Exim Internet Mailer

1. Scanning for viruses

The malware ACL condition lets you connect virus scanner software to Exim. It supports a “generic” interface to scanners called via the shell, and specialized interfaces for “daemon” type virus scanners, which are resident in memory and thus are much faster.

Since message data needs to have arrived, the condition may be only called in ACL defined by acl_smtp_data, acl_smtp_data_prdr, acl_smtp_mime or acl_smtp_dkim

A timeout of 2 minutes is applied to a scanner call (by default); if it expires then a defer action is taken.

You can set the av_scanner option in the main part of the configuration to specify which scanner to use, together with any additional options that are needed. The basic syntax is as follows:

av_scanner = <scanner-type>:<option1>:<option2>:[...]

If you do not set av_scanner, it defaults to

av_scanner = sophie:/var/run/sophie

If the value of av_scanner starts with a dollar character, it is expanded before use. The usual list-parsing of the content (see 6.20) applies. The following scanner types are supported in this release, though individual ones can be included or not at build time:

avast

This is the scanner daemon of Avast. It has been tested with Avast Core Security (currently at version 2.2.0). You can get a trial version at https://www.avast.com or for Linux at https://www.avast.com/linux-server-antivirus. This scanner type takes one option, which can be either a full path to a UNIX socket, or host and port specifiers separated by white space. The host may be a name or an IP address; the port is either a single number or a pair of numbers with a dash between. A list of options may follow. These options are interpreted on the Exim’s side of the malware scanner, or are given on separate lines to the daemon as options before the main scan command.

If pass_unscanned is set, any files the Avast scanner can’t scan (e.g. decompression bombs, or invalid archives) are considered clean. Use with care.

For example:

av_scanner = avast:/var/run/avast/scan.sock:FLAGS -fullfiles:SENSITIVITY -pup
av_scanner = avast:/var/run/avast/scan.sock:pass_unscanned:FLAGS -fullfiles:SENSITIVITY -pup
av_scanner = avast:192.168.2.22 5036

If you omit the argument, the default path /var/run/avast/scan.sock is used. If you use a remote host, you need to make Exim’s spool directory available to it, as the scanner is passed a file path, not file contents. For information about available commands and their options you may use

$ socat UNIX:/var/run/avast/scan.sock STDIO:
    FLAGS
    SENSITIVITY
    PACK

If the scanner returns a temporary failure (e.g. license issues, or permission problems), the message is deferred and a paniclog entry is written. The usual defer_ok option is available.

aveserver

This is the scanner daemon of Kaspersky Version 5. You can get a trial version at https://www.kaspersky.com/. This scanner type takes one option, which is the path to the daemon’s UNIX socket. The default is shown in this example:

av_scanner = aveserver:/var/run/aveserver

clamd

This daemon-type scanner is GPL and free. You can get it at https://www.clamav.net/. Some older versions of clamd do not seem to unpack MIME containers, so it used to be recommended to unpack MIME attachments in the MIME ACL. This is no longer believed to be necessary.

The options are a list of server specifiers, which may be a UNIX socket specification, a TCP socket specification, or a (global) option.

A socket specification consists of a space-separated list. For a Unix socket the first element is a full path for the socket, for a TCP socket the first element is the IP address and the second a port number, Any further elements are per-server (non-global) options. These per-server options are supported:

retry=<timespec>	Retry on connect fail

The retry option specifies a time after which a single retry for a failed connect is made. The default is to not retry.

If a Unix socket file is specified, only one server is supported.

Examples:

av_scanner = clamd:/opt/clamd/socket
av_scanner = clamd:192.0.2.3 1234
av_scanner = clamd:192.0.2.3 1234:local
av_scanner = clamd:192.0.2.3 1234 retry=10s
av_scanner = clamd:192.0.2.3 1234 : 192.0.2.4 1234

If the value of av_scanner points to a UNIX socket file or contains the local option, then the ClamAV interface will pass a filename containing the data to be scanned, which should normally result in less I/O happening and be more efficient. Normally in the TCP case, the data is streamed to ClamAV as Exim does not assume that there is a common filesystem with the remote host.

