This chapter describes library functions available in Mailfromd version 8.17.1. For the simplicity of explanation, we use the word ‘boolean’ to indicate variables of numeric type that are used as boolean values. For such variables, the term ‘False’ stands for the numeric 0, and ‘True’ for any non-zero value.
sed
functionstring
getmacro (string macro)
¶Returns the value of Sendmail macro macro. If macro is
not defined, raises the e_macroundef
exception.
Calling getmacro(name)
is completely equivalent to
referencing ${name}
, except that it allows to construct
macro names programmatically, e.g.:
if getmacro("auth_%var") = "foo" ... fi
boolean
macro_defined (string name)
¶Return true if Sendmail macro name is defined.
Notice, that if your MTA supports macro name negotiation23, you will have to export macro names used by these two functions using ‘#pragma miltermacros’ construct. Consider this example:
func authcheck(string name) do string macname "auth_%name" if macro_defined(macname) if getmacro(macname) ... fi fi done #pragma miltermacros envfrom auth_authen prog envfrom do authcheck("authen") done
In this case, the parser cannot deduce that the envfrom
handler
will attempt to reference the ‘auth_authen’ macro, therefore the
‘#pragma miltermacros’ is used to help it.
sed
function ¶The sed
function allows you to transform a string by replacing
parts of it that match a regular expression with another string. This
function is somewhat similar to the sed
command line utility
(hence its name) and bears similarities to analogous functions in
other programming languages (e.g. sub
in awk
or the
s//
operator in perl
).
string
sed (string subject, expr, …)
¶The expr argument is an s-expressions of the the form:
s/regexp/replacement/[flags]
where regexp is a regular expression, and replacement is a
replacement string for each part of the subject that matches
regexp. When sed
is invoked, it attempts to match
subject against the regexp. If the match succeeds, the
portion of subject which was matched is replaced with
replacement. Depending on the value of flags
(see global replace), this process may continue until the entire
subject has been scanned.
The resulting output serves as input for next argument, if such is supplied. The process continues until all arguments have been applied.
The function returns the output of the last s-expression.
Both regexp and replacement are described in detail in The ‘s’ Command in GNU sed.
Supported flags are:
Apply the replacement to all matches to the regexp, not just the first.
Use case-insensitive matching. In the absence of this flag, the value
set by the recent #pragma regex icase
is used (see icase).
regexp is an extended regular expression (see Extended regular expressions in GNU sed). In the absence of this flag, the value set by the
recent #pragma regex extended
(if any) is used (see extended).
Only replace the numberth match of the regexp.
Note: the POSIX standard does not specify what should happen
when you mix the ‘g’ and number modifiers. Mailfromd
follows the GNU sed
implementation in this regard, so
the interaction is defined to be: ignore matches before the
numberth, and then match and replace all matches from the
numberth on.
Any delimiter can be used in lieue of ‘/’, the only requirement being that it be used consistently throughout the expression. For example, the following two expressions are equivalent:
s/one/two/ s,one,two,
Changing delimiters is often useful when the regex contains
slashes. For instance, it is more convenient to write s,/,-,
than
s/\//-/
.
Here is an example of sed
usage:
set email sed(input, 's/^<(.*)>$/\1/x')
It removes angle quotes from the value of the ‘input’ variable and assigns the result to ‘email’.
To apply several s-expressions to the same input, you can either give
them as multiple arguments to the sed
function:
set email sed(input, 's/^<(.*)>$/\1/x', 's/(.+@)(.+)/\1\L\2\E/x')
or give them in a single argument separated with semicolons:
set email sed(input, 's/^<(.*)>$/\1/x;s/(.+@)(.+)/\1\L\2\E/x')
Both examples above remove optional angle quotes and convert the domain name part to lower case.
Regular expressions used in sed
arguments are controlled by
the #pragma regex
, as another expressions used throughout the
MFL source file. To avoid using the ‘x’ modifier in the above
example, one can write:
#pragma regex +extended set email sed(input, 's/^<(.*)>$/\1/', 's/(.+@)(.+)/\1\L\2\E/')
See Pragma regex, for details about that #pragma
.
So far all examples used constant s-expressions. However, this is
not a requirement. If necessary, the expression can be stored in a
variable or even constructed on the fly before passing it as argument
to sed
. For example, assume that you wish to remove the domain
part from the value, but only if that part matches one of predefined
domains. Let a regular expression that matches these domains be
stored in the variable domain_rx
. Then this can be done as
follows:
set email sed(input, "s/(.+)(@%domain_rx)/\1/")
If the constructed regular expression uses variables whose value should be matched exactly, such variables must be quoted before being used as part of the regexp. Mailfromd provides a convenience function for this:
string
qr (string str[; string delim])
¶Quote the string str as a regular expression. This function selects the characters to be escaped using the currently selected regular expression flavor (see Pragma regex). At most two additional characters that must be escaped can be supplied in the delim optional parameter. For example, to quote the variable ‘x’ for use in double-quoted s-expression:
qr(x, '/"')
string
escape (string str, [string chars])
¶Returns a copy of str with the characters from chars escaped, i.e. prefixed with a backslash. If chars is not specified, ‘\"’ is assumed.
escape('"a\tstr"ing') ⇒ '\"a\\tstr\"ing' escape('new "value"', '\" ') ⇒ 'new\ \"value\"'
string
unescape (string str)
¶Performs the reverse to ‘escape’, i.e. removes any prefix backslash characters.
unescape('a \"quoted\" string') ⇒ 'a "quoted" string'
string
unescape (string str, [string chars])
¶string
domainpart (string str)
¶Returns the domain part of str, if it is a valid email address, otherwise returns str itself.
domainpart("gray") ⇒ "gray" domainpart("gray@gnu.org.ua") ⇒ "gnu.org.ua"
number
index (string s, string t)
¶number
index (string s, string t, number start)
¶Returns the index of the first occurrence of the string t in the string s, or -1 if t is not present.
index("string of rings", "ring") ⇒ 2
Optional argument start, if supplied, indicates the position in string where to start searching.
index("string of rings", "ring", 3) ⇒ 10
To find the last occurrence of a substring, use the function rindex (see rindex).
number
interval (string str)
¶Converts str, which should be a valid time interval specification (see time interval specification), to seconds.
number
length (string str)
¶Returns the length of the string str in bytes.
length("string") ⇒ 6
string
dequote (string str)
¶Removes ‘<’ and ‘>’ surrounding str. If str is not enclosed by angle brackets or these are unbalanced, the argument is returned unchanged:
dequote("<root@gnu.org.ua>") ⇒ "root@gnu.org.ua" dequote("root@gnu.org.ua") ⇒ "root@gnu.org.ua" dequote("there>") ⇒ "there>"
string
localpart (string str)
¶Returns the local part of str if it is a valid email address, otherwise returns str unchanged.
localpart("gray") ⇒ "gray" localpart("gray@gnu.org.ua") ⇒ "gray"
string
replstr (string s, number n)
¶Replicate a string, i.e. return a string, consisting of s repeated n times:
replstr("12", 3) ⇒ "121212"
string
revstr (string s)
¶Returns the string composed of the characters from s in reversed order:
revstr("foobar") ⇒ "raboof"
number
rindex (string s, string t)
¶number
rindex (string s, string t, number start)
¶Returns the index of the last occurrence of the string t in the string s, or -1 if t is not present.
rindex("string of rings", "ring") ⇒ 10
Optional argument start, if supplied, indicates the position in string where to start searching. E.g.:
rindex("string of rings", "ring", 10) ⇒ 2
See also String manipulation.
string
substr (string str, number start)
¶string
substr (string str, number start, number length)
¶Returns the at most length-character substring of str starting at start. If length is omitted, the rest of str is used.
If length is greater than the actual length of the string, the
e_range
exception is signalled.
substr("mailfrom", 4) ⇒ "from" substr("mailfrom", 4, 2) ⇒ "fr"
string
substring (string str, number start, number end)
¶Returns a substring of str between offsets start and
end, inclusive. Negative end means offset from the end of
the string. In other words, yo obtain a substring from start to the
end of the string, use substring(str, start, -1)
:
substring("mailfrom", 0, 3) ⇒ "mail" substring("mailfrom", 2, 5) ⇒ "ilfr" substring("mailfrom", 4, -1) ⇒ "from" substring("mailfrom", 4, length("mailfrom") - 1) ⇒ "from" substring("mailfrom", 4, -2) ⇒ "fro"
This function signals e_range
exception if either start or
end are outside the string length.
string
tolower (string str)
¶Returns a copy of the string str, with all the upper-case characters translated to their corresponding lower-case counterparts. Non-alphabetic characters are left unchanged.
tolower("MAIL") ⇒ "mail"
string
toupper (string str)
¶Returns a copy of the string str, with all the lower-case characters translated to their corresponding upper-case counterparts. Non-alphabetic characters are left unchanged.
toupper("mail") ⇒ "MAIL"
string
ltrim (string str[, string cset)
¶Returns a copy of the input string str with any leading characters present in cset removed. If the latter is not given, white space is removed (spaces, tabs, newlines, carriage returns, and line feeds).
ltrim(" a string") ⇒ "a string" ltrim("089", "0") ⇒ "89"
Note the last example. It shows how ltrim
can be used to
convert decimal numbers in string representation that begins with
‘0’. Normally such strings will be treated as representing octal
numbers. If they are indeed decimal, use ltrim
to strip off
the leading zeros, e.g.:
set dayofyear ltrim(strftime('%j', time()), "0")
string
rtrim (string str[, string cset)
¶Returns a copy of the input string str with any trailing characters present in cset removed. If the latter is not given, white space is removed (spaces, tabs, newlines, carriage returns, and line feeds).
number
vercmp (string a, string b)
¶Compares two strings as mailfromd
version numbers. The
result is negative if b precedes a, zero if they refer to
the same version, and positive if b follows a:
vercmp("5.0", "5.1") ⇒ 1 vercmp("4.4", "4.3") ⇒ -1 vercmp("4.3.1", "4.3") ⇒ -1 vercmp("8.0", "8.0") ⇒ 0
string
sa_format_score (number code, number prec)
¶Format code as a floating-point number with prec decimal digits:
sa_format_score(5000, 3) ⇒ "5.000"
This function is convenient for formatting SpamAssassin scores for use in message headers and textual reports. It is defined in module sa.mfl.
See SpamAssassin, for examples of its use.
string
sa_format_report_header (string text)
¶Format a SpamAssassin report text in order to include it in a RFC 822 header. This function selects the score listing from text, and prefixes each line with ‘* ’. Its result looks like:
* 0.2 NO_REAL_NAME From: does not include a real name * 0.1 HTML_MESSAGE BODY: HTML included in message
See SpamAssassin, for examples of its use.
string
strip_domain_part (string domain, number n)
¶Returns at most n last components of the domain name domain. If n is 0 the function returns domain.
This function is defined in the module strip_domain_part.mfl (see Modules).
Examples:
require strip_domain_part strip_domain_part("puszcza.gnu.org.ua", 2) ⇒ "org.ua" strip_domain_part("puszcza.gnu.org.ua", 0) ⇒ "puszcza.gnu.org.ua"
boolean
is_ip (string str)
¶Returns ‘true’ if str is a valid IPv4 address. This function is defined in the module is_ip.mfl (see Modules).
For example:
require is_ip is_ip("1.2.3.4") ⇒ 1 is_ip("1.2.3.x") ⇒ 0 is_ip("blah") ⇒ 0 is_ip("255.255.255.255") ⇒ 1 is_ip("0.0.0.0") ⇒ 1
string
revip (string ip)
¶Reverses octets in ip, which must be a valid string representation of an IPv4 address.
Example:
revip("127.0.0.1") ⇒ "1.0.0.127"
string
verp_extract_user (string email, string domain)
¶If email is a valid VERP-style email address for domain, this function returns the user name, corresponding to that email. Otherwise, it returns empty string.
verp_extract_user("gray=gnu.org.ua@tuhs.org", 'gnu\..*') ⇒ "gray"
string
sprintf (string format, …)
¶The function sprintf
formats its argument according to
format (see below) and returns the resulting string. It takes
varying number of parameters, the only mandatory one being
format.
The format string is a simplified version of the format argument to
C printf
-family functions.
The format string is composed of zero or more directives: ordinary characters (not ‘%’), which are copied unchanged to the output stream; and conversion specifications, each of which results in fetching zero or more subsequent arguments. Each conversion specification is introduced by the character ‘%’, and ends with a conversion specifier. In between there may be (in this order) zero or more flags, an optional minimum field width, and an optional precision.
Notice, that in practice that means that you should use single quotes with the format arguments, to protect conversion specifications from being recognized as variable references (see singe-vs-double).
No type conversion is done on arguments, so it is important that the
supplied arguments match their corresponding conversion specifiers.
By default, the arguments are used in the order given, where each
‘*’ and each conversion specifier asks for the next argument. If
insufficiently many arguments are given, sprintf
raises
‘e_range’ exception. One can also specify explicitly which
argument is taken, at each place where an argument is required, by
writing ‘%m$’, instead of ‘%’ and ‘*m$’
instead of ‘*’, where the decimal integer m denotes the
position in the argument list of the desired argument, indexed
starting from 1. Thus,
sprintf('%*d', width, num);
and
sprintf('%2$*1$d', width, num);
are equivalent. The second style allows repeated references to the same argument.
The character ‘%’ is followed by zero or more of the following flags:
The value should be converted to an alternate form. For ‘o’ conversions, the first character of the output string is made zero (by prefixing a ‘0’ if it was not zero already). For ‘x’ and ‘X’ conversions, a non-zero result has the string ‘0x’ (or ‘0X’ for ‘X’ conversions) prepended to it. Other conversions are not affected by this flag.
The value should be zero padded. For ‘d’, ‘i’, ‘o’, ‘u’, ‘x’, and ‘X’ conversions, the converted value is padded on the left with zeros rather than blanks. If the ‘0’ and ‘-’ flags both appear, the ‘0’ flag is ignored. If a precision is given, the ‘0’ flag is ignored. Other conversions are not affected by this flag.
The converted value is to be left adjusted on the field boundary. (The default is right justification.) The converted value is padded on the right with blanks, rather than on the left with blanks or zeros. A ‘-’ overrides a ‘0’ if both are given.
A blank should be left before a positive number (or empty string) produced by a signed conversion.
A sign (‘+’ or ‘-’) always be placed before a number produced by a signed conversion. By default a sign is used only for negative numbers. A ‘+’ overrides a space if both are used.
An optional decimal digit string (with nonzero first digit) specifying a minimum field width. If the converted value has fewer characters than the field width, it will be padded with spaces on the left (or right, if the left-adjustment flag has been given). Instead of a decimal digit string one may write ‘*’ or ‘*m$’ (for some decimal integer m) to specify that the field width is given in the next argument, or in the m-th argument, respectively, which must be of numeric type. A negative field width is taken as a ‘-’ flag followed by a positive field width. In no case does a non-existent or small field width cause truncation of a field; if the result of a conversion is wider than the field width, the field is expanded to contain the conversion result.
An optional precision, in the form of a period (‘.’) followed by an optional decimal digit string. Instead of a decimal digit string one may write ‘*’ or ‘*m$’ (for some decimal integer m) to specify that the precision is given in the next argument, or in the m-th argument, respectively, which must be of numeric type. If the precision is given as just ‘.’, or the precision is negative, the precision is taken to be zero. This gives the minimum number of digits to appear for ‘d’, ‘i’, ‘o’, ‘u’, ‘x’, and ‘X’ conversions, or the maximum number of characters to be printed from a string for the ‘s’ conversion.
A character that specifies the type of conversion to be applied. The conversion specifiers and their meanings are:
The numeric argument is converted to signed decimal notation. The precision, if any, gives the minimum number of digits that must appear; if the converted value requires fewer digits, it is padded on the left with zeros. The default precision is ‘1’. When ‘0’ is printed with an explicit precision ‘0’, the output is empty.
The numeric argument is converted to unsigned octal (‘o’), unsigned decimal (‘u’), or unsigned hexadecimal (‘x’ and ‘X’) notation. The letters ‘abcdef’ are used for ‘x’ conversions; the letters ‘ABCDEF’ are used for ‘X’ conversions. The precision, if any, gives the minimum number of digits that must appear; if the converted value requires fewer digits, it is padded on the left with zeros. The default precision is ‘1’. When ‘0’ is printed with an explicit precision 0, the output is empty.
The string argument is written to the output. If a precision is specified, no more than the number specified of characters are written.
A ‘%’ is written. No argument is converted. The complete conversion specification is ‘%%’.
