else in /usr/lib/courier-imap is configurable. Do not simply overwrite 1.3.0 configuration files with configuration files from the previous version. It's tempting, but don't do it. It may work, but you will lose the automatic upgrade capability for future releases.

Upgrading from Courier-IMAP 1.1 or earlier

Note that Courier-IMAP 1.2 includes a compatible POP3 server, and the installation script will also install a POP3 server on your system. Even though it is installed, you are not required to use it, but you still need to be aware of its existence. If you install the RPM build of Courier-IMAP, you're going to get the POP3 server started at system boot. If you do not need POP3 services, edit both the pop3d.config and pop3d-ssl.config configuration files, and set POP3DSTART and POP3DSSLSTART to NO

Upgrading from Courier-IMAP 1.0 or earlier

If the server is running, manually stop the server before installing the new version.

INSTALLATION

To compile and install the Courier-IMAP server (this is the short version, a longer version follows):

 
$ ./configure [ options, see below ]
$ make
$ make check       # Note - the --enable-workarounds-for-imap-client-bugs
                   # option to configure will result in make check FAILING.
$ su root
# make install     # Or, make install-strip, to strip the executables.
# make install-configure   # Install configuration files.

                   # Start the authdaemond process

NOTE

You MUST run the configure script as normal user, not root. Did you extract the tarball as root? It won't work. Remove the extracted source code. Log in as a normal user. Extract the source code as a normal user, then run configure. You will do everything as a normal user, except for the final step of installing the compiled software.


As mentioned in "Requirements", above, if you are using xBSD, you must use gmake instead of make.


NOTE: The configure script may run as much as 5-10 minutes on slow machines. It may appear that configure is stuck in a loop, but that's an illusion. Courier-IMAP is built from a collection of modular components, each with its own configuration script. The configuration scripts share a lot of common code, leading to an initial impression that the same configuration script is being repeatedly run.

See below for a description of the options to the configure script.

WARNING: set your umask to 022 before running make install or make install-strip.

You should try make install-strip first. Use make install if make install-strip fails.

The configure script accepts certain options, but the defaults should be fine most of the time. make install puts everything in /usr/lib/courier-imap. If the directory /etc/pam.d exists, make install creates /etc/pam.d/imap and /etc/pam.d/pop3, overwriting any existing files. If you have some other IMAP server installed, this means that you will want to save your existing configuration in /etc/pam.d/{imap|pop3}.

"make check" performs some internal sanity checks. If make check fails, something is wrong, and Courier-IMAP may not work for you reliably. Certain options are documented to cause make check to fail, due to different IMAP protocol behavior. If you need to use those options, first compile Courier-IMAP without them, run make check, and if all goes well extract the source code again in a different directory, then build it for the second time using your options.

After installation, you will need to review the files in /usr/lib/courier-imap/etc and make any changes you deem necessary, including:

configure should automatically detect if you use vpopmail, and compile and install the authvchkpw authentication module.

After running make install or make install-strip you will then have to modify your system's startup scripts to run Courier-IMAP when your system boots.

Use the following command to start the Courier-IMAP server:

 
$ /usr/lib/courier-imap/libexec/imapd.rc start

This assumes that Courier-IMAP is installed in /usr/lib/courier-imap. Use the following command to stop Courier-IMAP:

 
$ /usr/lib/courier-imap/libexec/imapd.rc stop

You will have to add these commands to your system startup/shutdown scripts.

IMAP over SSL

To add SSL support you have to install OpenSSL before installing Courier-IMAP. Download OpenSSL from http://www.openssl.org/. Follow the instruction in OpenSSL package to install it and configure it. SSL support in Courier-IMAP has been tested with OpenSSL 0.9.5a.

The /usr/lib/courier-imap/lib/imapd-ssl configuration file sets some additional options for SSL support, which you may need to adjust. Consult that configuration file for additional information. Then, you also have to run the /usr/lib/courier-imap/libexec/imapd-ssl.rc script from your system startup and shutdown scripts, just like the /usr/lib/courier-imap/libexec/imapd.rc script. You may accept both SSL and non-SSL connections by running both scripts.