The final example shows that multiple TCP targets can be specified. Exim will randomly use one for each incoming email (i.e. it load balances them). Note that only TCP targets may be used if specifying a list of scanners; a UNIX socket cannot be mixed in with TCP targets. If one of the servers becomes unavailable, Exim will try the remaining one(s) until it finds one that works. When a clamd server becomes unreachable, Exim will log a message. Exim does not keep track of scanner state between multiple messages, and the scanner selection is random, so the message will get logged in the mainlog for each email that the down scanner gets chosen first (message wrapped to be readable):

2013-10-09 14:30:39 1VTumd-0000Y8-BQ malware acl condition:
   clamd: connection to localhost, port 3310 failed
   (Connection refused)

If the option is unset, the default is /tmp/clamd. Thanks to David Saez for contributing the code for this scanner.

cmdline

This is the keyword for the generic command line scanner interface. It can be used to attach virus scanners that are invoked from the shell. This scanner type takes 3 mandatory options:

1. The full path and name of the scanner binary, with all command line options, and a placeholder (%s) for the directory to scan.

2. A regular expression to match against the STDOUT and STDERR output of the virus scanner. If the expression matches, a virus was found. You must make absolutely sure that this expression matches on “virus found”. This is called the “trigger” expression.

3. Another regular expression, containing exactly one pair of parentheses, to match the name of the virus found in the scanners output. This is called the “name” expression.

For example, Sophos Sweep reports a virus on a line like this:

Virus 'W32/Magistr-B' found in file ./those.bat

For the trigger expression, we can match the phrase “found in file”. For the name expression, we want to extract the W32/Magistr-B string, so we can match for the single quotes left and right of it. Altogether, this makes the configuration setting:

av_scanner = cmdline:\
             /path/to/sweep -ss -all -rec -archive %s:\
             found in file:'(.+)'

drweb

The DrWeb daemon scanner (https://www.sald.ru/) interface takes one option, either a full path to a UNIX socket, or host and port specifiers separated by white space. The host may be a name or an IP address; the port is either a single number or a pair of numbers with a dash between. For example:

av_scanner = drweb:/var/run/drwebd.sock
av_scanner = drweb:192.168.2.20 31337

If you omit the argument, the default path /usr/local/drweb/run/drwebd.sock is used. Thanks to Alex Miller for contributing the code for this scanner.

f-protd

The f-protd scanner is accessed via HTTP over TCP. One argument is taken, being a space-separated hostname and port number (or port-range). For example:

av_scanner = f-protd:localhost 10200-10204

If you omit the argument, the default values shown above are used.

f-prot6d

The f-prot6d scanner is accessed using the FPSCAND protocol over TCP. One argument is taken, being a space-separated hostname and port number. For example:

av_scanner = f-prot6d:localhost 10200

If you omit the argument, the default values show above are used.

fsecure

The F-Secure daemon scanner (https://www.f-secure.com/) takes one argument which is the path to a UNIX socket. For example:

av_scanner = fsecure:/path/to/.fsav

If no argument is given, the default is /var/run/.fsav. Thanks to Johan Thelmen for contributing the code for this scanner.

kavdaemon

This is the scanner daemon of Kaspersky Version 4. This version of the Kaspersky scanner is outdated. Please upgrade (see aveserver above). This scanner type takes one option, which is the path to the daemon’s UNIX socket. For example:

av_scanner = kavdaemon:/opt/AVP/AvpCtl

The default path is /var/run/AvpCtl.

mksd

This was a daemon type scanner that is aimed mainly at Polish users, though some documentation was available in English. The history can be shown at https://en.wikipedia.org/wiki/Mks_vir and this appears to be a candidate for removal from Exim, unless we are informed of other virus scanners which use the same protocol to integrate. The only option for this scanner type is the maximum number of processes used simultaneously to scan the attachments, provided that mksd has been run with at least the same number of child processes. For example:

av_scanner = mksd:2

You can safely omit this option (the default value is 1).

sock

This is a general-purpose way of talking to simple scanner daemons running on the local machine. There are four options: an address (which may be an IP address and port, or the path of a Unix socket), a commandline to send (may include a single %s which will be replaced with the path to the mail file to be scanned), an RE to trigger on from the returned data, and an RE to extract malware_name from the returned data. For example:

av_scanner = sock:127.0.0.1 6001:%s:(SPAM|VIRUS):(.*)$

Note that surrounding whitespace is stripped from each option, meaning there is no way to specify a trailing newline. The socket specifier and both regular-expressions are required. Default for the commandline is %s\n (note this does have a trailing newline); specify an empty element to get this.

sophie

Sophie is a daemon that uses Sophos’ libsavi library to scan for viruses. You can get Sophie at http://sophie.sourceforge.net/. The only option for this scanner type is the path to the UNIX socket that Sophie uses for client communication. For example:

av_scanner = sophie:/tmp/sophie

The default path is /var/run/sophie, so if you are using this, you can omit the option.