These functions check whether all characters of str fall into a
certain character class according to the ‘C’ (‘POSIX’)
locale24. ‘True’ (1) is returned if they do, ‘false’ (0)
is returned otherwise. In the latter case, the global variable
ctype_mismatch
is set to the index of the first character that
is outside of the character class (characters are indexed from 0).
boolean
isalnum (string str)
¶Checks for alphanumeric characters:
isalnum("a123") ⇒ 1 isalnum("a.123") ⇒ 0 (ctype_mismatch = 1)
boolean
isalpha (string str)
¶Checks for an alphabetic character:
isalnum("abc") ⇒ 1 isalnum("a123") ⇒ 0
boolean
isascii (string str)
¶Checks whether all characters in str are 7-bit ones, that fit into the ASCII character set.
isascii("abc") ⇒ 1 isascii("ab\0200") ⇒ 0
boolean
isblank (string str)
¶Checks if str contains only blank characters; that is, spaces or tabs.
boolean
iscntrl (string str)
¶Checks for control characters.
boolean
isdigit (string str)
¶Checks for digits (0 through 9).
boolean
isgraph (string str)
¶Checks for any printable characters except spaces.
boolean
islower (string str)
¶Checks for lower-case characters.
boolean
isprint (string str)
¶Checks for printable characters including space.
boolean
ispunct (string str)
¶Checks for any printable characters which are not a spaces or alphanumeric characters.
boolean
isspace (string str)
¶Checks for white-space characters, i.e.: space, form-feed (‘\f’), newline (‘\n’), carriage return (‘\r’), horizontal tab (‘\t’), and vertical tab (‘\v’).
boolean
isupper (string str)
¶Checks for uppercase letters.
boolean
isxdigit (string str)
¶Checks for hexadecimal digits, i.e. one of ‘0’, ‘1’, ‘2’, ‘3’, ‘4’, ‘5’, ‘6’, ‘7’, ‘8’, ‘9’, ‘a’, ‘b’, ‘c’, ‘d’, ‘e’, ‘f’, ‘A’, ‘B’, ‘C’, ‘D’, ‘E’, ‘F’.
MFL provides a set of functions for writing to disk files,
pipes or sockets and reading from them. The idea behind them is the
same as in most other programming languages: first you open the
resource with a call to open
which returns a descriptor
i.e. an integer number uniquely identifying the resource. Then you
can write or read from it using this descriptor. Finally, when the
resource is no longer needed, you can close it with a call to
close
.
The number of available resource descriptors is limited. The
default limit is 1024. You can tailor it to your needs
using the max-streams
runtime configuration statement.
See max-streams, for a detailed description.
By default, all I/O operations are unbuffered. This can be changed by setting the following global variables:
number
io_buffering ¶Sets the default buffering type. Allowed values are (symbolic names are defined in status.mfl module):
0
BUFFER_NONE
No buffering. This is the default.
1
BUFFER_FULL
Full buffering. Size of the buffer is set by the
io_buffer_size
global variable (see below).
2
BUFFER_LINE
Line buffering. When reading, it is pretty much the same as
BUFFER_FULL
. When writing, the data are accumulated in buffer
and actually sent to the underlying transport stream when the newline
character is seen.
The initial size of the buffer is set by the io_buffer_size
variable. It will grow as needed during the I/O.
number
io_buffer_size ¶Set the buffer size if io_buffering
is set to
BUFFER_FULL
or BUFFER_LINE
. By default, this variable
is set to the size of the system page.
number
open (string name)
¶The name argument specifies the name of a resource to open and the access rights you need to have on it. The function returns a descriptor of the opened stream, which can subsequently be used as an argument to other I/O operations.
Buffering mode for the opened stream is defined by the
io_buffering
and io_buffer_size
global variables. It
can be changed using the setbuf
function (see setbuf).
First symbols of name determine the type of the resource to be opened and the access mode:
The rest of name is a name of a file. Open the file for read-write access. If the file exists, truncate it to zero length, otherwise create the file.
The rest of name is a name of a file. Open the file for appending (writing at end of file). The file is created if it does not exist.
Treat the rest of name as the command name and its arguments. Run this command and open its standard input for writing. The standard error is closed before launching the program. This can be altered by using the following versions of this construct:
Standard error is redirected to /dev/null.
Execute command with its standard error redirected to the file name. If the file exists, it will be truncated.
Standard error of the command is appended to the file name. If file does not exist, it will be created.
The ‘|2>null:’ construct described above is a shortcut for
|2>>file:/dev/null command
Standard error is redirected to the given syslog facility and, optionally, priority. If the latter is omitted, ‘LOG_ERR’ is assumed.
Valid values for facility are: ‘user’, ‘daemon’, ‘auth’, ‘authpriv’, ‘mail’, and ‘local0’ through ‘local7’. Valid values for priority are: ‘emerg’, ‘alert’, ‘crit’, ‘err’, ‘warning’, ‘notice’, ‘info’, ‘debug’. Both facility and priority may be given in upper, lower or mixed cases.
Notice, that no whitespace characters are allowed between ‘|’ and ‘2>’.
Treat the rest of name as the command name and its arguments. Run this command with its stdin closed and stdout open for reading.
The standard error is treated as described above (see ‘|’).
Treat the rest of name as the command name and its arguments.
Run this command and set up for two-way communication with it, i.e
writes to the descriptor returned by open
will send data to the
program’s standard input, reads from the descriptor will get data from
the program’s standard output.
The standard error is treated as described above (see ‘|’). For example, the following redirects it to syslog ‘mail.debug’:
|&2>syslog:mail.debug command
Treat the rest of name as the URL of a socket to connect to. Valid URL forms are described in milter port specification.
If none of these prefixes is used, name is treated as a name
of an existing file and open
will attempt to open this file for
reading.
The open
function will signal exception e_failure
if it
is unable to open the resource or get the required access to it.
number
spawn (string cmd [, number in, number out, number err])
¶Runs the supplied command cmd. The syntax of the cmd is
the same as for the name argument to open
(see above),
which begins with ‘|’, excepting that the ‘|’ sign is
optional. That is:
spawn("/bin/cat")
has exactly the same effect as
open("|/bin/cat")
Optional arguments specify file stream descriptors to be used for the
program standard input, output and error streams, correspondingly.
If supplied, these should be the values returned by a previous call to
open
or tempfile
. The value ‘-1’ means no
redirection.
Buffering mode for the opened stream is defined by the
io_buffering
and io_buffer_size
global variables. It
can be changed using the setbuf
function (see setbuf).
The example below starts the awk
program with a simple
expression as its argument and redirects the content of the
file /etc/passwd to its standard input. The returned
stream descriptor is bound to the command’s standard output
(see the description of ‘|<’ prefix above). The standard
error is closed:
number fd spawn("<awk -F: '{print $1}'", open("/etc/passwd"))
void
close (number rd)
¶The argument rd is a resource descriptor returned by a
previous call to open
. The function close
closes the
resource and deallocates any memory associated with it.
close
will signal e_range
exception if rd lies
outside of allowed range of resource descriptors. See max-streams.
Notice that you are not required to close resources opened by open
.
Any unclosed resource will be closed automatically upon the
termination of the filtering program.
void
shutdown (number rd, number how)
¶This function causes all or part of a full-duplex connection to be
closed. The rd must be either a socket descriptor (returned by
open(@...)
) or a two-way pipe socket descriptor (returned by
open(|&...)
), otherwise the call to shutdown
is
completely equivalent to close
.
The how
argument identifies which part of the connection to
shut down:
number
tempfile ([string tmpdir])
¶Creates a nameless temporary file and returns its descriptor. Optional tmpdir supplies the directory where to create the file, instead of the default /tmp.
void
rewind (number rd)
¶Rewinds the stream identified by rd to its beginning.
number
copy (number dst, number src)
¶Copies all data from the stream src to dst. Returns number of bytes copied.
The following functions provide basic read/write capabilities.
void
write (number rd, string str [, number size])
¶Writes the string str to the resource descriptor rd. If the size argument is given, writes this number of bytes.
This function always attempts to write the requested amount of
data. It will signal e_range
exception if rd lies
outside of allowed range of resource descriptors, e_io
exception if an I/O error occurs, and e_eof
exception
if it wrote 0 bytes (e.g. because the underlying device is full).
void
write_body (number rd, pointer bp , number size)
¶Write the body segment of length size from pointer bp to
the stream rd. This function can be used only in prog
body
(see body handler). Its second and third arguments
correspond exactly to the parameters of the body
handler, so
the following construct writes the message body to the resource
fd
, which should have been open prior to invoking the
body
handler:
prog body do write_body(fd, $1, $2) done
string
read (number rd, number n)
¶Read and return at most n bytes from the resource descriptor rd.
If there are less than n bytes in the stream, the remaining
bytes will be returned. Use length()
to obtain the actual size
of the returned data. If there are no bytes left, the e_eof
exception will be signalled.
The function may signal the following exceptions:
rd lies outside of allowed range of resource descriptors.
End of file encountered.
An I/O error occurred.
string
getdelim (number rd, string delim)
¶Read and return the next string terminated by delim from the resource descriptor rd.
The terminating delim string will be removed from the return value.
When using this function, it is highly recommended to enable full
buffering for fd, either by setting io_buffering
before
open
(see io_buffering) or by calling setbuf
after it (see setbuf). See getline, for an example.
This function may signal the following exceptions:
rd lies outside of allowed range of resource descriptors.
End of file encountered.
An I/O error occurred.
string
getline (number rd)
¶Read and return the next line from the resource
descriptor rd. A line is any sequence of characters terminated
with the default line delimiter. The default delimiter is
a property of rd, i.e. different descriptors can have different
line delimiters. The default value is ‘\n’ (ASCII 10), and can
be changed using the fd_set_delimiter
function (see below).
When using this function, it is highly recommended to enable full
buffering for fd, either by setting io_buffering
before
open
(see io_buffering) or by calling setbuf
after it (see setbuf), e.g.:
set fd open(input) setbuf(fd, BUFFER_FULL) set line getline(fd) ...
This function may signal the following exceptions:
rd lies outside of allowed range of resource descriptors.
End of file encountered.
An I/O error occurred.
void
fd_set_delimiter (number fd, string delim)
¶Set new line delimiter for the descriptor fd, which must be in opened state.
Default delimiter is a newline character (ASCII 10). The following example shows how to change it to CRLF sequence:
fd_set_delimiter(fd, "\r\n")
string
fd_delimiter (number fd)
¶Returns the line delimiter string for fd.
The following example shows how mailfromd
I/O functions can
be used to automatically add IP addresses to an RBL zone:
set nsupdate_cmd "/usr/bin/nsupdate -k /etc/bind/Kmail.+157+14657.private" func block_address(string addr) do number fd string domain set fd open "|%nsupdate_cmd" set domain revip(addr) . ".rbl.myzone.come" write(fd, "prereq nxrrset %domain A\n" "update add %domain 86400 A %addr\n\n" done
The function revip
is defined in revip.
void
setbuf(number fd, [number type, number size])
¶Changes the buffering mode of fd according to the remaining two
arguments. The type specifies buffering type
(see io_buffering), and size supplies the buffer size for
buffering types BUFFER_FULL
and BUFFER_LINE
. If
size is omitted, it defaults to io_buffer_size
(see io_buffer_size). Omitted type defaults to
io_buffering
(see io_buffering).
number
getbuftype(number fd)
¶Returns the type of buffering currently in effect for the descriptor fd. See io_buffering, for a list of possible return values.
If this function returns BUFFER_FULL
or BUFFER_LINE
, you
can use getbufsize
to get the associated buffer size.
number
getbufsize(number fd)
¶Returns the buffer size for the descriptor fd.
This section describes functions that transform data using
Mailutils filter pipes. Filter pipe is a string defining data
flow between several filters. Each filter takes input,
transforms it according to certain rules and produces the transformed
data on its output. As in shell, multiple filters are connected
using pipe characters (‘|’). For example, the crlf
filter
inserts a carriage return character before each newline character. A
filter doing that kind of transformation is defined as:
"crlf"
Another filter, base64
, converts its input to a BASE64 encoded
string. To transform each newline into carriage return + newline pair
and encode the resulting stream in BASE64, one would write:
"crlf | base64"
Some filters take one or more arguments. These are specified as
a comma-delimited list in parentheses after the filter name. For
example, the linelen
filter limits the length of each output
line to the given number of octets. The following filter pipe will
limit the length of base64 lines in the filter above to 62 octets:
"crlf | base64 | linelen(62)"
Many filters operate in two modes: encode and decode. By
default all MFL functions apply filters in encode mode. The desired
mode can be stated explicitly in the filter string by using
encode()
and decode()
functions. They take a filter
pipe line as their argument. For example, the following will decode
the stream produced by the example filter above:
"decode(base64 | crlf)"
See Filters and Filter Pipes, for a discussion of available filters and their arguments.
string
filter_string (string input, string filter_pipe)
¶Transforms the string input using filters in filter_pipe and returns the result. Example:
set input "test\ninput\n" filter_string(input, "crlf|base64") ⇒ "dGVzdA0KaW5wdXQNCg=="
void
filter_fd (number src_fd, number dst_fd, string filter_pipe)
¶Given two I/O descriptors, reads data from src_fd, transforms it using filter_pipe and writes the result to descriptor dst_fd.
Both descriptors must be obtained using functions described in I/O functions.
A filter pipe is a string consisting of filter invocations
delimited by pipe characters (‘|’). Each invocation
is a filter name optionally followed by a comma-separated list of
parameters. Most filters can operate in two modes: encode and
decode. Unless specified otherwise, filters are invoked in
encode mode. To change the mode, the encode
and decode
meta-filters are provided. Argments to these filters are filter pipes
that will be executed in the corresponding mode.
The following Mailutils filters are available:
In encode mode, converts its input into 7-bit ASCII, by clearing the 8th bit on each processed byte.
In decode mode, it operates exactly as the 8bit filter, i.e. copies its input to the output verbatim.
The filter takes no arguments.
Encodes or decodes the input using the base64
encoding.
The only difference between BASE64
and B
is that, in
encode mode, the former limits each ouput line length to 76 octets,
whereas the latter produces a contiguous stream of base64 data.
In decode mode, both filters operate exactly the same way.
A convenience interface to the iconv
filter, available for use
only in the message_body_to_stream
function. It decodes the
part of a MIME message from its original character set, which is
determined from the value of the Content-Type
header, to the
destination character set cset. Optional fallback
parameter specifies the representation fallback to be used for octets
that cannot be converted between the charater sets. Its use is
described in See iconv.
This filter is normally takes its input from the mimedecode
filter, as in:
message_body_to_stream(fd, msg, 'mimedecode|charset(utf-8)')
See mimedecode, for a detailed discussion.
Converts line separators from LF (ASCII 10) to CRLF (ASCII 13 10) and vice-versa.
In decode mode, translates each CRLF to LF. Takes no arguments.
In encode mode, translates each LF to CRLF. If an optional argument ‘-n’ is given, produces a normalized output, by preserving each input CRLF sequence untouched (otherwise such sequences will be are translated to CR CR LF).
In encode mode, replaces each LF (‘\n’ or ASCII 10) character with CRLF (‘\r\n’, ASCII 13 10), and byte-stuffs the output by producing an additional ‘.’ in front of any ‘.’ appearing at the beginning of a line in input. Upon end of input, it outputs additional ‘.\r\n’, if the last output character was ‘\n’, or ‘\r\n.\r\n’ otherwise.
If supplied the ‘-n’ argument, it preserves each CRLF input
sequence untranslated (see the CRLF
above).
In decode mode, the reverse is performed: each CRLF is replaced with a single LF byte, and additional dots are removed from beginning of lines. A single dot on a line by itself marks the end of the stream and causes the filter to return EOF.
In encode mode, byte-stuffs the input by outputting an additional dot (‘.’) in front of any dot appearing at the beginning of a line. Upon encountering end of input, it outputs additional ‘.\n’.
In decode mode, the reverse is performed: additional dots are removed from beginning of lines. A single dot on a line by itself (i.e. the sequence ‘\n.\n’) marks the end of the stream and causes the filter to return EOF.
This filter doesn’t take arguments.
Performs a traditional UNIX processing of lines starting with a ‘From’ followed by a space character.
In encode mode, each ‘From ’ at the beginning of a line is replaced by ‘>From ’.
In decode mode, the reverse operation is performed: initial greater-then sign (‘>’) is removed from any line starting with ‘>From ’.
The filter takes no arguments.
MBOXRD-compatible processing of envelope lines.
In encode mode, each ‘From ’ optionally preceded by any number of contiguous ‘>’ characters and appearing at the beginning of a line is prefixed by another ‘>’ character on output.
In decode mode, the reverse operation is performed: initial greater-then sign (‘>’) is removed from any line starting with one or more ‘>’ characters followed by ‘From ’.
This filter treats its input as a RFC-2822 email message. It extracts its header part (i.e. everything up to the first empty line) and copies it to the output. The body of the message is ignored.
The filter operates only in decode mode and takes no arguments.
Converts input from character set src to dst. The filter works the same way in both decode and encode modes.
It takes two mandatory arguments: the names of the input (src) and output (dst) charset. Optional third argument specifies what to do when an illegal character sequence is encountered in the input stream. Its possible values are:
Raise a e_ilseq
exception.
Copy the offending octet to the output verbatim and continue conversion from the next octet.
Print the offending octet to the output using the C octal conversion and continue conversion from the next octet.
The default is copy-octal
.
The following example creates a iconv
filter for converting from
iso-8859-2
to utf-8
, raising the e_ilseq
exception on the first conversion error:
iconv(iso-8859-2, utf-8, none)
In decode mode, the filter removes from the input all lines beginning with a given inline comment sequence str. The default comment sequence is ‘;’ (a semicolon).
The following options modify the default behavior:
Emit line number information after each contiguous sequence of removed lines. The argument str supplies an information starter – a sequence of characters which is output before the actual line number.
Remove empty lines, i.e. the lines that contain only whitespace characters.
Squeeze whitespace. Each sequence of two or more whitespace characters encountered on input is replaced by a single space character on output.
A whitespace-must-follow mode. A comment sequence is recognized only if followed by a whitespace character. The character itself is retained on output.
In encode mode the inline-comment
filter adds a comment-starter
sequence at the beginning of each line. The default comment-starter
is ‘;’ and can be changed by specifying the desired comment
starter as the first argument.
The only option supported in this mode is -S, which enables the whitespace-must-follow mode, in which a single space character (ASCII 20) is output after each comment sequence.