Note that SSL requires a valid, signed, X.509 certificate to be installed where Courier-IMAP expects to find it. The default location for the X.509 certificate, in PEM format, is /usr/lib/courier-imap/share/imapd.pem. The X.509 certificate must be signed by a certificate authority that is known to the IMAP client. You can generate your own self-signed certificate by running the script /usr/lib/courier-imap/share/mkimapdcert which will work too, except that IMAP clients using SSL will display a warning message the first time they connect to the server. To get rid of the warning message you'll have to pay for a signed X.509 certificate. The gory details of setting up SSL is beyond the scope of this document, and you should consult the OpenSSL documentation for more information.

The mkimapdcert script will not overwrite an existing imapd.pem certificate, in order to allow precompiled packages to simply call mkimapdcert after installation, without worry.

The bundled POP3 server

The POP3 server included with Courier-IMAP provides POP3 access to INBOX, and that's about it. Enabling the POP3 server is very similar to enabling the IMAP server, with the following differences:

The configuration files are /usr/lib/courier-imap/etc/pop3dand /usr/lib/courier-imap/etc/pop3d-ssl.

The startup/shutdown scripts are /usr/lib/courier-imap/libexec/pop3d.rcand /usr/lib/courier-imap/libexec/pop3d-ssl.rc.

The SSL certificate is /usr/lib/courier-imap/share/pop3d.pem, and the /usr/lib/courier-imap/share/mkpop3dcert script can be used to create a self-signed SSL certificate for testing purposes.

System-V style startup

If your system uses System-V style startup scripts, take a look at courier-imap.sysvinit - this is a sample /etc/init.d script. courier-imap.sysvinit is created by configure. In most cases it can be merely copied to /etc/init.d and /etc/rc?.d directories (with the execute permission bit turned on).

The sample startup script will check if IMAP or POP3 over SSL is enabled. The sample startup script automatically creates dummy SSL certificates the first time it is executed.

Options to configure:

Foreign character set sorting/searching

The Courier-IMAP server can search and sort messages using other than the default us-ascii/iso-8859-1 character set. You can find the list of available character sets in the file unicode/charsetlist.txt.

The default is to include only the ISO-8859-1/US-ASCII character set. Use the --enable-unicode option to include all available character sets.

It is also possible to include translation tables only for selected character sets. Example:

--enable-unicode=iso-8859-1,utf-8,iso-8859-10

Technically, IMAP servers must support the UTF-8 character set, however few IMAP clients (I've yet to see one, actually) care about UTF-8, so the UTF-8 character set is optional in Courier-IMAP. The only required character set - which is always included, explicitly or implicitly - is ISO-8859-1/US-ASCII.

Note that character set translation tables need substantial memory. This should not be a problem in most cases. Most compilers will place the read-only character set tables into a shared text segment, that's shared by all running servers. --enable-unicode should not really be much of a burden for most modern operating systems.

Attentive individuals will observe that all character set tables are compiled even without the --enable-unicode option. That is normal -- only the explicitly selected character set tables will actually make it into the final executable.

Installation directories

Unless the options --prefix, --bindir, or --mandir are used, everything will be installed in the directory /usr/lib/courier-imap.

Use the --prefix option to specify a different directory. This directory will have the following subdirectories:

Having everything installed underneath one directory allows its contents to be easily backed up, before a newer version of courier-imap is installed. Reverting to a previous version is as simple as restoring from backup.

Because some binaries in bin and sbin may be executed from the command line, it will be necessary to change your systemwide global startup script to add this directory to the default PATH. Additionally, it will also be necessary to modify the configuration of the man(1) command so that it can find Courier-IMAP's manual pages in this directory:

        PATH="/usr/lib/courier-imap/bin:$PATH"
        if test -w /etc
        then
                PATH="/usr/lib/courier-imap/sbin:$PATH"
        fi
        export PATH
        MANPATH="/usr/lib/courier-imap/man:$MANPATH"
        export MANPATH

As an alternative, you may use the --bindir and --mandir options in order to install binaries to /usr/local/bin and the manual pages to /usr/local/man, which should already be searched by default:

        ./configure --bindir=/usr/local/bin --mandir=/usr/local/man

Other familiar configure options, such as --sysconfdir and --datadir work too, for those who know how to properly use them.