When av_scanner is correctly set, you can use the malware condition in the DATA ACL. Note: You cannot use the malware condition in the MIME ACL.

The av_scanner option is expanded each time malware is called. This makes it possible to use different scanners. See further below for an example. The malware condition caches its results, so when you use it multiple times for the same message, the actual scanning process is only carried out once. However, using expandable items in av_scanner disables this caching, in which case each use of the malware condition causes a new scan of the message.

The malware condition takes a right-hand argument that is expanded before use and taken as a list, slash-separated by default. The first element can then be one of

• “true”, “*”, or “1”, in which case the message is scanned for viruses. The condition succeeds if a virus was found, and fail otherwise. This is the recommended usage.

• “false” or “0” or an empty string, in which case no scanning is done and the condition fails immediately.

• A regular expression, in which case the message is scanned for viruses. The condition succeeds if a virus is found and its name matches the regular expression. This allows you to take special actions on certain types of virus. Note that “/” characters in the RE must be doubled due to the list-processing, unless the separator is changed (in the usual way 6.21).

You can append a defer_ok element to the malware argument list to accept messages even if there is a problem with the virus scanner. Otherwise, such a problem causes the ACL to defer.

You can append a tmo=<val> element to the malware argument list to specify a non-default timeout. The default is two minutes. For example:

malware = * / defer_ok / tmo=10s

A timeout causes the ACL to defer.

When a connection is made to the scanner the expansion variable $callout_address is set to record the actual address used.

When a virus is found, the condition sets up an expansion variable called $malware_name that contains the name of the virus. You can use it in a message modifier that specifies the error returned to the sender, and/or in logging data.

Beware the interaction of Exim’s message_size_limit with any size limits imposed by your anti-virus scanner.

Here is a very simple scanning example:

deny message = This message contains malware ($malware_name)
     malware = *

The next example accepts messages when there is a problem with the scanner:

deny message = This message contains malware ($malware_name)
     malware = */defer_ok

The next example shows how to use an ACL variable to scan with both sophie and aveserver. It assumes you have set:

av_scanner = $acl_m0

in the main Exim configuration.

deny message = This message contains malware ($malware_name)
     set acl_m0 = sophie
     malware = *

deny message = This message contains malware ($malware_name)
     set acl_m0 = aveserver
     malware = *

2. Scanning with SpamAssassin and Rspamd

The spam ACL condition calls SpamAssassin’s spamd daemon to get a spam score and a report for the message. Support is also provided for Rspamd.

For more information about installation and configuration of SpamAssassin or Rspamd refer to their respective websites at https://spamassassin.apache.org/ and https://www.rspamd.com/

SpamAssassin can be installed with CPAN by running:

perl -MCPAN -e 'install Mail::SpamAssassin'

SpamAssassin has its own set of configuration files. Please review its documentation to see how you can tweak it. The default installation should work nicely, however.

By default, SpamAssassin listens on 127.0.0.1, TCP port 783 and if you intend to use an instance running on the local host you do not need to set spamd_address. If you intend to use another host or port for SpamAssassin, you must set the spamd_address option in the global part of the Exim configuration as follows (example):

spamd_address = 192.168.99.45 783

The SpamAssassin protocol relies on a TCP half-close from the client. If your SpamAssassin client side is running a Linux system with an iptables firewall, consider setting net.netfilter.nf_conntrack_tcp_timeout_close_wait to at least the timeout, Exim uses when waiting for a response from the SpamAssassin server (currently defaulting to 120s). With a lower value the Linux connection tracking may consider your half-closed connection as dead too soon.

To use Rspamd (which by default listens on all local addresses on TCP port 11333) you should add variant=rspamd after the address/port pair, for example:

spamd_address = 127.0.0.1 11333 variant=rspamd

As of version 2.60, SpamAssassin also supports communication over UNIX sockets. If you want to us these, supply spamd_address with an absolute filename instead of an address/port pair:

spamd_address = /var/run/spamd_socket

You can have multiple spamd servers to improve scalability. These can reside on other hardware reachable over the network. To specify multiple spamd servers, put multiple address/port pairs in the spamd_address option, separated with colons (the separator can be changed in the usual way 6.21):

spamd_address = 192.168.2.10 783 : \
                192.168.2.11 783 : \
                192.168.2.12 783

Up to 32 spamd servers are supported. When a server fails to respond to the connection attempt, all other servers are tried until one succeeds. If no server responds, the spam condition defers.