Implements a familiar UNIX line-continuation facility. The filter removes from itsinput stream any newline character immediately preceded by a backslash. This filter operates only in decode mode.
If given the arguments (‘-i’, str), enables the line number information facility. This facility emits current input line number (prefixed with str) after each contiguous sequence of one or more removed newline characters. It is useful for implementing parsers which are normally supposed to identify eventual erroneous lines with their input line numbers.
Limits the length of each output line to a certain number of octets. It operates in encode mode only and requires a single parameter: the desired output length in octets. This filter makes no attempt to analyze the lexical structure of the input: the newline caracters are inserted when the length of the output line reaches a predefined maximum. Any newline characters present in the input are taken into account when computing the input line length.
This is a domain-specific filter available for use only with the
message_body_to_stream
function. It decodes the part of
a MIME message from whatever encoding that was used to store it
in the message to a stream of bytes. See mimedecode.
Encodes or decodes the input using the quoted-printable encoding.
In encode mode, the xml
filter converts input stream (which must
contain valid UTF-8 characters) into a form suitable for inclusion into
a XML or HTML document, i.e. it replaces ‘<’, ‘>’, and
‘&’ with ‘<’, ‘>’, and ‘&’,
correspondingly, and replaces invalid characters with their numeric
character reference representation.
In decode mode, a reverse operation is performed.
The filter does not take arguments.
number
email_map (string email)
¶Parses email and returns a bitmap, consisting of zero or more of the following flags:
email has more than one email address.
email has comment parts.
email has personal part.
email has local part.
email has domain part.
email has route part.
These constants are declared in the email.mfl module. The
function email_map
returns 0 if its argument is not a valid
email address.
boolean
email_valid (string email)
¶Returns ‘True’ (1) if email is a valid email address, consisting of local and domain parts only. E.g.:
email_valid("gray@gnu.org") ⇒ 1 email_valid("gray") ⇒ 0 email_valid('"Sergey Poznyakoff <gray@gnu.org>') ⇒ 0
This function is defined in email.mfl (see Modules).
Envelope modification functions set sender and add or delete recipient addresses from the message envelope. This allows MFL scripts to redirect messages to another addresses.
void
set_from (string email [, string args])
¶Sets envelope sender address to email, which must be a valid email address. Optional args supply arguments to ESMTP ‘MAIL FROM’ command.
void
rcpt_add (string address)
¶Add the e-mail address to the envelope.
void
rcpt_delete (string address)
¶Remove address from the envelope.
The following example code uses these functions to implement a simple alias-like capability:
prog envrcpt do string alias dbget(aliasdb, $1, "NULL", 1) if alias != "NULL" rcpt_delete($1) rcpt_add(alias) fi done
There are two ways to modify message headers in a MFL script. First is to use header actions, described in Action Statements, and the second way is to use message modification functions. Compared with the actions, the functions offer a series of advantages. For example, using functions you can construct the name of the header to operate upon (e.g. by concatenating several arguments), something which is impossible when using actions. Moreover, apart from three basic operations (add, modify and remove), as supported by header actions, header functions allow to insert a new header into a particular place.
void
header_add (string name, string value)
¶Adds a header ‘name: value’ to the message.
In contrast to the add
action, this function allows to construct
the header name using arbitrary MFL expressions.
void
header_add (string name, string value, number idx)
¶This syntax is preserved for backward compatibility. It is equivalent
to header_insert
, which see.
void
header_insert (string name, string value, number idx)
¶This function inserts a header ‘name: ‘value’’ at
idxth header position in the internal list of headers maintained
by the MTA. That list contains headers added to the message either by
the filter or by the MTA itself, but not the headers included in the
message itself. Some of the headers in this list are conditional,
e.g. the ones added by the ‘H?cond?’ directive in
sendmail.cf. MTA evaluates them after all header modifications
have been done and removes those of headers for which they yield false.
This means that the position at which the header added by
header_insert
will appear in the final message will differ from
idx.
void
header_delete (string name [, number index])
¶Delete header name from the envelope. If index is given, delete indexth instance of the header name.
Notice the differences between this function and the delete
action:
delete
requires it to be a literal string.
void
header_replace (string name, string value [, number index])
¶Replace the value of the header name with value. If index is given, replace indexth instance of header name.
Notice the differences between this function and the replace
action:
replace
requires it to be a literal string.
void
header_rename (string name, string newname[, number idx])
¶Defined in the module header_rename.mfl.
Available only in the ‘eom’ handler.
Renames the idxth instance of header name to newname. If idx is not given, assumes 1.
If the specified header or the idx instance of it is not present in the current message, the function silently returns. All other errors cause run-time exception.
The position of the renamed header in the header list is not preserved.
The example below renames ‘Subject’ header to ‘X-Old-Subject’:
require 'header_rename' prog eom do header_rename("Subject", "X-Old-Subject") done
void
header_prefix_all (string name [, string prefix])
¶Defined in the module header_rename.mfl.
Available only in the ‘eom’ handler.
Renames all headers named name by prefixing them with prefix. If prefix is not supplied, removes all such headers.
All renamed headers will be placed in a continuous block in the header list. The absolute position in the header list will change. Relative ordering of renamed headers will be preserved.
void
header_prefix_pattern (string pattern, string prefix)
¶Defined in the module header_rename.mfl.
Available only in the ‘eom’ handler.
Renames all headers with names matching pattern (in the sense of
fnmatch
, see fnmatches) by prefixing
them with prefix.
All renamed headers will be placed in a continuous block in the header list. The absolute position in the header list will change. Relative ordering of renamed headers will be preserved.
If called with one argument, removes all headers matching pattern.
For example, to prefix all headers beginning with ‘X-Spamd-’ with an additional ‘X-’:
require 'header_rename' prog eom do header_prefix_pattern("X-Spamd-*", "X-") done
Body modification is an experimental feature of MFL. The version 8.17.1 provides only one function for that purpose.
void
replbody (string text)
¶Replace the body of the message with text. Notice, that text must not contain RFC 822 headers. See the previous section if you want to manipulate message headers.
Example:
replbody("Body of this message has been removed by the mail filter.")
No restrictions are imposed on the format of text.
void
replbody_fd (number fd)
¶Replaces the body of the message with the content of the stream fd. Use this function if the body is very big, or if it is returned by an external program.
Notice that this function starts reading from the current position in
fd. Use rewind
if you wish to read from the beginning of
the stream.
The example below shows how to preprocess the body of the message using external program /usr/bin/mailproc, which is supposed to read the body from its standard input and write the processed text to its standard output:
number fd # Temporary file descriptor
prog data
do
# Open the temporary file
set fd tempfile()
done
prog body
do
# Write the body to it.
write_body(fd, $1, $2)
done
prog eom
do
# Use the resulting stream as the stdin to the mailproc
# command and read the new body from its standard output.
rewind(fd)
replbody_fd(spawn("</usr/bin/mailproc", fd))
done
Message modification functions described in the previous subsections do not take effect immediately, in the moment they are called. Instead they store the requested changes in the internal message modification queue. These changes are applied at the end of processing, before ‘eom’ stage finishes (see Figure 3.1).
One important consequence of this way of operation is that calling
any MTA action (see Action Statements), causes all prior
modifications to the message to be ignored. That is because after
receiving the action command, MTA will not call filter
for that message any more. In particular, the ‘eom’ handler will
not be called, and the message modification queue will not be flushed.
While it is logical for such actions as reject
or
tempfail
, it may be quite confusing for accept
.
Consider, for example, the following code:
prog envfrom do if $1 == "" header_add("X-Filter", "foo") accept fi done
Obviously, the intention was to add a ‘X-Filter’ header and accept the message if it was sent from the null address. What happens in reality, however, is a bit different: the message is accepted, but no header is added to it. If you need to accept the message and retain any modifications you have done to it, you need to use an auxiliary variable, e.g.:
number accepted 0 prog envfrom do if $1 == "" header_add("X-Filter", "foo") set accepted 1 fi done
Then, test this variable for non-zero value at the beginning of each subsequent handler, e.g.:
prog data do if accepted continue fi ... done
To help you trace such problematic usages of accept
,
mailfromd
emits the following warning:
RUNTIME WARNING near /etc/mailfromd.mfl:36: `accept' causes previous message modification commands to be ignored; call mmq_purge() prior to `accept', to suppress this warning
If it is OK to lose all modifications, call mmq_purge
, as
suggested in this message.
void
mmq_purge ()
¶Remove all modification requests from the queue. This function undoes
the effect of any of the following functions, if they had been called
previously: rcpt_add
, rcpt_delete
, header_add
,
header_insert
, header_delete
, header_replace
,
replbody
, quarantine
.
string
message_header_encode (string text, [string enc, string charset])
¶Encode text in accordance with RFC 2047. Optional arguments:
Encoding to use. Valid values are ‘quoted-printable’, or ‘Q’ (the default) and ‘base64’, or ‘B’.
Character set. By default ‘UTF-8’.
If the function is unable to encode the string, it raises the
exception e_failure
.
For example:
set string "Keld Jørn Simonsen <keld@dkuug.dk>" message_header_encode(string, "ISO-8859-1") ⇒ "=?ISO-8859-1?Q?Keld_J=F8rn_Simonsen?= <keld@dkuug.dk>"
string
message_header_decode (string text, [string charset])
¶text must be a header value encoded in accordance with RFC
2047. The function returns the decoded string. If the decoding fails,
it raises e_failure
exception. The optional argument
charset specifies the character set to use (default –
‘UTF-8’).
set string "=?ISO-8859-1?Q?Keld_J=F8rn_Simonsen?= <keld@dkuug.dk>" message_header_decode(string) ⇒ "Keld Jørn Simonsen <keld@dkuug.dk>"
string
unfold (string text)
¶If text is a “folded” multi-line RFC 2822 header value, unfold it. If text is a single-line string, return its unchanged copy.
For example, suppose that the message being processed contained the following header:
List-Id: Sent bugreports to <some-address@some.net>
Then, applying unfold
to its value25 will produce:
Sent bugreports to <some-address@some.net>
string
body_string (pointer text, number count)
¶Converts first count bytes from the memory location pointed to by text into a regular string.
This function is intended to convert the $1
argument passed to
a body
handler to a regular MFL string. For more
information about its use, see body handler.
bool
body_has_nulls (pointer text, number count)
¶Returns ‘True’ if first count bytes of the string pointed to by text contain ASCII NUL characters.
Example:
prog body do if body_has_nulls($1, $2) reject fi done
The following function is available only in the ‘eom’ handler:
void
progress ()
¶Notify the MTA that the filter is still processing the message. This causes MTA to restart its timeouts and allows additional amount of time for execution of ‘eom’.
Use this function if your ‘eom’ handler needs additional time for processing the message (e.g. for scanning a very big MIME message). You may call it several times, if the need be, although such usage is not recommended.
number
current_message ()
¶This function can be used in eom
handlers only. It returns a
message descriptor referring to the current message. See Message Functions, for a description of functions for accessing messages.
The functions below access the headers from the current message. They
are available in the following handlers: eoh
, body
, eom
.
number
current_header_count ([string name])
¶Return number of headers in the current message. If name is specified, return number of headers that have this name.
current_header_count() ⇒ 6 current_header_count("Subject") ⇒ 1
string
current_header_nth_name (number n)
¶Return the name of the nth header. The index n is 1-based.
string
current_header_nth_value (number n)
¶Return the value of the nth header. The index n is 1-based.
string
current_header (string name [, number n])
¶Return the value of the named header, e.g.:
set s current_header("Subject")
Optional second argument specifies the header instance, if there are more than 1 header of the same name, e.g.:
set s current_header("Received", 2)
Header indices are 1-based.
All current_header function raise the e_not_found
exception if the
requested header was not found.
A set of functions is provided for accessing mailboxes and messages within them. In this subsection we describe the functions for accessing mailboxes.
A mailbox is opened using mailbox_open
function:
number
mailbox_open (string url [, string mode, string perms])
¶Open a mailbox identified by url. Return a mailbox descriptor: a unique numeric identifier that can subsequently be used to access this mailbox.
The optional mode argument specifies the access mode for the mailbox. Its valid values are:
Value | Meaning |
---|---|
r | Open mailbox for reading. This is the default. |
w | Open mailbox for writing. If the mailbox does not exist, it is created. |
rw | Open mailbox for reading and writing. If the mailbox does not exist, it is created. |
wr | Same as ‘rw’. |
w+ | Open mailbox for reading and writing. If the mailbox does not exist, it is created. |
a | Open mailbox for appending messages to it. If the mailbox does not exist, an exception is signalled. |
a+ | Open mailbox for appending messages to it. If the mailbox does not exist, it is created. |
The optional perms argument specifies the permissions to use in case a new file (or files) is created. It is a comma-separated list of:
[go](+|=)[wr]+
The initial letter controls which users’ access is to be set: users in the file’s group (‘g’) or other users not in the file’s group (‘o’). The following character controls whether the permissions are added to the default ones (‘+’) or applied instead of them (‘=’). The remaining letters specify the permissions: ‘r’ for read access and ‘w’ for write access. For example:
g=rw,o+r
The number of mailbox descriptors available for simultaneous opening
is 64. This value can be changed using the
max-open-mailboxes
runtime configuration statement
(see max-open-mailboxes).
number
mailbox_messages_count (number nmbx)
¶Return the number of messages in mailbox. The argument nmbx is
a valid mailbox descriptor as returned by a previous call to
mailbox_open
.
number
mailbox_get_message (number mbx, number n)
¶Retrieve nth message from the mailbox identified by descriptor mbx. On success, the function returns a message descriptor, an integer number that can subsequently be used to access that message (see Message Functions). On error, an exception is raised.
Messages in a mailbox are numbered starting from 1.
void
mailbox_close (number nmbx)
¶Close a mailbox previously opened by mailbox_open
.
void
mailbox_append_message (number nmbx, number nmsg)
¶Append message nmsg to mailbox nmbx. The message
descriptor nsmg must be obtained from a previous call to
mailbox_get_message
or current_message
(see current_message).
The functions described below retrieve information from RFC822
messages. The message to operate upon is identified by its
descriptor, an integer number returned by the previous call to
mailbox_get_message
(see mailbox_get_message) or current_message
(see current_message) function. The maximum number of message
descriptors is limited by 1024. You can change this limit
using the max-open-messages
runtime configuration statement
(see max-open-messages).
number
message_size (number nmsg)
¶Return the size of the message nmsg, in bytes. Notice,
that if nmsg refers to current message
(see current_message), the returned value is less than the size
seen by the MTA, because mailfromd
recodes
CR-LF sequences to LF, i.e. removes carriage
returns (ASCII 13) occurring before line feeds
(ASCII 10. To obtain actual message length as seen by the
MTA, add the number of lines in the message:
set actual_length message_size(nmsg) + message_lines(nmsg)
boolean
message_body_is_empty (number nmsg)
¶Returns true
if the body of message nmsg has zero size or
contains only whitespace characters. If the
‘Content-Transfer-Encoding’ header is present, it is used to
decode body before processing.
void
message_close (number nmsg)
¶Close the message identified by descriptor nmsg.
number
message_lines (number nmsg)
¶Return total number of lines in message nmsg. The following relation holds true:
message_lines(x) = message_body_lines(x) + message_header_lines(x) + 1
string
message_read_line (number nmsg)
¶Read and return next line from the message nmsg. If
there are no more lines to read, raise the eof
exception.
Use message_rewind
to rewind the message stream and read its
contents again.
void
message_rewind (number nmsg)
¶Rewind the stream associated with message referred to by descriptor nmsg.
number
message_from_stream (number fd; string filter_chain)
¶Converts contents of the stream identified by fd to a mail message. Returns identifier of the created message.
Optional filter_chain supplies the name of a Mailutils filter chain, through which the data will be passed before converting. See http://mailutils.org/wiki/Filter_chain, for a description of filter chains.
void
message_to_stream (number fd, number nmsg; string filter_chain)
¶Copies message nsmg to stream descriptor fd. The
descriptor must be obtained by a previous call to open
.
Optional filter_chain supplies the name of a Mailutils filter chain, through which the data will be passed before writing them to fd. See http://mailutils.org/wiki/Filter_chain, for a description of filter chains.
number
message_header_size (number nmsg)
¶Return the size, in bytes of the headers of message nmsg. See the
note to the message_size
, above.
number
message_header_lines (number nmsg)
¶Return number of lines occupied by headers in message nmsg.
number
message_header_count (number nmsg, [string name])
¶Return number of headers in message nmsg.
If name is supplied, count only headers with that name.
string
message_find_header (number nmsg, string name [, number idx])
¶Return value of header name from the message nmsg. If the message contains several headers with the same name, optional parameter idx may be used to select one of them. Headers are numbered from ‘1’.
If no matching header is not found, the not_found
exception is
raised. If another error occurs, the failure
exception is
raised.
The returned string is a verbatim copy of the message contents (except
for eventual CR-LF -> LF translation, see above).
You might need to apply the unfold
function to it (see unfold).
string
message_nth_header_name (number nmsg, number n)
¶Returns the name of the nth header in message nmsg. If
there is no such header, e_range
exception is raised.
string
message_nth_header_value (number msg, number n)
¶Returns the value of the nth header in message nmsg. If
there is no such header, e_range
exception is raised.
boolean
message_has_header (number nmsg, string name [, number idx])
¶Return true
if message nmsg contains header with the
given name. If there are several headers with the same name,
optional parameter idx may be used to select one of them.
number
message_body_size (number nmsg)
¶Return the size, in bytes, of the body of message nmsg. See the
note to the message_size
, above.
number
message_body_lines (number nmsg)
¶Return number of lines in the body of message referred to by descriptor nmsg.
void
message_body_rewind (number nmsg)
¶Rewind the stream associated with the body of message referred to by descriptor nmsg.