AUTHENTICATION MODULES - GENERAL NOTES

An authentication module does a few things besides checking if a userid and password are valid. It's job also includes specifying the location of the primary maildir, and its system user and group id.

There are several authentication modules available. Each authentication module implements a different way of authenticating logins, and not all authentication modules can be used by everyone. Some authentication modules can be used only on systems that have certain libraries or software installed separately.

The configure script checks if the required libraries and software are available for each authentication module. If the required libraries and support files are available, the configure script includes the corresponding authentication module. The option --without-module instructs configure not to compile an authentication module, even if it could. Multiple --without options are, of course, allowed (each one specifying a different authentication module).

It is possible to include more than one authentication module. For example, if you select both authuserdb and authpam, each login will first be authenticated against /etc/userdb. If Courier-IMAP cannot find the account in /etc/userdb, it will then attempt to authenticate using the PAM library.

configure uses the following logic to determine which authentication modules will be included by default:

Confirming selected authentication options

You can find out what authentication modules were actually used, but you must first run make successfully. Running make compiles the authinfo program in the authlib subdirectory. This program tells you what authentication modules were compiled.

USERDB AUTHENTICATION MODULE

userdb is a way to implement many virtual mailboxes - many mailboxes that do not have to have a separate system userid allocated for each one, and there is no system login associated with each mailbox. userdb uses a database for mapping virtual addresses to physical maildirs. It should be scalable to thousands of mailboxes. It can also be used to replace linear searches of /etc/passwd with a database lookup, see pw2userdb(8).

Note - you still MUST use some valid system userid and groupid that is shared by all virtual mailboxes. Instead of allocating a single userid and groupid per each mailbox, the same userid and groupid is used for all of them.

This is a rough overview of using userdb. For additional information, read makeuserdb(8), and userdb(8). All the scripts will be installed in /usr/lib/courier-imap, so look for them there.

Courier-IMAP can use the userdb database simply by the virtue of installing the authuserdbauthentication module, which happens by default. The tricky part is creating the database. Hopefully, it's only tricky once, and the second time it won't be so tricky any more. The userdb configuration is extremely flexible, and there are many ways of actually setting up the database.

The best way to describe how userdb works is to try to create one virtual mail account. As mentioned before, virtual mailboxes still need one system account to be used for uid/gid purposes. Let's call this system account "vmail".

Simple userdb setup.

This approach should be used if you do not have many virtual mailboxes. It's very simple, but quickly becomes cumbersome if you administer many virtual mailboxes.

Create an empty /etc/userdb.

 
# cp /dev/null /etc/userdb
# chmod 700 /etc/userdb

/etc/userdb must have 700 permissions.

Now, run the script pw2userdb. This script converts the contents of your /etc/passwd to the /etc/userdb format (including the contents of /etc/shadow, this is why permissions on /etc/userdb must be 700). This script is usually used where you want to convert a very large /etc/passwd to /etc/userdb. userdb applications can now use a fast userdb database instead of a linear scan of /etc/passwd in order to look up system accounts. However, you probably don't want to do use this feature right now, so what you simply want to do is take the output of pw2userdb, and find the entry for the vmail account that you created earlier. Look for a line that starts with 'vmail' followed by tab, followed by familiar fields from /etc/passwd. Save the output of pw2userdb in a temporary file, edit it, and remove everything except the line containing vmail, and the very next line, which is a special entry that maps vmail's userid back to the vmail record.

Here's what you might find in the output of pw2userdb:

 
vmail   uid=1012|gid=1012|home=/home/vmail|systempw=*
1012=   vmail

The actual numerical values and the home directory location may vary. Save these two lines as /etc/userdb, and set the permissions on /etc/userdb to 700:

 
$ chmod 700 /etc/userdb

Now, with that out of the way, let's really create a virtual account. In this example we'll create a virtual mailbox for 'john@example.com'.

 
# su vmail
$ cd $HOME
$ maildirmake Maildir-john-example
$ exit
#

You may need to specify a full path to your maildirmake program. The end result is that you created $HOME/Maildir-john-example in vmail's account.