Unix and TCP socket specifications may be mixed in any order. Each element of the list is a list itself, space-separated by default and changeable in the usual way (6.21); take care to not double the separator.

For TCP socket specifications a host name or IP (v4 or v6, but subject to list-separator quoting rules) address can be used, and the port can be one or a dash-separated pair. In the latter case, the range is tried in strict order.

Elements after the first for Unix sockets, or second for TCP socket, are options. The supported options are:

pri=<priority>      Selection priority
weight=<value>      Selection bias
time=<start>-<end>  Use only between these times of day
retry=<timespec>    Retry on connect fail
tmo=<timespec>      Connection time limit
variant=rspamd      Use Rspamd rather than SpamAssassin protocol

The pri option specifies a priority for the server within the list, higher values being tried first. The default priority is 1.

The weight option specifies a selection bias. Within a priority set servers are queried in a random fashion, weighted by this value. The default value for selection bias is 1.

Time specifications for the time option are <hour>.<minute>.<second> in the local time zone; each element being one or more digits. Either the seconds or both minutes and seconds, plus the leading . characters, may be omitted and will be taken as zero.

Timeout specifications for the retry and tmo options are the usual Exim time interval standard, e.g. 20s or 1m.

The tmo option specifies an overall timeout for communication. The default value is two minutes.

The retry option specifies a time after which a single retry for a failed connect is made. The default is to not retry.

The spamd_address variable is expanded before use if it starts with a dollar sign. In this case, the expansion may return a string that is used as the list so that multiple spamd servers can be the result of an expansion.

When a connection is made to the server the expansion variable $callout_address is set to record the actual address used.

3. Calling SpamAssassin from an Exim ACL

Here is a simple example of the use of the spam condition in a DATA ACL:

deny message = This message was classified as SPAM
     spam = joe

The right-hand side of the spam condition specifies a name. This is relevant if you have set up multiple SpamAssassin profiles. If you do not want to scan using a specific profile, but rather use the SpamAssassin system-wide default profile, you can scan for an unknown name, or simply use “nobody”. Rspamd does not use this setting. However, you must put something on the right-hand side.

The name allows you to use per-domain or per-user antispam profiles in principle, but this is not straightforward in practice, because a message may have multiple recipients, not necessarily all in the same domain. Because the spam condition has to be called from a DATA-time ACL in order to be able to read the contents of the message, the variables $local_part and $domain are not set. Careful enforcement of single-recipient messages (e.g. by responding with defer in the recipient ACL for all recipients after the first), or the use of PRDR, are needed to use this feature.

The right-hand side of the spam condition is expanded before being used, so you can put lookups or conditions there. When the right-hand side evaluates to “0” or “false”, no scanning is done and the condition fails immediately.

Scanning with SpamAssassin uses a lot of resources. If you scan every message, large ones may cause significant performance degradation. As most spam messages are quite small, it is recommended that you do not scan the big ones. For example:

deny message = This message was classified as SPAM
     condition = ${if < {$message_size}{10K}}
     spam = nobody

The spam condition returns true if the threshold specified in the user’s SpamAssassin profile has been matched or exceeded. If you want to use the spam condition for its side effects (see the variables below), you can make it always return “true” by appending :true to the username.

When the spam condition is run, it sets up a number of expansion variables. Except for $spam_report, these variables are saved with the received message so are available for use at delivery time.

$spam_score

The spam score of the message, for example, “3.4” or “30.5”. This is useful for inclusion in log or reject messages.

$spam_score_int

The spam score of the message, multiplied by ten, as an integer value. For example “34” or “305”. It may appear to disagree with $spam_score because $spam_score is rounded and $spam_score_int is truncated. The integer value is useful for numeric comparisons in conditions.

$spam_bar

A string consisting of a number of “+” or “-” characters, representing the integer part of the spam score value. A spam score of 4.4 would have a $spam_bar value of “++++”. This is useful for inclusion in warning headers, since MUAs can match on such strings. The maximum length of the spam bar is 50 characters.

$spam_report

A multiline text table, containing the full SpamAssassin report for the message. Useful for inclusion in headers or reject messages. This variable is only usable in a DATA-time ACL. Beware that SpamAssassin may return non-ASCII characters, especially when running in country-specific locales, which are not legal unencoded in headers.

$spam_action

For SpamAssassin either ’reject’ or ’no action’ depending on the spam score versus threshold. For Rspamd, the recommended action.

The spam condition caches its results unless expansion in spamd_address was used. If you call it again with the same user name, it does not scan again, but rather returns the same values as before.