A call to message_body_read_line
(see below) after calling this
function will return the first line from the message body.
string
message_read_body_line (number nmsg)
¶Read and return next line from the body of the message nmsg. If
there are no more lines to read, raise the eof
exception.
Use message_body_rewind
(see above) to rewind the body stream
and read its contents again.
void
message_body_to_stream (number fd, number nmsg; string filter_pipe)
¶Copies the body of the message nsmg to stream descriptor
fd. The descriptor must be obtained by a previous call to
open
.
Optional filter_pipe supplies a sequence of Mailutils filters, through which the data will be passed before writing them to fd. See Filtering functions, for a discussion of filter pipe syntax.
In addition to filters described in See Filters and Filter Pipes, two special
filters are provided for use with this function: mimedecode
and
charset
. The mimedecode
filter instructs the function
to decode the message body by reverting the encoding specified by its
Content-Transfer-Encoding
header. It is normally used as the
very first filter in chain. The charset
filter recodes the
message body from it original character set to the character set
specified as its argument.
See mimedecode, for a detailed discussion of this feature.
boolean
message_is_multipart (number nmsg)
¶Return true
if message nmsg is a multipart
(MIME) message.
number
message_count_parts (number nmsg)
¶Return number of parts in message nmsg, if it is a multipart (MIME) message. If it is not, return ‘1’.
Use message_is_multipart
to check whether the message is a
multipart one.
number
message_get_part (number nmsg, number n)
¶Extract nth part from the multipart message nmsg. Numeration of parts begins from ‘1’. Return message descriptor referring to the extracted part. Message parts are regarded as messages, so any message functions can be applied to them.
string
message_content_type (number nmsg)
¶Returns content type for the message nmsg. The returned string is composed of content type and subtype, delimited by slash.
If nmsg is not a multipart message, the function returns ‘text/plain’.
Several functions are provided for decoding multi-part messages. Such
decoding is governed by Content-Transfer-Encoding
and
Content-Type
headers of the message. The
Content-Transfer-Encoding
header defines the method used to
encode the message. The value of Content-Type
header is used
to determine the character set the body is written in.
Basic MIME decoding facilities are provided by the built-in function
message_body_to_stream
, described in the previous subsection.
To instruct it to decode the content, pass it the filter_chain
argument beginning with the word mimedecode
. The usual
sequence is:
set fd open("> outfile") message_body_to_stream(fd, msg, "mimedecode")
To ensure that the produced stream is represented in a specific
character set, use the charset
special filter. Its argument is
the name of the character set to recode the text to:
set fd open("> outfile") message_body_to_stream(fd, msg, "mimedecode|charset(utf-8)")
The charset
filter takes also an optional second argument – a
fallback method, specifying what to do when an octet sequence is
encountered that cannot be represented in the requested character set.
Possible values for this argument are:
Stop further conversion and signal the e_ilseq
exception.
Copy the offending character to the output verbatim.
Represent the offending character as a C octal sequence (‘\nnn’, where n is an octal digit). This is the default.
To decode a particular part of the message, first extract it using the
message_get_part
function. Recall that message parts are
messages as well, and as such can be passed to
message_body_to_stream
. For example, the following code
fragment extracts all top-level parts of a multi-part message to files
named ‘part.N’:
if message_is_multipart(msg) set n message_count_parts(msg) loop for set i 1, while i <= n, set i i + 1 do set fd open("> part.%i") message_body_to_stream(fd, message_get_part(msg, i), "mimedecode") close(fd) done fi
The mime.mfl module provides additional functions for decoding multi-part messages:
number
message_body_decode (number nmsg; string charset, string fallback)
¶Decodes the body of the message (or message part) nmsg, optionally converting it to the given charset. The fallback argument specifies what to do if a byte sequence cannot be converted to the specified character set. See iconv fallback, for a detailed discussion.
The function returns a descriptor of the I/O stream that contains the decoded material. See I/O functions for a discussion of functions available for reading from it.
number
message_part_decode(number nmsg, number part; string charset, string fallback)
¶Decodes the body of the given part of a MIME message nmsg. The
argument part is a 1-based index of the part in the message.
Optional arguments charset and fallback have the same
meaning as in message_body_decode
(see above).
Returns a descriptor of the I/O stream that contains the decoded material.
This function is equivalent to:
message_body_decode(message_get_part(nmsg, part, charset, fallback))
Message digests are specially formatted messages that contain certain number of mail messages, encapsulated using the method described in RFC 934. Such digests are often used in mailing lists to reduce the frequency of sending mails. Messages of this format are also produced by the forward function in most MUA’s.
The usual way to handle a message digest in MFL is to convert it first to a MIME message, and then to use functions for accessing its parts (see MIME functions).
number
message_burst (number nmsg ; number flags)
¶Converts the message identified by the descriptor nmsg to a multi-part message. Returns a descriptor of the created message.
Optional argument flags controls the behavior of the bursting agent. It is a bitwise OR of error action and bursting flags.
Error action defines what to do if a part of the digest is
not in RFC822 message format. If it is ‘BURST_ERR_FAIL’ (the
default), the function will raise the ‘e_format’ exception. If
onerr is ‘BURST_ERR_IGNORE’, the improperly formatted part
will be ignored. Finally, the value ‘BURST_ERR_BODY’ instructs
message_burst
to create a replacement part with empty headers
and the text of the offending part as its body.
Bursting flags control various aspects of the agent behavior. Currently only one flag is defined, ‘BURST_DECODE’, which instructs the agent to decode any MIME parts (according to the ‘Content-Transfer-Encoding’ header) it encounters while bursting the message.
Parts of a message digest are separated by so-called encapsulation boundaries, which are in essence lines beginning with at least one dash followed by a non-whitespace character. A dash followed by a whitespace serves as a byte-stuffing character, a sort of escape for lines which begin with a dash themselves. Unfortunately, there are mail agents which do not follow byte-stuffing rules and pass lines beginning with dashes unmodified into resulting digests. To help handle such cases a global variable is provided which controls how much dashes should the line begin with for it to be recognized as an encapsulation boundary.
number
burst_eb_min_length ¶Minimal number of consecutive dashes an encapsulation boundary must begin with.
The default is 2.
The following example shows a function which saves all parts of a digest message to separate disk files. The argument orig is a message descriptor. The resulting files are named by concatenating the string supplied by the stem argument and the ordinal number (1-based) of the message part.
func burst_digest(number orig, string stem) do number msg message_burst(orig) number nparts message_count_parts(msg) loop for number i 1, while i <= nparts, set i i + 1 do number part message_get_part(msg, i) number out open(sprintf('>%s%02d', stem, i)) message_to_stream(out, part) done message_close(msg) done
void
quarantine (string text)
¶Place the message to the quarantine queue, using text as explanatory reason.
number
callout_open (string url)
¶Opens connection to the callout server listening at url. Returns the descriptor of the connection.
void
callout_close (number fd)
¶Closes the connection. fd is the file descriptor returned by
the previous call to callout_open
.
number
callout_do (number fd, string email [, string rest])
¶Instructs the callout server identified by fd (a file descriptor
returned by a previous call to callout_open
) to verify the
validity of the email. Optional rest argument supplies
additional parameters for the server. It is formatted as a sequence
of parameter settings delimited by whitespaces. Each setting is
a parameter name and value separated by a ‘=’ sign.
See callout parameters, for a discussion of available callout
parameters.
Possible return values:
0
Success. The email is found to be valid.
e_not_found
email does not exist.
e_temp_failure
The email validity cannot be determined right now, e.g. because remote SMTP server returned temporary failure. The caller should retry verification later.
e_failure
Some error occurred.
The function will throw the e_callout_proto
exception if the
remote host doesn’t speak the correct callout protocol.
Upon return, callout_do
modifies the following variables:
last_poll_host
Host name or IP address of the last polled SMTP server.
last_poll_greeting
Initial SMTP reply from the last polled host.
last_poll_helo
The reply to the HELO
(EHLO
) command, received from the
last polled host.
last_poll_sent
Last SMTP command sent to the polled host. If nothing was
sent, last_poll_sent
contains the string ‘nothing’.
last_poll_recv
Last SMTP reply received from the remote host. In case of multi-line replies, only the first line is stored. If nothing was received the variable contains the string ‘nothing’.
The default callout server is defined by the callout-url
statement in the configuration file, or by the callout
statement in the server milter
section (see configuring default callout server. The following functions operate on that server.
string
default_callout_server_url ()
¶Returns URL of the default callout server.
number
callout (string email)
¶Verifies the validity of the email using the default callout server.
The following functions are wrappers over the callout functions described in the previous section. They are provided for backward compatibility.
These functions are defined in the module poll.mfl, which you must require prior to using any of them.
boolean
_pollhost (string ip, string email, string domain, string mailfrom)
¶Poll SMTP host ip for email address email,
using domain as EHLO
domain and mailfrom as
MAIL FROM
. Returns 0 or 1 depending on the result of the test.
In contrast to the strictpoll
function, this function does not use
cache database and does not fall back to polling MX servers if the
main poll tempfails. The function can throw one of the following
exceptions: e_failure
, e_temp_failure
.
boolean
_pollmx (string ip, string email, string domain, string mailfrom)
¶Poll MXs of the domain for email address email, using
domain as EHLO
domain and mailfrom as MAIL
FROM
address. Returns 0 or 1 depending on the result of the test.
In contrast to the stdpoll
function, _pollmx
does
not use cache database and does not fall back to polling the ip
if the poll fails. The function can throw one of the following
exceptions: e_failure
, e_temp_failure
.
boolean
stdpoll (string email, string domain, string mailfrom)
¶Performs standard poll for email, using domain as
EHLO
domain and mailfrom as MAIL FROM
address.
Returns 0 or 1 depending on the result of the test. Can raise one of
the following exceptions: e_failure
, e_temp_failure
.
In on
statement context, it is synonymous to poll
without explicit host.
boolean
strictpoll (string host, string email, string domain, string mailfrom)
¶Performs strict poll for email on host host.
See the description of stdpoll
for the detailed information.
In on
context, it is synonymous to poll host host
.
The mailfrom argument can be a comma-separated list of email addresses, which can be useful for servers that are unusually picky about sender addresses. It is advised, however, that this list always contain the ‘<>’ address. For example:
_pollhost($client_addr, $f, "domain", "postmaster@my.net,<>")
See also mail-from-address.
Before returning, all described functions set the following built-in variables:
Variable | Contains |
---|---|
last_poll_host | Host name or IP address of the last polled host. |
last_poll_sent | Last SMTP command, sent to this host. If nothing was sent, it contains literal string ‘nothing’. |
last_poll_recv | Last SMTP reply received from this host. In case of multi-line replies, only the first line is stored. If nothing was received the variable contains the string ‘nothing’. |
cache_used | 1 if cached data were used instead of
polling, 0 otherwise. This variable is set by stdpoll
and strictpoll . If it equals 1 , none of the above
variables are modified. See cache_used example, for an example. |
Table 5.1: Variables set by polling functions
Following functions operate on IPv4 addresses and CIDRs.
number
ntohl (number n)
¶Converts the number n, from host to network byte order. The argument n is treated as an unsigned 32-bit number.
number
htonl (number n)
¶Converts the number n, from network to host byte order. The argument n is treated as an unsigned 32-bit number.
number
ntohs (number n)
¶The argument n is treated as an unsigned 16-bit number. The function converts this number from network to host order.
number
htons (number n)
¶The argument n is treated as an unsigned 16-bit number. The function converts this number from host to network order.
number
inet_aton (string s)
¶Converts the Internet host address s from the standard numbers-and-dots notation into the equivalent integer in host byte order.
inet_aton("127.0.0.1") ⇒ 2130706433
The numeric data type in MFL is signed, therefore on machines with 32 bit integers, this conversion can result in a negative number:
inet_aton("255.255.255.255") ⇒ -1
However, this does not affect arithmetical operations on IP addresses.
string
inet_ntoa (number n)
¶Converts the Internet host address n, given in host byte order to string in standard numbers-and-dots notation:
inet_ntoa(2130706433) ⇒ "127.0.0.1"
number
len_to_netmask (number n)
¶Convert number of masked bits n to IPv4 netmask:
inet_ntoa(len_to_netmask(24)) ⇒ 255.255.255.0 inet_ntoa(len_to_netmask(7)) ⇒ 254.0.0.0
If n is greater than 32 the function raises e_range
exception.
number
netmask_to_len (number mask)
¶Convert IPv4 netmask mask into netmask length (number of bits preserved by the mask):
netmask_to_len(inet_aton("255.255.255.0")) ⇒ 24 netmask_to_len(inet_aton("254.0.0.0")) ⇒ 7
boolean
match_cidr (string ip, string cidr)
¶This function is defined in the module match_cidr.mfl (see Modules).
It returns true
if the IP address ip pertains to the
IP range cidr. The first argument, ip, is a string
representation of an IP address. The second argument, cidr, is
a string representation of a IP range in CIDR notation, i.e.
"A.B.C.D/N"
, where A.B.C.D is an IPv4
address and N specifies the prefix length – the number of
shared initial bits, counting from the left side of the address.
The following example will reject the mail if the IP address of
the sending machine does not belong to the block 10.10.1.0/19
:
if not match_cidr(${client_addr}, "10.10.1.0/19") reject fi
MFL offers two sets of functions for querying the Domain Name
System. The dns_query
function and associated
dns_reply_
functions provide a generalized DNS API.
Other functions provide a simplified API.
number
dns_query (number type, string domain; number sort, number resolve)
¶This function looks up the domain name name. The type argument specifies type of the query to perform. On success, the function returns DNS reply descriptor, a non-negative integer number identifying the reply. It can then be passed to any of the ‘dns_reply_’ functions discussed below in order to retrieve the information from it.
If no matching records were found, the function returns ‘-1’.
On error, it throws a corresponding exception.
The type argument is one of the following constants (defined in the module ‘dns’):
Query the ‘A’ record. The domain should be the hostname to look up.
Query the ‘NS’ records.
Query the ‘PTR’ record. The domain address should be the IP address in dotted-quad form.
Query the ‘MX’ records.
Query the ‘TXT’ records.
If the query returns multiple RR sets, the optional argument sort controls whether they should be returned in the same order as obtained from the DNS (0, the default), or should be sorted (1).
The optional argument resolve is consulted for type values
DNS_TYPE_MX
and DNS_TYPE_NS
. By default, queries for
these types return hostnames. If resolve is set, such queries
return IP addresses instead.
To extract actual data from the dns_query
return value, use the
functions dns_reply_count
and dns_reply_string
. The
usual processing sequence is:
require dns # Send the query and save the reply descriptor set n dns_query(DNS_TYPE_NS, domain_name) if n >= 0 # If non-empty set is returned, iterate over each value in it: loop for set i 0, while i < dns_reply_count(n), set i i + 1 do # Get the actual data: echo dns_reply_string(n, i) done # Release the memory associated with the reply. dns_reply_release(n) fi
void
dns_reply_release (number rd)
¶Release the memory associated with the reply rd. If rd is -1, the function does nothing.
number
dns_reply_count (number rd)
¶Return the number of records in the reply rd. For convenience, if rd is -1, the function returns 0. If rd is negative (excepting -1), a ‘e_failure’ exception is thrown.
string
dns_reply_string (number rd, number n)
¶Returns nth record from the DNS reply rd.
number
dns_reply_ip (number rd, number n)
¶Returns nth record from the DNS reply rd, if the reply contains IP addresses.
These functions are implemented in two layers: primitive built-in functions which raise exceptions if the lookup fails, and library calls that are warranted to always return meaningful value without throwing exceptions.
The built-in layer is always available. The library calls become available after requesting the dns module (see Modules):
require dns
string
dns_getaddr (string domain)
¶Returns a whitespace-separated list of IP addresses (A
records) for domain.
string
dns_getname (string ipstr)
¶Returns a whitespace-separated list of domain names (PTR
records) for the IPv4 address ipstr.
string
getmx (string domain [, boolean ip])
¶Returns a whitespace-separated list of ‘MX’ names (if ip is not
given or if it is 0
) or ‘MX’ IP addresses (if
ip!=0
)) for domain. Within the returned
string, items are sorted in order of increasing ‘MX’ priority.
If domain has no ‘MX’ records, an empty string is returned.
If the DNS query fails, getmx
raises an appropriate
exception.
Examples:
getmx("mafra.cz") ⇒ "smtp1.mafra.cz smtp2.mafra.cz relay.iol.cz" getmx("idnes.cz") ⇒ "smtp1.mafra.cz smtp2.mafra.cz relay.iol.cz" getmx("gnu.org") ⇒ "mx10.gnu.org mx20.gnu.org" getmx("org.pl") ⇒ ""
Notes:
getmx(domain)
can
differ from that obtained from getmx(domain, 1)
, e.g.:
getmx("aol.com") ⇒ mailin-01.mx.aol.com mailin-02.mx.aol.com mailin-03.mx.aol.com mailin-04.mx.aol.com getmx("aol.com", 1) ⇒ 64.12.137.89 64.12.137.168 64.12.137.184 64.12.137.249 64.12.138.57 64.12.138.88 64.12.138.120 64.12.138.185 205.188.155.89 205.188.156.185 205.188.156.249 205.188.157.25 205.188.157.217 205.188.158.121 205.188.159.57 205.188.159.217
dns_query
.