The next step is to configure your mail server to deliver mail for john@example.com directly into the maildir ~vmail/Maildir-john-example. How that's done depends on the mail server that you use.

Now, let's connect the dots here, and create an entry in /etc/userdb for john@example.com:

 
# userdb "john@example.com" set home=/home/vmail \
                                mail=/home/vmail/Maildir-john-example \
                                uid=UUU gid=GGG

This command runs a Perl script named userdb , which is installed, by default in /usr/lib/courier-imap/sbin. Replace UUU and GGG with the userid and groupid of the vmail account. If you now look in /etc/userdb, you will see that a new record for john@example.com has been appended to the end of the file.

We're not finished yet. We need to set the IMAP password for this mailbox:

 
# userdbpw | userdb "john@example.com" set imappw

If you are running Red Hat Linux 6.0, or higher, you can specify the -md5 option to userdbpw, in order to use an MD5 password hash, instead of crypt. Specify "systempw" instead of "imappw" if you would like to use the same password for the POP3 server too. The imappw field is only checked by the IMAP server. If not defined, systempw is used instead. The field pop3pw will be checked by the POP3 server that's bundled with Courier-IMAP. If it is not defined the POP3 server will check systempw too

Finally, compile the database:

 
# makeuserdb

This command creates the actual database, /etc/userdb.dat and /etc/userdbshadow.dat from the plain text file /etc/userdb. Courier-IMAP will now start accepting logins to this mailbox. Adding and removing mailboxes can be done while Courier-IMAP is running.

Courier-IMAP reads /etc/userdb.dat and /etc/userdbshadow.dat only. The plain text source, /etc/userdb is not read by Courier-IMAP itself. Changes take effect only when makeuserdb runs.

Large virtual domain farm.

The previous approach uses a single flat file, /etc/userdb. This might be a bit cumbersome if you have thousands of virtual mailboxes in many domains. Here's an alternative approach that can scale to thousands of domains and mailboxes.

Instead of creating a /etc/userdb file, create a subdirectory:

 
# mkdir /etc/userdb
# chmod 700 /etc/userdb

Now, create /etc/userdb/default, containing pw2userdb's output for the vmail account, as previously described.

This time, you probably want to create all mailboxes for the same domain in a separate subdirectory:

 
# su vmail
$ cd $HOME
$ mkdir -p domains/example-com
$ maildirmake domains/example-com/john
$ exit

The idea is that all the maildirs for @example.com will be stored in ~vmail/domains/example-com. All maildirs for domain.org will be stored in ~vmail/domains/domain.org. The actual layout and naming conventions are entirely up to you to define.

Now, configure your mail server to deliver mail for john@example.com into this maildir.

Here's how configure /etc/userdb.

 
$ userdb "example-com/john@example.com" set home=/home/vmail \
                                mail=/home/vmail/domains/example-com/john \
                                uid=UUU gid=GGG

This creates the file /etc/userdb/example-com (the first parameter to the userdb command), and appends a record named "john@example.com". You will store all userdb entries for @example.com in the file /etc/userdb/example-com. All entries for @domain.org will be maintained in /etc/userdb/domain-org, and so on.

 
$ userdbpw | userdb "example-com/john@example.com" set imappw

This sets the IMAP access password for this account. Finally:

 
$ makeuserdb

VCHKPW AUTHENTICATION MODULE

This authentication module is provided for legacy support. New installs should use the authuserdb module. It is possible to convert vpopmail/vchkpw single userid authentication passwd files to /etc/userdb using the vchkpw2userdb script. See vchkpw2userdb(8) for more information.

LDAP AUTHENTICATION MODULE

This module attempts to authenticate against an LDAP server. This option installs a sample configuration file, which is /usr/lib/courier-imap/etc/authldaprc by default. You will need to edit this file in order to configure LDAP authentication. This configuration file defines the location of your LDAP server, as well as the names of attributes used to perform LDAP authentication. See the comments in the sample configuration file, and the authldap(7) manual page, for more information.