The spam condition returns DEFER if there is any error while running the message through SpamAssassin or if the expansion of spamd_address failed. If you want to treat DEFER as FAIL (to pass on to the next ACL statement block), append /defer_ok to the right-hand side of the spam condition, like this:

deny message = This message was classified as SPAM
     spam    = joe/defer_ok

This causes messages to be accepted even if there is a problem with spamd.

Here is a longer, commented example of the use of the spam condition:

# put headers in all messages (no matter if spam or not)
warn  spam = nobody:true
      add_header = X-Spam-Score: $spam_score ($spam_bar)
      add_header = X-Spam-Report: $spam_report

# add second subject line with *SPAM* marker when message
# is over threshold
warn  spam = nobody
      add_header = Subject: *SPAM* $h_Subject:

# reject spam at high scores (> 12)
deny  message = This message scored $spam_score spam points.
      spam = nobody:true
      condition = ${if >{$spam_score_int}{120}{1}{0}}

4. Scanning MIME parts

The acl_smtp_mime global option specifies an ACL that is called once for each MIME part of an SMTP message, including multipart types, in the sequence of their position in the message. Similarly, the acl_not_smtp_mime option specifies an ACL that is used for the MIME parts of non-SMTP messages. These options may both refer to the same ACL if you want the same processing in both cases.

These ACLs are called (possibly many times) just before the acl_smtp_data ACL in the case of an SMTP message, or just before the acl_not_smtp ACL in the case of a non-SMTP message. However, a MIME ACL is called only if the message contains a Content-Type: header line. When a call to a MIME ACL does not yield “accept”, ACL processing is aborted and the appropriate result code is sent to the client. In the case of an SMTP message, the acl_smtp_data ACL is not called when this happens.

You cannot use the malware or spam conditions in a MIME ACL; these can only be used in the DATA or non-SMTP ACLs. However, you can use the regex condition to match against the raw MIME part. You can also use the mime_regex condition to match against the decoded MIME part (see section 45.5).

At the start of a MIME ACL, a number of variables are set from the header information for the relevant MIME part. These are described below. The contents of the MIME part are not by default decoded into a disk file except for MIME parts whose content-type is “message/rfc822”. If you want to decode a MIME part into a disk file, you can use the decode condition. The general syntax is:

decode = [/<path>/]<filename>

The right hand side is expanded before use. After expansion, the value can be:

1. “0” or “false”, in which case no decoding is done.

2. The string “default”. In that case, the file is put in the temporary “default” directory <spool_directory>/scan/<message_id>/ with a sequential filename consisting of the message id and a sequence number. The full path and name is available in $mime_decoded_filename after decoding.

3. A full path name starting with a slash. If the full name is an existing directory, it is used as a replacement for the default directory. The filename is then sequentially assigned. If the path does not exist, it is used as the full path and filename.

4. If the string does not start with a slash, it is used as the filename, and the default path is then used.

The decode condition normally succeeds. It is only false for syntax errors or unusual circumstances such as memory shortages. You can easily decode a file with its original, proposed filename using

decode = $mime_filename

However, you should keep in mind that $mime_filename might contain anything. If you place files outside of the default path, they are not automatically unlinked.

For RFC822 attachments (these are messages attached to messages, with a content-type of “message/rfc822”), the ACL is called again in the same manner as for the primary message, only that the $mime_is_rfc822 expansion variable is set (see below). Attached messages are always decoded to disk before being checked, and the files are unlinked once the check is done.

The MIME ACL supports the regex and mime_regex conditions. These can be used to match regular expressions against raw and decoded MIME parts, respectively. They are described in section 45.5.

The following list describes all expansion variables that are available in the MIME ACL:

$mime_boundary

If the current part is a multipart (see $mime_is_multipart) below, it should have a boundary string, which is stored in this variable. If the current part has no boundary parameter in the Content-Type: header, this variable contains the empty string.

$mime_charset

This variable contains the character set identifier, if one was found in the Content-Type: header. Examples for charset identifiers are:

us-ascii
gb2312 (Chinese)
iso-8859-1

Please note that this value is not normalized, so you should do matches case-insensitively.

$mime_content_description

This variable contains the normalized content of the Content-Description: header. It can contain a human-readable description of the parts content. Some implementations repeat the filename for attachments here, but they are usually only used for display purposes.

$mime_content_disposition

This variable contains the normalized content of the Content-Disposition: header. You can expect strings like “attachment” or “inline” here.