If you intend to iterate over returned values, better use
dns_query
directly, e.g. instead of doing
string_list_iterate(getmx(domain), ` ', MX, `do_something(MX)')
use
set n dns_query(DNS_TYPE_MX, domain) if n >= 0 loop for set i 0, while i < dns_reply_count(n), set i i + 1 do do_something(dns_reply_string(n, i)) done dns_reply_release(n) fi
See dns_query, for details about the dns_query
function and
associated dns_reply_*
calls.
It will most probably be removed in future releases, when array data types are implemented.
boolean
primitive_hasmx (string domain)
¶Returns true
if the domain name given by its argument
has any ‘MX’ records.
If the DNS query fails, this function throws failure
or
temp_failure
.
boolean
hasmx (string domain)
¶Returns true
if the domain name given by its argument
has any ‘MX’ records.
Otherwise, if domain has no ‘MX’s or if the DNS query fails,
hasmx
returns false
.
string
primitive_hostname (string ip)
¶The ip argument should be a string representing an IP address in dotted-quad notation. The function returns the canonical name of the host with this IP address obtained from DNS lookup. For example
primitive_hostname (${client_addr})
returns the fully qualified domain name of the host represented by Sendmail variable ‘client_addr’.
If there is no ‘PTR’ record for ip, primitive_hostname
raises the exception e_not_found
.
If DNS query fails, the function raises failure
or
temp_failure
, depending on the character of the failure.
string
hostname (string ip)
¶The ip argument should be a string representing an IP address in dotted-quad notation. The function returns the canonical name of the host with this IP address obtained from DNS lookup.
If there is no ‘PTR’ record for ip, or if the lookup fails, the function returns ip unchanged.
The previous mailfromd
versions used the following
paradigm to check if an IP address resolves:
if hostname(ip) != ip ...
boolean
primitive_ismx (string domain, string host)
¶The domain argument is any valid domain name, the host is a host name or IP address.
The function returns true
if host is one of the ‘MX’
records for the domain.
If domain has no ‘MX’ records, primitive_ismx
raises
exception e_not_found
.
If DNS query fails, the function raises failure
or
temp_failure
, depending on the character of the failure.
boolean
ismx (string domain, string host)
¶The domain argument is any valid domain name, the host is a host name or IP address.
The function returns true
if host is one of the ‘MX’
records for the domain. Otherwise it returns false
.
If domain has no ‘MX’ records, or if the DNS query fails, the
function returns false
.
string
primitive_resolve (string host, [string domain])
¶Reverse of primitive_hostname
. The primitive_resolve
function
returns the IP address for the host name specified by host
argument. If host has no A records, the function raises the
exception e_not_found
.
If DNS lookup fails, the function raises failure
or
temp_failure
, depending on the character of the failure.
If the optional domain argument is given, it will be appended to host (with an intermediate dot), before querying the DNS. For example, the following two expressions will return the same value:
primitive_resolve("puszcza.gnu.org.ua") primitive_resolve("puszcza", "gnu.org.ua")
There is a considerable internal difference between one-argument and
two-argument forms of primitive_resolve
: the former
queries DNS for an ‘A’ record, whereas the latter
queries it for any record matching host in the domain
domain and then selects the most appropriate one. For example,
the following two calls are equivalent:
primitive_hostname("213.130.0.22") primitive_resolve("22.0.130.213", "in-addr.arpa")
This makes it possible to use primitive_resolve
for querying
DNS black listing domains. See match_dnsbl, for a
working example of this approach. See also match_rhsbl, for
another practical example of the use of the two-argument form.
string
resolve (string host, [string domain])
¶Reverse of hostname
. The resolve
function
returns IP address for the host name specified by host
argument. If the host name cannot be resolved, or a DNS failure
occurs, the function returns ‘"0"’.
This function is entirely equivalent to primitive_resolve
(see above), except that it never raises exceptions.
string
ptr_validate (string ip)
¶Tests whether the DNS reverse-mapping for ip exists and correctly points to a domain name within a particular domain.
First, it obtains all PTR records for ip. Then, for each record returned, a look up for A records is performed and IP addresses of each record are compared against ip. The function returns true if a matching A record is found.
boolean
primitive_hasns (string domain)
¶Returns ‘True’ if the domain domain has at least one ‘NS’ record. Throws exception if DNS lookup fails.
boolean
hasns (string domain)
¶Returns ‘True’ if the domain domain has at least one ‘NS’ record. Returns ‘False’ if there are no ‘NS’ records or if the DNS lookup fails.
string
getns (string domain ; boolean resolve, boolean sort)
¶Returns a whitespace-separated list of all the ‘NS’ records for the domain domain. Optional parameters resolve and sort control the formatting. If resolve is 0 (the default), the resulting string will contain IP addresses of the NS servers. If resolve is not 0, hostnames will be returned instead. If sort is 1, the returned items will be sorted.
If the DNS query fails, getns
raises an appropriate
exception.
Notes:
dns_query
.
If you intend to iterate over returned values, better use
dns_query
directly, e.g. instead of doing
string_list_iterate(getns(domain), ` ', NS, `do_something(NS)')
use
set n dns_query(DNS_TYPE_NS, domain) if n >= 0 loop for set i 0, while i < dns_reply_count(n), set i i + 1 do do_something(dns_reply_string(n, i)) done dns_reply_release(n) fi
See dns_query, for details about the dns_query
function and
associated dns_reply_*
calls.
It will most probably be removed in future releases, when array data types are implemented.
The geolocation functions allow you to identify the country where
the given IP address or host name is located. These functions are
available only if the libmaxminddb
library is installed and
mailfromd
is compiled with the ‘GeoIP2’ support.
The libmaxminddb
library is distributed by ‘MaxMind’ under
the terms of the Apache License Version 2.0. It is available
from https://dev.maxmind.com/geoip/geoip2/downloadable/#MaxMind_APIs.
void
geoip2_open (string filename)
¶Opens the geolocation database file filename. The database must be in GeoIP2 format.
If the database cannot be opened, geoip2_open
throws the
e_failure
exception.
If this function is not called, geolocation functions described below will try to open the database file ‘/usr/share/GeoIP/GeoLite2-City.mmdb’.
string
geoip2_dbname (void)
¶Returns the name of the geolocation database currently in use.
The geolocation database for each IP address, which serves as a look up key, stores a set of items describing this IP. This set is organized as a map of key-value pairs. Each key is a string value. A value can be a scalar, another map or array of values. Using JSON notation, the result of a look up in the database might look as:
{ "country":{ "geoname_id":2921044, "iso_code":"DE", "names":{ "en": "Germany", "de": "Deutschland", "fr":"Allemagne" }, }, "continent":{ "code":"EU", "geoname_id":6255148, "names":{ "en":"Europe", "de":"Europa", "fr":"Europe" } }, "location":{ "accuracy_radius":200, "latitude":49.4478, "longitude":11.0683, "time_zone":"Europe/Berlin" }, "city":{ "geoname_id":2861650, "names":{ "en":"Nuremberg", "de":"Nürnberg", "fr":"Nuremberg" } }, "subdivisions":[{ "geoname_id":2951839, "iso_code":"BY", "names":{ "en":"Bavaria", "de":"Bayern", "fr":"Bavière" } } }
Each particular data item in such structure is identified by its
search path, which is a dot-delimited list of key names leading
to that value. For example, using the above map, the name of the city
in English can be retrieved using the key city.names.en
.
string
geoip2_get (string ip, string path)
¶Looks up the IP address ip in the geolocation database. If found, returns data item identified by the search path path.
The function can throw the following exceptions:
The ip was not found in the database.
The path does not exist the returned map.
General error occurred. E.g. the database cannot be opened, ip is not a valid IP address, etc.
string
geoip2_get_json (string ip [; number indent)
¶Looks up the ip in the database and returns entire data set associated with it, formatted as a JSON object. If the optional parameter indent is supplied and is greater than zero, it gives the indentation for each nesting level in the JSON object.
Applications may test whether the GeoIP2 support is present and
enable the corresponding code blocks conditionally, by testing if
the ‘WITH_GEOIP2’ m4 macro is defined. For example, the
following code adds to the message the ‘X-Originator-Country’
header, containing the 2 letter code of the country where the client
machine is located. If mailfromd
is compiled without
the ‘GeoIP2’ support, it does nothing:
m4_ifdef(`WITH_GEOIP2',` try do header_add("X-Originator-Country", geoip2_get($client_addr, 'country.iso_code')) done catch e_not_found or e_range do pass done ')
The functions described below provide a user interface to DBM databases.
Each DBM database is a separate disk file that keeps key/value pairs. The interface allows to retrieve the value corresponding to a given key. Both ‘key’ and ‘value’ are null-terminated character strings. To lookup a key, it is important to know whether its length includes the terminating null byte. By default, it is assumed that it does not.
Another important database property is the file mode of the database file. The default file mode is ‘640’ (i.e. ‘rw-r----’, in symbolic notation).
These and other properties can be configured using the dbprop
pragma:
#pragma dbprop pattern prop [prop]
The pattern is the database name or shell-style globbing
pattern. Properties defined by that pragma apply to each database
whose name matches this pattern. If several dbprop
pragmas
match the database name, the one that matches exactly is preferred.
The rest of arguments define properties for that database. The valid values for prop are:
Setting ‘null’ property is necessary, for databases created with
makemap -N hash
command.
640 rw-r----
For example, consider the following pragmas:
#pragma dbprop /etc/mail/whitelist.db 640
It tells that the database file whitelist.db has privileges ‘640’ and do not include null in the key length.
Similarly, the following pragma:
#pragma dbprop `/etc/mail/*.db' null 600 bdb://
declares that all database files in directory /etc/mail are
Berkeley DB files, have privileges ‘640’, and include null terminator
in the key length. Notice, the use of m4
quoting
characters in the example below. Without them, the sequence ‘/*’
would have been taken as the beginning of a comment.
Additionally, for compatibility with previous versions (up to 5.0), the terminating null property can be requested via an optional argument to the database functions (in description below, marked as null).
boolean
dbmap (string db, string key, [boolean null])
¶Looks up key in the DBM file db and returns
true
if it is found.
See above for the meaning of null.
See whitelisting, for an example of using this function.
string
dbget (string db, string key [, string default, boolean null])
¶Looks up key in the database db and returns the value associated with it. If the key is not found returns default, if specified, or empty string otherwise.
See above for the meaning of null.
void
dbput (string db, string key, string value [, boolean null, number mode ])
¶Inserts in the database a record with the given key and value. If a record with the given key already exists, its value is replaced with the supplied one.
See above for the meaning of null. Optional mode allows
to explicitly specify the file mode for this database. See also
#pragma dbprop
, described above.
void
dbinsert (string db, string key, string value [, boolean replace, boolean null, number mode ])
¶This is an improved variant of dbput
, which provides a
better control on the actions to take if the key already exists in the
database. Namely, if replace is ‘True’, the old value is
replaced with the new one. Otherwise, the ‘e_exists’ exception
is thrown.
void
dbdel (string db, string key [, boolean null, number mode])
¶Delete from the database the record with the given key. If there are no such record, return without signalling error.
If the optional null argument is given and is not zero, the terminating null character will be included in key length.
Optional mode allows to explicitly specify the file mode for
this database. See also #pragma dbprop
, described above.
The functions above have also the corresponding exception-safe interfaces, which return cleanly if the ‘e_dbfailure’ exception occurs. To use these interfaces, request the safedb module:
require safedb
The exception-safe interfaces are:
number
safedbmap (string db, string key [, number default, boolean null])
¶This is an exception-safe interface to dbmap
. If a
database error occurs while attempting to retrieve the record,
safedbmap
returns default or ‘0’, if it is
not defined.
string
safedbget (string db, string key [, string default, boolean null])
¶This is an exception-safe interface to dbget
. If a
database error occurs while attempting to retrieve the record,
safedbget
returns default or empty string, if it is
not defined.
void
safedbput (string db, string key, string value [, boolean null])
¶This is an exception-safe interface to dbput
. If a
database error occurs while attempting to retrieve the record,
the function returns without raising exception.
void
safedbdel (string db, string key [, boolean null])
¶This is an exception-safe interface to dbdel
. If a
database error occurs while attempting to delete the record,
the function returns without raising exception.
The verbosity of ‘safedb’ interfaces in case of database error is
controlled by the value of safedb_verbose
variable. If it is
‘0’, these functions return silently. This is the default
behavior. Otherwise, if safedb_verbose
is not ‘0’, these
functions log the detailed diagnostics about the database error and
return.
The following functions provide a sequential access to the contents of a DBM database:
number
dbfirst (string name)
¶Start sequential access to the database name. The return value is an opaque identifier, which is used by the remaining sequential access functions. This number is ‘0’ if the database is empty.
number
dbnext (number dn)
¶Select next record form the database. The argument dn is the
access identifier, returned by a previous call to dbfirst
or
dbnext
.
Returns new access identifier. This number is ‘0’ if all records in the database have been visited.
The usual approach for iterating over all records in a database dbname is:
loop for number dbn dbfirst(dbname) do ... done while dbnext(dbn)
The following two functions can be used to access values of the
currently selected database record. Their argument, dn, is the
access identifier, returned by a previous call to dbfirst
or
dbnext
.
string
dbkey (number dn)
¶Return the key from the selected database record.
string
dbvalue (number dn)
¶Return the value from the selected database record.
void
dbbreak (number dn)
¶Stop sequential access to the database and deallocate all associated resources. Use this function if you need to break from the sequential access loop, as in the example below:
loop for number dbn dbfirst(dbname) do if some_condition dbbreak(dbn) break fi done while dbnext(dbn)
number
db_expire_interval (string fmt)
¶The fmt argument is a database format identifier
(see Database Formats). If it is valid, the function returns the
expiration interval for that format. Otherwise, db_expire_interval
raises the
e_not_found
exception.
string
db_name (string fmtid)
¶The fmtid argument is a database format identifier
(see Database Formats). The function returns the file name
for that format. If fmtid does not match any known format,
db_name
raises the e_not_found
exception.
number
db_get_active (string fmtid)
¶Returns the flag indicating whether the cache database fmtid
is currently enabled. If fmtid does not match any known format,
db_name
raises the e_not_found
exception.
void
db_set_active (string fmtid, boolean enable)
¶Enables the cache database fmtid if enable is ‘True’, or disables it otherwise. For example, to disable DNS caching, do:
db_set_active("dns", 0)
boolean
relayed (string domain)
¶Returns true
if the string domain is found in one of
relayed domain files (see relayed-domain-file). The
usual construct is:
if relayed(hostname(${client_addr})) ...
which yields true
if the IP address from Sendmail
variable
‘client_addr’ is relayed by the local machine.
Control database is a DBM file whose records define actions to be taken for mails coming from particular IP or email addresses. Functions and variables for using the control database are defined in module cdb.mfl.
string
cdb_name ¶Name of the database file to use as control database. By default it is /etc/mail/mfctl.db.
number
cdb_greylist_interval ¶Greylisting interval, for database records that prescribe greylisting (see below). Defaults to 900 seconds.
void
cdb_check (string prefix, string addr)
¶Perform a look up in the database. If the value is found, take the action it indicates.
The key to use for the look up depends on the value of prefix:
The addr argument must be an IP address. The look up key is ‘ip:addr’.
The addr argument is an email address. The key is ‘email:cemail’, where cemail is addr in canonical form, obtained by converting its domain part to lower case.
The addr argument is an email address. The key is formed as ‘domain:dom’, where dom is domain part of addr converted to lower case.
Initial key value is obtained as for ‘domain’. If the key is found, the requested action is performed. Otherwise, the shortest hostname prefix (sequence of characters up to and including the first dot) is stripped off the domain and the process is retried.
Thus, the action is determined by the longest subdomain of addr, x, for which the key ‘domain:x’ exists in the database.
If addr is an email address, its domain part is selected, otherwise it is used as is. The list of MX servers for this domain is obtained. For each host from that list the key ‘mx:host’ is looked up. First entry found is used.
The function cdb_check
returns if the key was not found in the
database or if the value found was ‘OK’ (case-insensitive) or
empty. Otherwise, the looked up value determines the action
(see reply actions), as described in the following table. The
action is returned and execution of the filter program stops.
Continue to the next milter state. See continue.
Accept the mail. See accept.
Reject the mail. See reject.
Return a temporary failure. See tempfail.
Greylist the mail using the interval defined by the
cdb_greylist_interval
variable. If called in envrcpt
handler, the action is taken immediately. Otherwise, if called
in connect
, helo
or envfrom
, the action is
delayed until envrcpt
is invoked. Otherwise, if called
in any other handler an error is reported.
Here, code is SMTP response code, xcode is extended STMP response code, and text is an arbitrary text. If code is given, it must begin with ‘4’ or ‘5’. Its first digit defines the action to be taken: temporary failure (if ‘4’) or reject (if ‘5’). If xcode is given, its first digit must match that of code.
If only text is supplied, it is equivalent to
reject(550, 5.1.0, text)
boolean
access (string pathname, number mode)
¶Checks whether the calling process can access the file pathname. If pathname is a symbolic link, it is dereferenced. The function returns ‘True’ if the file can be accessed and ‘False’ otherwise26.
Symbolic values for mode are provided in module status:
string
getenv (string name)
¶Searches the environment list for the variable name and returns its value. If the variable is not defined, the function raises the exception ‘e_not_found’.
string
gethostname ([bool fqn])
¶Return the host name of this machine.
If the optional fqn is given and is ‘true’, the function will attempt to return fully-qualified host name, by attempting to resolve it using DNS.
string
getdomainname ()
¶Return the domain name of this machine. Note, that it does not necessarily coincide with the actual machine name in DNS.