DAEMON AUTHENTICATION MODULE

Selecting the authdaemon module (it is selected by default) will compile and install a separate process, called "authdaemond". The "authdaemond" process is started and stopped by the imapd.rc, imapd-ssl.rc, pop3d.rc and pop3d-ssl.rc. ATTENTION: all four scripts stop authdaemond. Typically all the scripts are called together, at system startup and shutdown. If you need to stop just ONE service, find the pid and kill it manually. Otherwise authdaemond stops and nobody can log in anymore.

The authdaemon module (and the authdaemond process that goes with it) is very useful when a database is used as an authentication back-end (like MySQL, PostgreSQL, or OpenLDAP). The authdaemon module offers an alternative to compiling all the authentication code as standalone modules. Enabling authdaemon in addition to any other modules will result in authdaemon being built as the only "official" authentication module. All other modules are compiled into a separate process, "authdaemond".

Certain authentication modules -- such as authldap, authpgsql and authmysql -- connect to an external database in order to validate an authentication request. The database connection is created and destroyed for every login request, and with a large number of connections and authentications it is better to have a single daemon process running in a background, with a semi-permanent connection to the database, handling authentication requests. That's exactly what authdaemond does.

The authdaemon authentication module takes an authentication request, and forwards it to the permanently running authdaemond process, waits for the answer, and returns the result to the application. Courier-IMAP, by the virtue of its modular design, "sees" only the authdaemon authentication module. Behind the scenes, authdaemon takes every authentication request and passes it along to the authdaemond for processing, where all the real authentication happens.

The /usr/lib/courier-imap/etc/authdaemonrc configuration file sets several parameters for the authdaemond process. See the comments in this file for more information. Currently, authdaemonrc sets two parameters: number of daemon processes, and which available authentication modules are used.

Although authdaemond might be built with several authentication modules, not all of them must be used. This allows for a single authdaemond binary to be made that gets installed on multiple systems with different authentication needs. The default module list specified in authdaemonrc is a list of all the available authentication modules.

All available authentication modules are compiled by default. Although it is possible to disable unwanted authentication modules via the configure script, it is better to disable them at runtime. The list of all active authentication modules is read from the authdaemonrc configuration file. Removing an authentication module from this configuration file disables it.

The number of authdaemond processes is also set in this configuration file. The more processes that are started, the more authentication requests can be handled. If authdaemon does not receive an answer within some amount of time, it will assume an authentication failure, and abort. Try increasing the number of processes if you start seeing random authentication failures. However, that should only be used as a stop-gap measure. If the default number of authdaemond processes proves to be insufficient, it is far more likely that more resources are needed for the server (more RAM, a faster disk, or a faster CPU) in the humble opinion of the author. Increasing the number of processes should only be used as a stop-gap measure, until a more thorough analysis of the system performance.

After making any changes to authdaemonrc, run the following command for these changes to take effect:

    /usr/lib/courier-imap/libexec/authlib/authdaemond restart

Alternative authdaemond modules

/usr/lib/courier-imap/libexec/authlib/authdaemond is actually a script. Depending on your system configuration, there may be one or more different authdaemond binaries installed, and here's why.

authdaemond.plain will include all available authentication modules except for certain "heavy" authentication modules. The current list of "heavy" authentication modules is authldap, authpgsql, and authmysql.

If support for a "heavy" authentication module is selected, there will be an additional binary installed, such as authdaemond.mysql, authdaemond.pgsql or authdaemond.ldap. The /usr/lib/courier-imap/libexec/authlib/authdaemond script checks if any "heavy" authentication daemon is installed, and, if so, runs that. Otherwise, the default authdaemond.plain binary goes in.

This allows an easy way to create binary Courier-IMAP distributions with and without LDAP, MySQL, or PostgreSQL support. The distributor would simply build Courier-IMAP on a machine that contains both LDAP/MySQL/PostgreSQL development libraries, then take everything but authdaemond.mysql, authdaemond.pgsql and authdaemond.ldap and roll it into the base Courier-IMAP package. authdaemond.mysql, authdaemond.pgsql, and authdaemond.ldap are rolled into separate sub-packages. Loading a base package installs basic system authentication services. Adding LDAP, MySQL, or PostgreSQL support is as simple as loading the corresponding sub-package.

USING SHARED FOLDERS

Courier-IMAP supports shared folders. See the file README.sharedfolders.html for information on how to set up shared folders.