$mime_content_id

This variable contains the normalized content of the Content-ID: header. This is a unique ID that can be used to reference a part from another part.

$mime_content_size

This variable is set only after the decode modifier (see above) has been successfully run. It contains the size of the decoded part in kilobytes. The size is always rounded up to full kilobytes, so only a completely empty part has a $mime_content_size of zero.

$mime_content_transfer_encoding

This variable contains the normalized content of the Content-transfer-encoding: header. This is a symbolic name for an encoding type. Typical values are “base64” and “quoted-printable”.

$mime_content_type

If the MIME part has a Content-Type: header, this variable contains its value, lowercased, and without any options (like “name” or “charset”). Here are some examples of popular MIME types, as they may appear in this variable:

text/plain
text/html
application/octet-stream
image/jpeg
audio/midi

If the MIME part has no Content-Type: header, this variable contains the empty string.

$mime_decoded_filename

This variable is set only after the decode modifier (see above) has been successfully run. It contains the full path and filename of the file containing the decoded data.

$mime_filename

This is perhaps the most important of the MIME variables. It contains a proposed filename for an attachment, if one was found in either the Content-Type: or Content-Disposition: headers. The filename will be RFC2047 or RFC2231 decoded, but no additional sanity checks are done. If no filename was found, this variable contains the empty string.

$mime_is_coverletter

This variable attempts to differentiate the “cover letter” of an e-mail from attached data. It can be used to clamp down on flashy or unnecessarily encoded content in the cover letter, while not restricting attachments at all.

The variable contains 1 (true) for a MIME part believed to be part of the cover letter, and 0 (false) for an attachment. At present, the algorithm is as follows:

1. The outermost MIME part of a message is always a cover letter.

2. If a multipart/alternative or multipart/related MIME part is a cover letter, so are all MIME subparts within that multipart.

3. If any other multipart is a cover letter, the first subpart is a cover letter, and the rest are attachments.

4. All parts contained within an attachment multipart are attachments.

As an example, the following will ban “HTML mail” (including that sent with alternative plain text), while allowing HTML files to be attached. HTML coverletter mail attached to non-HTML coverletter mail will also be allowed:

deny message = HTML mail is not accepted here
!condition = $mime_is_rfc822
condition = $mime_is_coverletter
condition = ${if eq{$mime_content_type}{text/html}{1}{0}}

$mime_is_multipart

This variable has the value 1 (true) when the current part has the main type “multipart”, for example, “multipart/alternative” or “multipart/mixed”. Since multipart entities only serve as containers for other parts, you may not want to carry out specific actions on them.

$mime_is_rfc822

This variable has the value 1 (true) if the current part is not a part of the checked message itself, but part of an attached message. Attached message decoding is fully recursive.

$mime_part_count

This variable is a counter that is raised for each processed MIME part. It starts at zero for the very first part (which is usually a multipart). The counter is per-message, so it is reset when processing RFC822 attachments (see $mime_is_rfc822). The counter stays set after acl_smtp_mime is complete, so you can use it in the DATA ACL to determine the number of MIME parts of a message. For non-MIME messages, this variable contains the value -1.

5. Scanning with regular expressions

You can specify your own custom regular expression matches on the full body of the message, or on individual MIME parts.

The regex condition takes one or more regular expressions as arguments and matches them against the full message (when called in the DATA ACL) or a raw MIME part (when called in the MIME ACL). The regex condition matches linewise, with a maximum line length of 32K characters. That means you cannot have multiline matches with the regex condition.

The mime_regex condition can be called only in the MIME ACL. It matches up to 32K of decoded content (the whole content at once, not linewise). If the part has not been decoded with the decode modifier earlier in the ACL, it is decoded automatically when mime_regex is executed (using default path and filename values). If the decoded data is larger than 32K, only the first 32K characters are checked.

The regular expressions are passed as a colon-separated list. To include a literal colon, you must double it. Since the whole right-hand side string is expanded before being used, you must also escape dollar signs and backslashes with more backslashes, or use the \N facility to disable expansion. Here is a simple example that contains two regular expressions:

deny message = contains blacklisted regex ($regex_match_string)
     regex = [Mm]ortgage : URGENT BUSINESS PROPOSAL

The conditions returns true if any one of the regular expressions matches. The $regex_match_string expansion variable is then set up and contains the matching regular expression. The expansion variables $regex1 $regex2 etc are set to any substrings captured by the regular expression.

Warning: With large messages, these conditions can be fairly CPU-intensive.