Depending on the underlying ‘libc’ implementation, this call may
return empty string or the string ‘(none)’. Do not rely on it to
get the real domain name of the box mailfromd
runs on, use
localdomain
(see below) instead.
string
localdomain ()
¶Return the local domain name of this machine.
This function first uses getdomainname
to make a first guess.
If it does not return a meaningful value, localdomain
calls
gethostname(1)
to determine the fully qualified host name of
the machine, and returns its domain part.
To use this function, require the localdomain module
(see Modules), e.g.: require localdomain
.
number
time ()
¶Return the time since the Epoch (00:00:00 UTC, January 1, 1970), measured in seconds.
string
strftime (string fmt, number timestamp)
¶string
strftime (string fmt, number timestamp, boolean gmt)
¶Formats the time timestamp (seconds since the Epoch) according
to the format specification format. Ordinary characters placed
in the format string are copied to the output without conversion.
Conversion specifiers are introduced by a ‘%’ character.
See Time and Date Formats, for a detailed description of the
conversion specifiers. We recommend using single quotes
around fmt to prevent ‘%’ specifiers from being interpreted
as Mailfromd
variables (See Literals, for a
discussion of quoted literals and variable interpretation within
them).
The timestamp argument can be a return value of time
function (see above).
For example:
strftime('%Y-%m-%d %H:%M:%S %Z', 1164477564) ⇒ 2006-11-25 19:59:24 EET strftime('%Y-%m-%d %H:%M:%S %Z', 1164477564, 1) ⇒ 2006-11-25 17:59:24 GMT
string
uname (string format)
¶This function returns system information formatted according to the format specification format. Ordinary characters placed in the format string are copied to the output without conversion. Conversion specifiers are introduced by a ‘%’ character. The following conversions are defined:
Name of this system.
Name of this node within the communications network to which this node is attached. Note, that it does not necessarily coincide with the actual machine name in DNS.
Kernel release.
Kernel version.
Name of the hardware type on which the system is running.
For example:
uname('%n runs %s, release %r on %m') ⇒ "Trurl runs Linux, release 2.6.26 on i686"
Notice the use of single quotes.
void
unlink (string name)
¶Unlinks (deletes) the file name. On error, throws the
e_failure
exception.
number
system (string str)
¶The function system
executes a command specified in str
by calling /bin/sh -c string
, and returns -1 on error or
the return status of the command otherwise.
void
sleep (number secs[, usec])
¶Sleep for secs seconds. If optional usec argument is given, it specifies additional number of microseconds to wait for. For example, to suspend execution of the filter for 1.5 seconds:
sleep(1,500000)
This function is intended mostly for debugging and experimental purposes.
number
umask (number mask)
¶Set the umask to mask & 0777. Return the previous value of the mask.
string
getpwnam (string name)
¶string
getpwuid (number uid)
¶Look for the user name (getpwnam
) or user ID uid
(getpwuid
) in the system password database and return the
corresponding record, if found. If not found, raise the
‘e_not_found’ exception.
The returned record consists of six fields, separated by colon sign:
uname:passwd:uid:gid:gecos:dir:shell
Field | Meaning |
---|---|
uname | user name |
passwd | user password |
uid | user ID |
gid | group ID |
gecos | real name |
dir | home directory |
shell | shell program |
For example:
getpwnam("gray") ⇒ "gray:x:1000:1000:Sergey Poznyakoff:/home/gray:/bin/bash"
Following two functions can be used to test for existence of a key in the user database:
boolean
mappwnam (string name)
¶boolean
mappwuid (number uid)
¶Return ‘true’ if name (or uid) is found in the system user database.
‘Sieve’ is a powerful mail filtering language, defined in
RFC 3028. Mailfromd
supports an extended form
of this language. For a description of the language and available
extensions, see Sieve Language in GNU Mailutils Manual.
boolean
sieve (number msg, string script [, number flags, string file, number line])
¶Compile the Sieve program script and execute it over the message identified by the descriptor nmsg.
Optional flags modify the behavior of the function. It is a
bit-mask field, consisting of a bitwise or
of one or more of
the following flags, defined in sieve.mfl:
MF_SIEVE_FILE
¶The script argument specifies the name of a Sieve program file. This is the default.
MF_SIEVE_TEXT
¶The script argument is a string containing entire Sieve program. Optional arguments file and line can be used to fix source locations in Sieve diagnostic messages (see below).
MF_SIEVE_LOG
¶Log every executed ‘Sieve’ action.
MF_SIEVE_DEBUG_TRACE
¶Trace execution of ‘Sieve’ tests.
MF_SIEVE_DEBUG_INSTR
¶Log every instruction, executed in the compiled ‘Sieve’ code. This produces huge amounts of output and is rarely useful, unless you suspect some bug in ‘Sieve’ implementation and wish to trace it.
For example, MF_SIEVE_LOG|MF_SIEVE_DEBUG_TRACE
enables logging
‘Sieve’ actions and tests.
The sieve
function returns true
if the message was
accepted by the script program, and false
otherwise.
Here, the word accepted means that some form of ‘KEEP’
action (see Actions in GNU Mailutils
Manual) was executed over the message.
While executing the Sieve script, Sieve environment (RFC 5183) is initialized as follows:
The domain name of the server Sieve is running on.
Host name of the server Sieve is running on.
The string ‘MTA’.
The string ‘GNU Mailutils’.
The string ‘pre’.
Defined to the value of ‘client_ptr’ macro, if it was required.
Defined to the value of ‘client_addr’ macro, if it was required.
The version of GNU Mailutils.
The following example discards each message not accepted by the ‘Sieve’ program /etc/mail/filter.siv:
require 'sieve' group eom do if not sieve(current_message(), "/etc/mail/filter.siv", MF_SIEVE_LOG) discard fi done
The Sieve program can be embedded in the MFL filter, as shown in the example below:
require 'sieve' prog eom do if not sieve(current_message(), "require \"fileinto\";\n" "fileinto \"/tmp/sieved.mbox\";", MF_SIEVE_TEXT | MF_SIEVE_LOG) discard fi done
In such cases, any Sieve diagnostics (error messages, traces, etc.)
will be marked with the locations relative to the line where the call
to sieve
appears. For example, the above program produces the
following in the log:
prog.mfl:7: FILEINTO; delivering into /tmp/sieved.mbox
Notice, that the line number correctly refers to the line where the
fileinto
action appears in the source. However, there are
cases where the reported line number is incorrect. This happens,
for instance, if script is a string variable defined elsewhere.
To handle such cases, sieve
accepts two optional parameters
which are used to compute the location in the Sieve program. The
file parameter specifies the file name where the definition of
the program appears, and the line parameter gives the number of
line in that file where the program begins. For example:
require 'sieve' const sieve_prog_line __line__ + 2 string sieve_prog <<EOT require "fileinto"; fileinto "/tmp/sieved.mbox"; EOT prog eom do if not sieve(current_message(), sieve_prog, MF_SIEVE_TEXT | MF_SIEVE_LOG, __file__, sieve_prog_line) discard fi done
The actual Sieve program begins two lines below the
sieve_prog_line
constant definition, which is reflected in its
initialization.
A set of functions is defined for interfacing with other filters via
TCP. Currently implemented are interfaces with
SpamAssassin
spamd
daemon and with
ClamAV
anti-virus.
Both interfaces work much the same way: the remote filter is
connected and the message is passed to it. If the remote filter
confirms that the message matches its requirements, the function
returns true
. Notice that in practice that means that such a
message should be rejected or deferred.
The address of the remote filter is supplied as the second argument in the form of a standard URL:
proto://path[:port]
The proto part specifies the connection protocol. It should be ‘tcp’ for the TCP connection and ‘file’ or ‘socket’ for the connection via UNIX socket. In the latter case the proto part can be omitted. When using TCP connection, the path part gives the remote host name or IP address and the optional port specifies the port number or service name to use. For example:
# connect to ‘remote.filter.net’ on port 3314: tcp://remote.filter.net:3314 # the same, using symbolic service name (must be defined in # /etc/services): tcp://remote.filter.net:spamd # Connect via a local UNIX socket (equivalent forms): /var/run/filter.sock file:///var/run/filter.sock socket:///var/run/filter.sock
The description of the interface functions follows.
boolean
spamc (number msg, string url, number prec, number command)
¶Send the message msgt to the SpamAssassin daemon (spamd
)
listening on the given url. The command argument
identifies what kind of processing is needed for the message. Allowed
values are:
Process the message and return 1 or 0 depending on whether it is
diagnosed as spam or not. Store SpamAssassin keywords in the global
variable sa_keywords
(see below).
Process the message and return 1 or 0 depending on whether it is
diagnosed as spam or not. Store entire SpamAssassin report in the
global variable sa_keywords
.
Learn the supplied message as spam.
Learn the supplied message as ham.
Forget any prior classification of the message.
The second argument, prec, gives the precision, in decimal
digits, to be used when converting SpamAssassin diagnostic data and
storing them into mailfromd
variables.
The floating point SpamAssassin data are converted to the integer
mailfromd
variables using the following relation:
var = int(sa-var * 10**prec)
where sa-var stands for the SpamAssassin value and var
stands for the corresponding mailfromd
one. int()
means taking the integer part and ‘**’ denotes the exponentiation
operator.
The function returns additional information via the following variables:
sa_score
¶The spam score, converted to integer as described above. To convert
it to a floating-point representation, use sa_format_score
function (see sa_format_score). See also the
example below.
sa_threshold
¶The threshold, converted to integer form.
sa_keywords
¶If command is ‘SA_SYMBOLS’, this variable contains a string of comma-separated SpamAssassin keywords identifying this message, e.g.:
ADVANCE_FEE_1,AWL,BAYES_99
If command is ‘SA_REPORT’, the value of this variable is a spam report message. It is a multi-line textual message, containing detailed description of spam scores in a tabular form. It consists of the following parts:
The words ‘Content preview’, followed by a colon and an excerpt of the message body.
It has the following form:
Content analysis details: (score points, max required)
where score and max are spam score and threshold in floating point.
The score table is formatted in three columns:
The score, as a floating point number with one decimal digit.
SpamAssassin rule name that contributed this score.
Textual description of the rule
The score table can be extracted from sa_keywords
using
sa_format_report_header
function (see sa_format_report_header), as illustrated in the example below.
The value of this variable is undefined if command is ‘SA_LEARN_SPAM’, ‘SA_LEARN_HAM’ or ‘SA_FORGET’.
The spamc
function can signal the following exceptions:
e_failure
if the connection fails, e_url
if the supplied
URL is invalid and e_range
if the supplied port number
is out of the range 1–65535.
An example of using this function:
prog eom do if spamc(current_message(), "tcp://192.168.10.1:3333", 3, SA_SYMBOLS) reject 550 5.7.0 "Spam detected, score %sa_score with threshold %sa_threshold" fi done
Here is a more advanced example:
prog eom do set prec 3 if spamc(current_message(), "tcp://192.168.10.1:3333", prec, SA_REPORT) add "X-Spamd-Status" "SPAM" else add "X-Spamd-Status" "OK" fi add "X-Spamd-Score" sa_format_score(sa_score, prec) add "X-Spamd-Threshold" sa_format_score(sa_threshold, prec) add "X-Spamd-Keywords" sa_format_report_header(sa_keywords) done
boolean
sa (string url, number prec; number command)
¶Additional interface to the spamc
function, provided for
backward compatibility. It is equivalent to
spamc(current_message(), url, prec, command)
If command is not supplied, ‘SA_SYMBOLS’ is used.
boolean
clamav (number msg, string url)
¶Pass the message msg to the ClamAV daemon at url. Return
true
if it detects a virus in it. Return virus name in
clamav_virus_name
global variable.
The clamav
function can signal the following exceptions:
e_failure
if failed to connect to the server, e_url
if
the supplied URL is invalid and e_range
if the
supplied port number is out of the range 1–65535.
An example usage:
prog eom do if clamav(current_message(), "tcp://192.168.10.1:6300") reject 550 5.7.0 "Infected with %clamav_virus_name" fi done
number
rate (string key, number sample-interval, [number min-samples, number threshold])
¶Returns the mail sending rate for key per sample-interval. Optional min-samples, if supplied, specifies the minimal number of mails needed to obtain the statistics. The default is 2. Optional threshold controls rate database updates. If the observed rate (per sample-interval seconds) is higher than the threshold, the hit counters for that key are not incremented and the database is not updated. Although the threshold argument is optional27, its use is strongly encouraged. Normally, the value of threshold equals the value compared with the return from rate, as in:
if rate("$f-$client_addr", rate_interval, 4, maxrate) > maxrate tempfail 450 4.7.0 "Mail sending rate exceeded. Try again later" fi
This function is a low-level interface. Instead of using it directly,
we advise to use the rateok
function, described below.
boolean
rateok (string key, number sample-interval, number threshold,
¶[number min-samples])
To use this function, require the rateok module
(see Modules), e.g.: require rateok
.
The rateok
function returns ‘True’ if the mail sending
rate for key, computed for the interval of sample-interval
seconds is less than the threshold. Optional min-samples
parameter supplies the minimal number of mails needed to obtain the
statistics. It defaults to 4.
See Sending Rate, for a detailed description of the rateok
and
its use. The interval
function (see interval) is often
used in the second argument to rateok
or rate
.
boolean
tbf_rate (string key, number cost, number sample-interval, number burst-size)
¶This function implements a classical token bucket filter algorithm.
Tokens are added to the bucket identified by the key at constant
rate of 1 token per sample-interval microseconds, to a maximum
of burst-size tokens. If no bucket is found for the specified key, a
new bucket is created and initialized to contain burst-size
tokens. If the bucket contains cost or more tokens, cost
tokens are removed from it and tbf_rate
returns ‘True’.
Otherwise, the function returns ‘False’.
For a detailed description of the Token Bucket Algorithm and its use to limit mail rates, see TBF.
boolean
greylist (string key, number interval)
¶Returns ‘True’ if the key is found in the greylist database
(controlled by database greylist
configuration file statement,
see Database Configuration). The argument interval gives the greylisting
interval in seconds. The function stores the number of seconds left to the end
of greylisting period in the global variable
greylist_seconds_left
. See Greylisting, for a detailed
explanation.
The function greylist
can signal e_dbfailure
exception.
boolean
is_greylisted (string key
¶Returns ‘True’ if the key is still greylisted. If
‘true’ is returned, the function also stores the number of
seconds left to the end of greylisting period in the global variable
greylist_seconds_left
.
This function is available only if Con Tassios implementation of greylisting is used. See greylisting types, for a discussion of available greylisting implementations. See Pragma greylist, for a way to switch to Con Tassios implementation.
boolean
portprobe (string host, [number port])
¶boolean
listens (string host, [number port])
¶Returns true
if the IP address or host name given by
host argument listens on the port number port (default
25).
This function is defined in the module portprobe.
boolean
validuser (string name)
¶Returns true
if authenticity of the user name is
confirmed using mailutils authentication system. See Local Account Verification, for more details.
boolean
valid_domain (string domain)
¶Returns true
if the domain name domain has a
corresponding A record or if it has any ‘MX’ records, i.e. if it
is possible to send mail to it.
To use this function, require the valid_domain module (see Modules):
require valid_domain
number
heloarg_test (string arg, string remote_ip, string local_ip)
¶Verify if an argument of ‘HELO’ (‘EHLO’) command is valid. To use this function, require the heloarg_test module (see Modules).
Arguments:
‘HELO’ (‘EHLO’) argument. Typically, the value of $s
Sendmail macro;
IP address of the remote client. Typically, the value of
$client_addr
Sendmail macro;
IP address of this SMTP server;
The function returns a number describing the result of the test, as described in the following table.
Code | Meaning |
---|---|
HELO_SUCCESS | arg successfully passes all tests. |
HELO_MYIP | arg is our IP address. |
HELO_IPNOMATCH | arg is an IP, but it does not match the remote party IP address. |
HELO_ARGNORESOLVE | arg is an IP, but it does not resolve. |
HELO_ARGNOIP | arg is in square brackets, but it is not an IP address. |
HELO_ARGINVALID | arg is not an IP address and does not resolve to one. |
HELO_MYSERVERIP | arg resolves to our server IP. |
HELO_IPMISMATCH | arg does not resolve to the remote client IP address. |
The mail sending functions are new interfaces, introduced in version 3.1.
The underlying mechanism for sending mail, called mailer, is specified by --mailer command line option. This global setting can be overridden using the last optional argument to a particular function. In any case, the mailer is specified in the form of a URL.
Mailer URL begins with a protocol specification.
Two protocol specifications are currently supported: ‘sendmail’
and ‘smtp’. The former means to use a
sendmail
-compatible program to send mails. Such a program
must be able to read mail from its standard input and must support the
following options:
Do not treat ‘.’ as message terminator.
Use addr as the address of the sender.
Get recipient addresses from the message.
These conditions are met by most existing MTA programs, such
as exim
or postfix
(to say nothing of
sendmail
itself).
Following the protocol specification is the mailer location, which is separated from it with a colon. For the ‘sendmail’ protocol, the mailer location sets the full file name of the Sendmail-compatible MTA binary, for example:
sendmail:/usr/sbin/sendmail
A special form of a sendmail URL, consisting of protocol
specification only (‘sendmail:’) is also allowed. It means
“use the sendmail binary from the _PATH_SENDMAIL
macro in your /usr/include/paths.h file”. This is the default
mailer.