CRAM-MD5 AUTHENTICATION

CRAM-MD5 authentication allows IMAP clients to authenticate themselves without sending the password in clear-text over the network. Courier-IMAP now supports CRAM-MD5 by default, but is not enabled for reasons explained below. CRAM-MD5 support is implemented by the authcram module, with one exception - authldap, authpgsql, and authmysql support CRAM-MD5 authentication if the LDAP or the MySQL/PostgreSQL server stores clear-text passwords, and not crypt-ed passwords.

To use CRAM-MD5 it is necessary to use an IMAP client that support CRAM-MD5 authentication, of course. That's the easy part.

The problem is that it is not possible to use the system password when logging in using CRAM-MD5. That's because CRAM-MD5 requires the knowledge of the actual password, in the clear, in order to calculate authentication tokens (even though that the password itself is not sent in the clear over the network).

So, implementation of CRAM-MD5 is an advanced task that should be attempted only when you are comfortable with, and fully understand how Courier-IMAP works in general. Here's an overview of this procedure:

Enabling CRAM-MD5 authentication

Because of these unfortunate complexities, CRAM-MD5 authentication is disabled after installation. When you're ready to use CRAM-MD5, edit the imapd configuration file and add the "AUTH=CRAM-MD5" keyword to the IMAP_CAPABILITY environment variable, then restart Courier-IMAP. There are instructions in the imapd configuration file to that effect.

If you do not intend to ever use CRAM-MD5 authentication, you can either specify --without-authcram option to the configure script, or simply edit imapd and remove authcram from the AUTHMODULES setting.

SENDING MAIL VIA AN IMAP CONNECTION

This server allows using the IMAP connection to send E-mail. Normally, the IMAP protocol provides only access to mail in an existing mail account, and mail clients must use SMTP in order to send mail. The Courier-IMAP server has an optional setting to enable mail to be send via an IMAP connection in a manner that should work with all existing IMAP mail clients. This can be useful when an account is logged in from a shared access pool which normally blocks most access to the SMTP port.

This is implemented by enabling a setting in the imapd configuration file that designates a folder as a special "Outbox" folder. The default setting is a folder called "Outbox" (IMAP path INBOX.Outbox), but the name can be changed to anything. This folder, for the most part, is no different than any other folder. If a folder by that name doesn't exist, it needs to be created, just like any other IMAP folder. It looks and acts like any other folder, except that each message added to the folder, via IMAP's APPEND or COPY command, will also be mailed out by the Courier-IMAP server to the addresses listed in the To:, Cc:, and Bcc: headers.

It should be possible to use this to send mail from any IMAP client by:

  1. Composing a draft message, telling the IMAP client to save the draft message in its drafts folder on the IMAP server.

  2. Opening the drafts folder, and moving or copying the message to the Outbox folder.

  3. The act of copying the message into the Outbox folder will send the mail. There won't be any explicit notification to the fact that the message was sent, so it's a good idea to include your own E-mail address on the Cc: list.

NOTE: it is tempting to configure the IMAP mail client to use Outbox as its default folder for saving drafts. Resist the temptation. If you forget, you'll save a partially completed draft, which will be then obediently mailed out.

NOTE: the message, in addition to being sent, will be saved in the folder in the normal fashion. After saving the message, reopen the Outbox folder and delete the sent message, or move it someplace else.

NOTE: when enabled, the Courier-IMAP server will advertize a private XCOURIEROUTBOX IMAP capability. It is theoretically possible to code an IMAP mail client that reads this capability and automatically configures itself accordingly -- when this IMAP capability is present -- to send E-mail in the normal way but using the IMAP connection. At this time, I'm not aware of any actual mail clients that know how to do this.

NOTE: many mail clients save some additional internal information in headers of draft messages. The internal information is normally removed before the mail client sends the message. Make sure that none of this extra information is something that should not be mailed out.

REALTIME FOLDER STATUS UPDATES

If the FAM, the File Alteration Monitor (http://oss.sgi.com/projects/fam/) is installed it will be possible to allow multiple clients to open the same folder, and have all clients to be simultaneously notified of any changes to the folder contents.

After installing the server see the imapd(8) manual page for more information.