The ‘smtp’ protocol means to use an SMTP server directly. In this case the mailer location consists of two slashes, followed by the IP address or host name of the SMTP server, and, optionally, the port number. If the port number is present, it is separated from the rest of URL by a colon. For example:
smtp://remote.server.net smtp://remote.server.net:24
void
send_mail (string msg [, string to, string from, string mailer])
¶Sends message msg to the email address to. The value of msg must be a valid RFC 2822 message, consisting of headers and body. Optional argument to can contain several email addresses. In this case the message will be sent to each recipient specified in to. If it is not specified, recipient addresses will be obtained from the message headers.
Other optional arguments are:
Sets the sender address. By default ‘<>’ is used.
The URL of the mailer to use
Sample usage:
set message <<- EOT Subject: Test message To: Postmaster <postmaster@gnu.org.ua> From: Mailfromd <devnull@gnu.org.ua> X-Agent: %__package__ (%__version__) Dear postmaster, This is to notify you that our /etc/mailfromd.mfl needs a revision. -- Mailfromd filter administrator EOT send_mail(message, "postmaster@gnu.org.ua")
void
send_text (string text, string headers [, string to, string from, string mailer])
¶A more complex interface to mail sending functions.
Mandatory arguments:
Text of the message to be sent.
Headers for the message.
Optional arguments:
Recipient email addresses.
Sender email address.
URL of the mailer to use.
The above example can be rewritten using send_text
as follows:
set headers << -EOT Subject: Test message To: Postmaster <postmaster@gnu.org.ua> From: Mailfromd <devnull@gnu.org.ua> X-Agent: %__package__ (%__version__) EOT set text <<- EOT Dear postmaster, This is to notify you that our /etc/mailfromd.mfl needs a revision. -- Mailfromd filter administrator EOT send_text(text, headers, "postmaster@gnu.org.ua")
void
send_message (number msg [string to, string from, string mailer])
¶Send the message identified by descriptor msg (see Message Functions).
Optional arguments are:
Recipient email addresses.
Sender email address.
URL of the mailer to use.
void
send_dsn (string to, string sender, string rcpt, string text [, string headers, string from, string mailer])
¶This is an experimental interface which will change in the future versions. It sends a message disposition notification (RFC 2298, RFC 1894), of type ‘deleted’ to the email address to. Arguments are:
Recipient email address.
Original sender email address.
Original recipient email address.
Notification text.
Optional arguments:
Message headers
Sender address.
URL of the mailer to use.
void
create_dsn (string sender, string rcpt, string text [, string headers, string from])
¶Creates DSN message and returns its descriptor. Arguments are:
Original sender email address.
Original recipient email address.
Notification text.
Message headers
Sender address.
The functions described in this subsection allow to check whether the given IP address is listed in certain black list DNS zone.
boolean
match_dnsbl (string address, string zone, string range)
¶This function looks up address in the DNS blacklist zone zone and checks if the return falls into the given range of IP addresses.
It is intended as a replacement for the Sendmail macros ‘dnsbl’ and ‘enhdnsbl’.
To use match_dnsbl
, require the match_dnsbl module
(see Modules).
Arguments:
IP address of the SMTP server to be tested.
FQDN of the DNSbl zone to test against.
The range of IP addresses in CIDR notation or the word ‘ANY’, which stands for ‘127.0.0.0/8’.
The function returns true
if dns lookup for address in
the zone dnsbl yields an IP that falls within the range,
specified by cidr. Otherwise, it returns false
.
This function raises the following exceptions: e_invip
if
address is invalid and e_invcidr
if cidr is invalid.
boolean
match_rhsbl (string email, string zone, string range)
¶This function checks if the IP address, corresponding to the domain part of email is listed in the RHS DNS blacklist zone zone, and if so, whether its record falls into the given range of IP addresses range.
It is intended as a replacement for the Sendmail macro ‘rhsbl’ by Derek J. Balling.
To use this function, require the match_rhsbl module (see Modules).
Arguments:
E-mail address, whose domain name should be tested (usually, it is
$f
)
Domain name of the RHS DNS blacklist zone.
The range of IP addresses in CIDR notation.
Sender Policy Framework, or SPF for short, is an
extension to SMTP protocol that allows to identify forged
identities supplied with the MAIL FROM
and HELO
commands. The framework is explained in detail in RFC 4408
(http://tools.ietf.org/html/rfc4408) and on the
SPF Project Site. The following
description is a short introduction only, and the users are encouraged
to refer to the original specification for the detailed description of
the framework.
The domain holder publishes an SPF record – a special
DNS resource record that contains a set of rules declaring
which hosts are, and are not, authorized to use a domain name for
HELO
and MAIL FROM
identities. This resource record is
usually of type TXT
.28
The MFL script can verify if the identity matches the
published SPF record by calling check_host
function
and analyzing its return code. The function can be called either in
helo
or in envfrom
handler. Its arguments are:
The IP address of the SMTP client that is emitting the mail.
Usually it is $client_addr
.
The domain that provides the sought-after authorization information;
Normally it is the domain portion of the MAIL FROM
or
HELO
identity.
The MAIL FROM
identity.
The HELO
identity.
The SMTP domain served by the local server.
The function returns a numeric result code. For convenience, all possible return values are defined as macros in module spf.mfl. The table below describes each value along with the recommended actions for it:
None
¶A result of None
means that no records were published by the domain
or that no checkable sender domain could be determined from the given
identity. The checking software cannot ascertain whether or not the
client host is authorized. Such a message can be subject to
further checks that will decide about its fate.
Neutral
¶The domain owner has explicitly stated that he cannot or does not
want to assert whether or not the IP address is authorized. This
result must be treated exactly like None
; the distinction
between them exists only for informational purposes
Pass
¶The client is authorized to send mail with the given identity. The message can be subject to further policy checks with confidence in the legitimate use of the identity or it can be accepted in the absence of such checks.
Fail
¶The client is not authorized to use the domain in the given identity. The proper action in this case can be to mark the message with a header explicitly stating it is spam, or to reject it outright.
If you choose to reject such mails, we suggest to use reject
550 5.7.1
, as recommended by RFC 4408. The reject can return either
a default explanation string, or the one supplied by the domain that
published the SPF records, as in the example below:
reject 550 5.7.1 "SPF check failed:\n%spf_explanation"
(for the description of spf_explanation
, see spf_explanation)
SoftFail
¶The domain believes the host is not authorized but is not willing to
make that strong of a statement. This result code should be treated
as somewhere in between a Fail
and a Neutral
. It is not
recommended to reject the message based solely on this result.
TempError
A transient error occurred while performing SPF check. The proper action in this case is to accept or temporarily reject the message. If you choose the latter, we suggest to use SMTP reply code of ‘451’ and DSN code ‘4.4.3’, for example:
tempfail 451 4.4.3 "Transient error while performing SPF verification"
PermError
This result means that the domain’s published records could not be
correctly interpreted. This signals an error condition that requires
manual intervention to be resolved, as opposed to the TempError
result.
The following example illustrates the use of SPF
verification in envfrom
handler:
require 'status' require 'spf' prog envfrom do switch check_host($client_addr, domainpart($f), $f, $s) do case Fail: string text "" if spf_explanation != "" set text "%text\n%spf_explanation" fi reject 550 5.7.1 "SPF MAIL FROM check failed: %text" case Pass: accept case TempError: tempfail 451 4.4.3 "Transient error while performing SPF verification" default: on poll $f do when success: accept when not_found or failure: reject 550 5.1.0 "Sender validity not confirmed" when temp_failure: tempfail 450 4.7.0 "Temporary failure during sender verification" done done done
The SPF support is implemented in MFL in two layers: a built-in layer that provides basic support, and a library layer that provides a convenience wrapper over the library function.
The library layer is implemented in the module spf.mfl (see Modules).
The rest of this node describes available SPF functions and variables.
number
spf_check_host (string ip, string domain, string sender, string helo_domain, string my_domain)
¶This function is the basic implementation of the check_host
function, defined in RFC 4408, chapter 4. It fetches SPF
records, parses them, and evaluates them to determine whether a
particular host (ip) is or is not permitted to send mail from a
given email address (sender). The function returns an SPF
result code.
Arguments are:
The IP address of the SMTP client that is emitting the mail.
Usually it is $client_addr
.
The domain that provides the sought-after authorization information;
Normally it is the domain portion of the MAIL FROM
or
HELO
identity.
The MAIL FROM
identity.
The HELO
identity.
The SMTP domain served by the local server.
Before returning the spf_check_host
function stores
additional information in global variables:
spf_explanation
If the result code is Fail
, this variable contains the
explanation string as returned by the publishing domain, prefixed with
the value of the global variable spf_explanation_prefix
.
For example, if spf_explanation_prefix
contains ‘The
domain %{o} explains: ’, and the publishing domain
‘example.com’ returns the explanation string ‘Please see
http://www.example.com/mailpolicy.html’, than the value of
spf_explanation
will be:
The domain example.com explains: Please see http://www.example.com/mailpolicy.html
(see RFC 4408, chapter 8, for the description of SPF macro facility).
spf_mechanism
Name of the SPF mechanism that decided about the result code of the SPF record. If one or more ‘include’ or ‘redirect’ mechanisms were traversed before arriving at that mechanism, their values are appended in the reverse order.
number
spf_test_record (string record, string ip, string domain, string sender, string helo_domain, string my_domain)
¶Evaluate SPF record record as if it were published
by domain. The rest of arguments are the same as for
spf_check_host
above.
This function is designed primarily for testing and debugging purposes. You would hardly need to use it.
The spf_test_record
function sets the same global variables
as spf_check_host
.
number
check_host (string ip, string domain, string sender, string helo)
¶This function implements the check_host
function, defined in
RFC 4408, chapter 4. It fetches SPF records, parses them,
and evaluates them to determine whether a particular host (ip)
is or is not permitted to send mail from a given email address
(sender). The function returns an SPF result code.
This function is a wrapper over the built-in spf_check_host
.
The arguments are:
The IP address of the SMTP client that is emitting the mail.
Usually it is the same as the value of $client_addr
.
The domain that provides the sought-after authorization information;
Normally it is the domain portion of the MAIL FROM
or
HELO
identity.
The MAIL FROM
identity.
The HELO
identity.
string
spf_status_string (number code)
¶Converts numeric SPF result code to its string representation.
string
spf_explanation ¶If check_host
(or spf_check_host
or
spf_test_record
) returned Fail
, this variable contains the
explanation string as returned by the publishing domain, prefixed with
the value of the global variable spf_explanation_prefix
.
For example, if spf_explanation_prefix
contains ‘The
domain %{o} explains: ’, and the publishing domain
‘example.com’ returns the explanation string ‘Please see
http://www.example.com/mailpolicy.html’, than the value of
spf_explanation
will be:
The domain example.com explains: Please see http://www.example.com/mailpolicy.html
string
spf_mechanism ¶Set to the name of a SPF mechanism that decided about the result code of the SPF record.
string
spf_explanation_prefix ¶The prefix to be appended to the explanation string before storing
it in the spf_explanation
variable. This string can contain
valid SPF macros (see
RFC 4408, chapter 8), for
example:
set spf_explanation_prefix "%{o} explains: "
The default value is ‘""’ (an empty string).
DKIM or DomainKeys Identified Mail is an email authentication method that allows recipients to verify if an email was authorized by the owner of the domain that email claims to originate from. It does so by adding a digital signature which is verified using a public key published as a DNS TXT record. For technical details about DKIM, please refer to RFC 6376 (http://tools.ietf.org/html/rfc6376).
MFL provides functions for DKIM signing and verification.
number
dkim_verify (number msg)
¶Verifies the message msg (a message descriptor, obtained from
a call to current_message
, mailbox_get_message
,
message_from_stream
or a similar function).
Return value (constants defined in the ‘status’ module):
The message contains one or more ‘DKIM-Signature’ headers and one of them verified successfully.
The message contains one or more ‘DKIM-Signature’ headers, all of which failed to verify.
The message was not signed using DKIM, or the DNS query to obtain the public key failed, or an internal software error occurred during verification.
The following two global variables are always set upon return from this
function: dkim_explanation
and dkim_explanation_code
.
These can be used to clarify the verification result to the end user.
The variable dkim_signing_algorithm
is initialized with the
name of the algorithm used to sign the message.
Upon successful return, the variable dkim_verified_signature
is
set to the value of the successfully verified DKIM signature.
string
dkim_signing_algorithm ¶Name of the algorithm used to sign the message (either ‘rsa-sha1’ or ‘rsa-sha256’). If the algorithm was not specified (e.g. the signature is malformed), this variable is assigned an empty value.
string
dkim_explanation ¶An explanatory message clarifying the verification result.
number
dkim_explanation_code ¶A numeric code corresponding to the ‘dkim_explanation’ string. Its possible values are defined in ‘status.mfl’:
‘DKIM verification passed’
‘DKIM verification passed’
‘No DKIM signature’
‘internal error’
‘signature syntax error’
‘signature is missing required tag’
According to the DKIM specification, required tags are: a=
,
b=
, bh=
, d=
, h=
, s=
, v=
.
‘domain mismatch’
The domain part of the i=
tag does not match and is not a
subdomain of the domain listed in the d=
tag.
‘incompatible version’
Incompatible DKIM version listed in the v=
tag.
‘unsupported signing algorithm’
Either the a=
tag of the DKIM signature contains an unsupported
algorithm (currently supported algorithms are: ‘rsa-sha1’ and
‘rsa-sha256’) or this algorithm, while being supported by
mailfromd
, is not listed in the h=
tag of the public
DKIM key.
‘unsupported query method’
The q=
tag of the public DKIM key contains something other than
‘dns/txt’.
‘From field not signed’
‘signature expired’
‘public key unavailable’
‘public key not found’
‘key syntax error’
‘key revoked’
‘body hash did not verify’
‘can't decode b= tag’
Base64 decoding of the b=
tag failed.
‘signature did not verify’
‘unsupported public key type’
The k=
tag of the public DKIM signature contains a value, other
than ‘rsa’.
string
dkim_verified_signature ¶Upon successful return from the dkim_verify
function, this
variable holds the value of the successfully verified DKIM header.
This value is unfolded and all whitespace is removed from it.
An example of using the ‘dkim_verify’ function:
require status require dkim prog eom do string result switch dkim_verify(current_message()) do case DKIM_VERIFY_OK: set result "pass; verified for " . dkim_verified_signature_tag('i') case DKIM_VERIFY_PERMFAIL: set result "fail (%dkim_explanation)" case DKIM_VERIFY_TEMPFAIL: set result "neutral" done header_add("X-Verification-Result", "dkim=%result") done
The ‘dkim’ module defines convenience functions for manipulating with DKIM signatures:
dkim_signature_tag
(string sig, string tag)
¶Extracts the value of the tag tag from the DKIM signature sig. Signature must be normalized by performing the header unwrapping and removing whitespace characters.
If the tag was not found, returns empty string, unless tag is one of the tags listed in the table below. If any of these tags are absent, the following values are returned instead:
Tag | Default value |
---|---|
c | ‘simple/simple’ |
q | ‘dns/txt’ |
i | ‘@’ + the value of the ‘d’ tag. |
string
dkim_verified_signature_tag (string tag)
¶Returns the value of tag tag from the ‘dkim_verified_signature’ variable.
void
dkim_sign (string d, string s, string keyfile, [ string ch, string cb, string headers, string algo ])
¶This function is available only in the eom
handler.
Signs the current message. Notice, that no other modification should be attempted on the message after calling this function. Doing so would make the signature invalid.
Mandatory arguments:
Name of the domain claiming responsibility for an introduction of a message into the mail stream. It is also known as the signing domain identifier (SDID).
The selector name. This value, along with d identifies the location of the DKIM public key necessary for verifying the message. The public key is stored in the DNS TXT record for
s._domainkey.d
Name of the disk file that keeps the private key for signing the message. The file must be in PKCS#1 or PKCS#8 format (PEM formatted).
Optional arguments:
Canonicalization algorithm for message headers. Valid values are: ‘simple’ and ‘relaxed’. ‘simple’ is the default.
Canonicalization algorithm for message body. Valid and default values are the same as for ch.
A colon-separated list of header field names that identify the header fields that must be signed. Optional whitespace is allowed at either side of each colon separator. Header names are case-insensitive. This list must contain at least the ‘From’ header.
It may contain names of headers that are not present in the message being signed. This provides a way to explicitly assert the absence of a header field. For example, if headers contained ‘X-Mailer’ and that header is not present in the message being signed, but is added by a third party later, the signature verification will fail.
Similarly, listing a header field name once more than the actual number of its occurrences in a message allows you to prevent any further additions. For example, if there is a single ‘Comments’ header field at the time of signing, putting ‘Comments:Comments:’ in the headers parameter is sufficient to prevent any surplus ‘Comments’ headers from being added later on.
Multiple instances of the same header name are allowed. They mean that multiple occurrences of the corresponding header field will be included in the header hash. When such multiple header occurrences are referenced, they will be presented to the hashing algorithm in the reverse order. E.g. if the header list contained ‘Received:Received’) and the current message contained three ‘Received’ headers:
Received: A Received: B Received: C
then these headers will be signed in the following order:
Received: C Received: B
The default value for this parameter, split over multiple lines for readability, is as follows:
Signing algorithm: either ‘rsa-sha256’ or ‘rsa-sha1’. Default is ‘rsa-sha256’.
An example of using this function:
precious string domain "example.org" precious string selector "s2048" prog eom do dkim_sign("example.org", "s2048", "/etc/pem/my-private.pem", "relaxed", "relaxed", "from:to:subject") done
When sending a signed message, it is critical that no other
modifications be applied to the message after it has been signed.
Unfortunately, it is not always the case when mailfromd
is
used with Sendmail
. Before sending the message over SMTP,
Sendmail
reformats the headers that contain a list of email
addresses, by applying to them a procedure called in its parlance
commaization. The following headers are modified:
Apparently-To
, Bcc
, Cc
,
Disposition-Notification-To
, Errors-To
, From
,
Reply-To
, Resent-Bcc
, Resent-Cc
,
Resent-From
, Resent-Reply-To
, Resent-Sender
,
Resent-To
, Sender
, To
. Thus, if your
dkim_sign
includes any of these in the signature (which is the
default) and some of them happen to be formatted other way than the
one Sendmail
prefers, the DKIM signature would not verify on
the recipient side. To prevent this from happening, dkim_sign
mimics the Sendmail behavior and reformats those headers before
signing the message. This should ensure that the message signed and
the message actually sent are the same. This default behavior is
controlled by the following global variable:
number
dkim_sendmail_commaize ¶“Commaize” the address headers (see the list above) of the message the same way Sendmail does, and then sign the resulting message.
The default value is 1 (true
). You can set it to 0
(false
) if this behavior is not what you want (e.g. if you are
using postfix
or some other MTA).
The functions header_add
and header_insert
(see Header Modification Functions) as well as the add
action (see header manipulation) cannot interact properly with
dkim_sign
due to the shortcomings of the Milter API. If any of
these was called, dkim_sign
will throw the e_badmmq
exception with the diagnostics following diagnostics:
MMQ incompatible with dkim_sign: op on h, value v
where op is the operation code (‘ADD HEADER’ or ‘INSERT HEADER’), h is the header name and v is its value.
The following example shows one graceful way of handling such exception:
prog eom do try do dkim_sign("example.org", "s2048", "/etc/pem/my-private.pem") done catch e_badmmq do # Purge the message modification queue mmq_purge() # and retry dkim_sign("example.org", "s2048", "/etc/pem/my-private.pem") done done
See Message Modification Queue, for a discussion of the message modification queue.
Follow these steps to set up your own DKIM record:
Use the openssl genrsa
command. Run:
openssl genrsa -out private.pem 2048
The last argument is the size of the private key to generate in bits.
openssl rsa -in private.pem -pubout -outform PEM -out public.pem
A DKIM record is a TXT type DNS record that holds the public key part for verifying messages. Its format is defined in RFC 487129. The label for this record is composed as follows:
s._domainkey.d
where d is your domain name, and s is the selector you
chose to use. You will use these two values as parameters to the
dkim_sign
function in your eom
handler. E.g. if your
domain in ‘example.com’ and selector is ‘s2048’, then the
DKIM TXT record label is ‘s2048._domainkey.example.com’.
The public key file generated in step 2 will have the following contents:
-----BEGIN PUBLIC KEY----- base64 -----END PUBLIC KEY-----
where base64 is the key itself in base64 encoding. The minimal DKIM TXT record will be:
"v=DKIM1; p=base64"
The only mandatory tag is in fact ‘p=’. The use of ‘v=’ is recommended. More tags can be added as needed. In particular, while testing the DKIM support, it is advisable to add the ‘t=y’ tag.
Socket map (sockmap for short) is a special type of database used in Sendmail and MeTA1. It uses a simple server/client protocol over INET or UNIX stream sockets. The server listens on a socket for queries. The client connects to the server and sends it a query, consisting of a map name and a key separated by a single space. Both map name and key are sequences of non-whitespace characters. The map name serves to identify the type of the query. The server replies with a response consisting of a status indicator and result, separated by a single space. The result part is optional.
For example, following is the query for key ‘smith’ in map ‘aliases’:
11:aliases news,
A possible reply is:
18:OK root@domain.net,
This reply means that the key ‘news’ was found in the map, and the value corresponding to that key is ‘root@domain.net’.
The following reply means the key was not found:
8:NOTFOUND,
For a detailed description of the sockmap protocol, see Protocol in Smap manual.
The MFL library provides two primitives for dealing with sockmaps. Both primitives become available after requiring the sockmap module.
string
sockmap_lookup (number fd, string map, string key)
¶This function look ups the key in the map. The fd
refers to the sockmap to use. It must be obtained as a result of a
previous call to open
with the URL of the sockmap as
its first argument (see open). For example:
number fd open("@ unix:///var/spool/meta1/smap/socket") string ret sockmap_query(fd, "aliases", $rcpt_to) if ret matches "OK (.+)" set alias \1 fi close(fd)
string
sockmap_single_lookup (string url, string map, string key)
¶This function connects to the sockmap identified by the url, queries for key in map and closes the connection. It is useful when you need to perform only a single lookup on the sockmap.
The National Language Support functions allow you to write your scripts in such a way, that any textual messages they display are automatically translated to your native language, or, more precisely, to the language required by your current locale.
This section assumes the reader is familiar with the concepts of program
internationalization and localization. If not, please
refer to The Purpose of GNU
gettext
in GNU gettext manual, before reading further.
In general, internationalization of any MFL script
follows the same rules as described in the GNU gettext manual.
First of all, you select the program message domain, i.e. the
identifier of a set of translatable messages your script contain.
This identifier is then used to select appropriate translation.
The message domain is set using textdomain
function. For the
purposes of this section, let’s suppose the domain name is
‘myfilter’. All NLS functions are provided in the
nls module, which you need to require prior to using
any of them.
To find translations of textual message to the current locale, the
underlying gettext
mechanism will look for file
dirname/locale/LC_MESSAGES/domainname.mo,
where dirname is the message catalog hierarchy name,
locale is the locale name, and domainname is the name of
the message domain. By default dirname is
/usr/local/share/locale, but you may change it using
bindtextdomain
function. The right place for this initial
NLS setup is in the ‘begin’ block (see The ‘begin’ and ‘end’ special handlers).
To summarize all the above, the usual NLS setup will look like:
require nls begin do textdomain("myfilter") bindtextdomain("myfilter", "/usr/share/locale"); done
For example, given the settings above, and supposing the environment
variable LC_ALL
is set to ‘pl’, translations will be looked
in file /usr/share/locale/pl/LC_MESSAGES/myfilter.mo.
Once this preparatory work is done, you can request each message to
be translated by using gettext
function, or _
(underscore) macro. For example, the following statement will produce
translated textual description for ‘450’ response:
tempfail 450 4.1.0 _("Try again later")
Of course it assumes that the appropriate myfile.mo file
already exists. If it does not, nothing bad happens: in this case the
macro _
(as well as gettext
function) will simply
return its argument unchanged, so that the remote party will get the
textual message in English.
The ‘mo’ files are binary files created from ‘po’ source
files using msgfmt
utility, as described in Producing Binary MO Files in GNU
gettext manual. In turn, the format of ‘po’ files is described
in The Format of PO Files in GNU gettext manual.
string
bindtextdomain (string domain, string dirname)
¶This function sets the base directory of the hierarchy containing message catalogs for a given message domain.
domain is a string identifying the textual domain. If it is not empty, the base directory for message catalogs belonging to domain domain is set to dirname. It is important that dirname be an absolute pathname; otherwise it cannot be guaranteed that the message catalogs will be found.
If domain is ‘""’, bindtextdomain
returns the
previously set base directory for domain domain.
The rest of this section describes the NLS functions supplied in the nls module.
string
dgettext (string domain, string msgid)
¶dgettext
attempts to translate the string msgid into the
currently active locale, according to the settings of the textual
domain domain. If there is no translation available,
dgettext
returns msgid unchanged.
string
dngettext (string domain, string msgid, string msgid_plural, number n)
¶The dngettext
functions attempts to translate a text string
into the language specified by the current locale, by looking up the
appropriate singular or plural form of the translation in a message
catalog, set for the textual domain domain.
See Additional functions for plural forms in GNU gettext utilities, for a discussion of the plural form handling in different languages.
string
textdomain (string domain)
¶The textdomain
function sets the current message domain to
domain, if it is not empty. In any case the function returns
the current message domain. The current domain is ‘mailfromd’
initially. For example, the following sequence of textdomain
invocations will yield:
textdomain("") ⇒ "mailfromd" textdomain("myfilter") ⇒ "myfilter" textdomain("") ⇒ "myfilter"
string
gettext (string msgid)
¶gettext
attempts to translate the string msgid into the
currently active locale, according to the settings of the current textual
domain (set using textdomain
function). If there is no
translation available, gettext
returns msgid unchanged.
string
ngettext (string msgid, string msgid_plural, number n)
¶The ngettext
functions attempts to translate a text string
into the language specified by the current locale, by looking up the
appropriate singular or plural form of the translation in a message
catalog, set for the current textual domain.
See Additional functions for plural forms in GNU gettext utilities, for a discussion of the plural form handling in different languages.
The basic means for outputting diagnostic messages is the
‘echo’ instruction (see The echo
statement), which sends its arguments to
the currently established logging channel. In daemon mode, the latter
is normally connected to syslog, so any echoed messages are sent there
with the facility selected in mailfromd configuration and priority
‘info’.
If you want to send a message to another facility and/or priority, use the ‘syslog’ function:
void
syslog (number priority, string text)
¶Sends text to syslog. The priority argument is formed by ORing the facility and the level values (explained below). The facility level is optional. If not supplied, the currently selected logging facility is used.
The facility specifies what type of program is logging the message, and the level indicates its relative severity. The following symbolic facility values are declared in the syslog module: ‘LOG_KERN’, ‘LOG_USER’, ‘LOG_MAIL’, ‘LOG_DAEMON’, ‘LOG_AUTH’, ‘LOG_SYSLOG’, ‘LOG_LPR’, ‘LOG_NEWS’, ‘LOG_UUCP’, ‘LOG_CRON’, ‘LOG_AUTHPRIV’, ‘LOG_FTP’ and ‘LOG_LOCAL0’ through ‘LOG_LOCAL7’
The declared severity levels are: ‘LOG_EMERG’, ‘LOG_ALERT’, ‘LOG_CRIT’, ‘LOG_ERR’, ‘LOG_WARNING’, ‘LOG_NOTICE’, ‘LOG_INFO’ and ‘LOG_DEBUG’.
These functions are designed for debugging the MFL programs.
void
debug (string spec)
¶Enable debugging. The value of spec sets the debugging level. See debugging level specification, for a description of its format.
For compatibility with previous versions, this function is also available under the name ‘mailutils_set_debug_level’.
number
debug_level ([string srcname])
¶This function returns the debugging level currently in effect for the source module srcname, or the global debugging level, if called without arguments.
For example, if the program was started with --debug='all.trace5;engine.trace8' option, then:
debug_level() ⇒ 127 debug_level("engine") ⇒ 1023 debug_level("db") ⇒ 0
boolean
callout_transcript ([boolean value])
¶Returns the current state of the callout SMTP transcript. The result is 1 if the transcript is enabled and 0 otherwise. The transcript is normally enabled either by the use of the --transcript command line option (see SMTP transcript) or via the ‘transcript’ configuration statement (see transcript).
The optional value, supplies the new state for SMTP transcript. Thus, calling ‘callout_transcript(0)’ disables the transcript.
This function can be used in bracket-like fashion to enable transcript for a certain part of MFL program, e.g.:
number xstate callout_transcript(1) on poll $f do ... done set xstate callout_transcript(0)
Note, that the use of this function (as well as the use of the
--transcript option) makes sense only if callouts are
performed by the mailfromd
daemon itself. It will not
work if a dedicated callout server is used for that purpose
(see calloutd
).
string
debug_spec ([string catnames, bool showunset])
¶Returns the current debugging level specification, as given by
--debug command line option or by the debug
configuration
statement (see Logging and Debugging configuration).
If the argument srcnames is specified, it is treated as a semicolon-separated list of categories for which the debugging specification is to be returned.
For example, if mailfromd
was started with
--debug=all.trace5;spf.trace1;engine.trace8;db.trace0, then:
debug_spec() ⇒ "all.trace5,engine.trace8" debug_spec("all;engine") ⇒ "all.trace5,engine.trace8" debug_spec("engine;db") ⇒ "db.trace0;engine.trace8" debug_spec("prog") ⇒ ""
When called without arguments, debug_spec
returns only
those categories which have been set, as shown in the first example
above.
Optional showunset parameters controls whether to return unset module specifications. To print all debugging specifications, whether set or not, use
debug_spec("", 1)
These three functions are intended to complement each other. The
calls to debug
can be placed around some piece of code you wish
to debug, to enable specific debugging information for this code
fragment only. For example:
/* Save debugging level for dns.c source */ set dlev debug_spec("dns", 1) /* Set new debugging level */ debug("dns.trace8") . . . /* Restore previous level */ debug(dlev)
void
program_trace (string module)
¶Enable tracing for a set of modules given in module argument. See --trace-program, for a description of its format.
void
cancel_program_trace (string module)
¶Disable tracing for given modules.
This pair of functions is also designed to be used together in
a bracket-like fashion. They are useful for debugging
mailfromd
, but are not advised to use otherwise, since
tracing slows down the execution considerably.
void
stack_trace ()
¶Generate a stack trace in this point. See tracing runtime errors, for the detailed description of stack traces.
The functions below are intended mainly for debugging MFL run-time
engine and for use in mailfromd
testsuite. You will hardly
need to use them in your programs.
void
_expand_dataseg (number n)
¶Expands the run-time data segment by at least n words.
number
_reg (number r)
¶Returns the value of the register r at the moment of the call. Symbolic names for run-time registers are provided in the module _register:
Name | Register |
---|---|
REG_PC | Program counter |
REG_TOS | Top of stack |
REG_TOH | Top of heap |
REG_BASE | Frame base |
REG_REG | General-purpose accumulator |
REG_MATCHSTR | Last matched string pointer |
number
_stack_free ()
¶Returns number of words available for use in stack. This is the same as
_reg(REG_TOS) - _reg(REG_TOH)
number
_heap_reserve (number n)
¶Use up next n words in the heap. Return the address of the first word.
void
_wd ([number n])
¶Enters a time-consuming loop and waits there for n seconds (by
default – indefinitely). The intention is to facilitate attaching
to mailfromd
with the debugger. Before entering the loop,
a diagnostic message is printed on the ‘crit’ facility, informing
about the PID of the process and suggesting the command to be used to
attach to it, e.g.:
mailfromd: process 21831 is waiting for debug mailfromd: to attach: gdb -ex 'set variable mu_wd::_count_down=0' /usr/sbib/mailfromd 21831
These functions convert numeric identifiers of various MFL entities to strings and vice-versa.
string
milter_state_name (number code)
¶Convert the numeric identifier of a milter state to textual form.
It is normally used to convert the milter_state
variable
(see milter state) to string, e.g.:
milter_state_name(5) ⇒ "connect"
If code does not refer to a valid milter state, the e_inval
exception is raised.
number
milter_state_code (string name)
¶Returns numeric code of the milter state name. If name
does not refer to a valid milter state, returns 0
(milter_state_none
from the milter.mfl module).
string
milter_action_name (number code)
¶Convert the numeric identifier of a reply action (see reply actions) to textual name.
If code does not refer to a valid reply action, the e_inval
exception is raised.
milter_state_code("connect") ⇒ 5
This function is useful in action hooks. See Action Hook, for details.
number
milter_action_name (string name)
¶Returns numeric code of the reply action identified by name. If name does not refer to a valid action, returns -1.
The calls described in this section provide interface for invoking functions defined in a dynamically loaded library and retrieving their return values. For a detailed description of this interface and a discussion of its use, see Dynamically Loaded Modules.
number
dlopen (filename)
¶Loads the dynamically loaded library filename and returns a numeric handle that can be used to call functions from that library.
Unless filename is an absolute pathname, it will be looked up in
mfmod search path, defined by the configuration variable
runtime.mfmod-path
(see mfmod-path).
Maximum number of dynamic libraries that can be loaded simultaneously
is limited by the configuration variable runtime.max-mfmods
.
Once open, the library remains loaded until mailfromd
exits.
There is no dlclose
function, since it is not deemed necessary
(at the time of this writing, at least). Therefore, the common
practice is to call this function in a begin
section
(see The ‘begin’ and ‘end’ special handlers) and assign its return value to a global or static
variable, which will then be used by further dl*
calls in this
module.
any
dlcall (number dlh, string symbol, string types, ...)
¶Calls a mfmod function symbol from the library identified
by handle dlh (a value returned by a previous call to
dlopen
). The types parameter defines types of the
remaining arguments. It is a list of type letters, one for each argument:
String value.
Numeric value.
Message.
An example usage:
set dlh dlopen("mfmod_test.so") string x dlcall(dlh, "runtest", "sn", "input string", 3)
This example calls the function ‘runtest’ from the
mfmod_test.so library with two arguments: a string ‘input
string’ and numeric value ‘3’ and assings the return value to the
variable x
.
Type of the return value from dlcall
is determined by the
value of retval.type
upon return from the underlying library
function. See Loadable Library, for a detailed description.
For more details about using this function, Interface Module.
That is, if it supports Milter protocol 6 and
upper. Sendmail 8.14.0 and Postfix 2.6 and newer do. MeTA1 (via
pmult
) does as well. See Using mailfromd
with Various MTAs, for more
details.
Support for other locales is planned for future versions.
For example:
prog header do echo unfold($2) done
Note, that the return code is inverted in respect to the system function ‘access(2)’.
It is made optional in order to provide backward compatibility with the releases of mailfromd prior to 5.0.93.
Although RFC 4408
introduces a special SPF
record type for this purpose, it is
not yet widely used. As of version 8.17.1, MFL
does not support SPF
DNS records.