NAME

NAME
MIME::Lite - low-calorie MIME generator

SYNOPSIS
use MIME::Lite;

Create a single-part message:

### Create a new single-part message, to send a GIF file:
$msg = MIME::Lite->new(
From =>'[email protected]',
To =>'[email protected]',
Cc =>'[email protected], [email protected]',
Subject =>'Helloooooo, nurse!',
Type =>'image/gif',
Encoding =>'base64',
Path =>'hellonurse.gif'
);

Create a multipart message (i.e., one with attachments):

### Create a new multipart message:
$msg = MIME::Lite->new(
From =>'[email protected]',
To =>'[email protected]',
Cc =>'[email protected], [email protected]',
Subject =>'A message with 2 parts...',
Type =>'multipart/mixed'
);

### Add parts (each "attach" has same arguments as "new"):
$msg->attach(Type =>'TEXT',
Data =>"Here's the GIF file you wanted"
);
$msg->attach(Type =>'image/gif',
Path =>'aaa000123.gif',
Filename =>'logo.gif',
Disposition => 'attachment'
);

Output a message:

### Format as a string:
$str = $msg->as_string;

### Print to a filehandle (say, a "sendmail" stream):
$msg->print(\*SENDMAIL);

Send a message:

### Send in the "best" way (the default is to use "sendmail"):
$msg->send;

DESCRIPTION
In the never-ending quest for great taste with fewer calories, we
proudly present: *MIME::Lite*.

MIME::Lite is intended as a simple, standalone module for generating
(not parsing!) MIME messages... specifically, it allows you to output a
simple, decent single- or multi-part message with text or binary
attachments. It does not require that you have the Mail:: or MIME::
modules installed.

You can specify each message part as either the literal data itself (in
a scalar or array), or as a string which can be given to open() to get a
readable filehandle (e.g., "<filename" or "somecommand|").

You don't need to worry about encoding your message data: this module
will do that for you. It handles the 5 standard MIME encodings.

If you need more sophisticated behavior, please get the MIME-tools
package instead. I will be more likely to add stuff to that toolkit over
this one.

EXAMPLES
Create a simple message containing just text

$msg = MIME::Lite->new(
From =>'[email protected]',
To =>'[email protected]',
Cc =>'[email protected], [email protected]',
Subject =>'Helloooooo, nurse!',
Data =>"How's it goin', eh?"
);

Create a simple message containing just an image

$msg = MIME::Lite->new(
From =>'[email protected]',
To =>'[email protected]',
Cc =>'[email protected], [email protected]',
Subject =>'Helloooooo, nurse!',
Type =>'image/gif',
Encoding =>'base64',
Path =>'hellonurse.gif'
);

Create a multipart message

### Create the multipart "container":
$msg = MIME::Lite->new(
From =>'[email protected]',
To =>'[email protected]',
Cc =>'[email protected], [email protected]',
Subject =>'A message with 2 parts...',
Type =>'multipart/mixed'
);

### Add the text message part:
### (Note that "attach" has same arguments as "new"):
$msg->attach(Type =>'TEXT',
Data =>"Here's the GIF file you wanted"
);

### Add the image part:
$msg->attach(Type =>'image/gif',
Path =>'aaa000123.gif',
Filename =>'logo.gif',
Disposition => 'attachment'
);

Attach a GIF to a text message

This will create a multipart message exactly as above, but using the
"attach to singlepart" hack:

### Start with a simple text message:
$msg = MIME::Lite->new(
From =>'[email protected]',
To =>'[email protected]',
Cc =>'[email protected], [email protected]',
Subject =>'A message with 2 parts...',
Type =>'TEXT',
Data =>"Here's the GIF file you wanted"
);

### Attach a part... the make the message a multipart automatically:
$msg->attach(Type =>'image/gif',
Path =>'aaa000123.gif',
Filename =>'logo.gif'
);

Attach a pre-prepared part to a message

### Create a standalone part:
$part = MIME::Lite->new(
Type =>'text/html',
Data =>'<H1>Hello</H1>',
);
$part->attr('content-type.charset' => 'UTF8');
$part->add('X-Comment' => 'A message for you');

### Attach it to any message:
$msg->attach($part);

Print a message to a filehandle

### Write it to a filehandle:
$msg->print(\*STDOUT);

### Write just the header:
$msg->print_header(\*STDOUT);

### Write just the encoded body:
$msg->print_body(\*STDOUT);

Print a message into a string

### Get entire message as a string:
$str = $msg->as_string;

### Get just the header:
$str = $msg->header_as_string;

### Get just the encoded body:
$str = $msg->body_as_string;

Send a message

### Send in the "best" way (the default is to use "sendmail"):
$msg->send;

Send an HTML document... with images included!

$msg = MIME::Lite->new(
To =>'[email protected]',
Subject =>'HTML with in-line images!',
Type =>'multipart/related'
);
$msg->attach(Type => 'text/html',
Data => qq{ <body>
Here's <i>my</i> image:
<img src="cid:myimage.gif">
</body> }
);
$msg->attach(Type => 'image/gif',
Id => 'myimage.gif',
Path => '/path/to/somefile.gif',
);
$msg->send();

Change how messages are sent

### Do something like this in your 'main':
if ($I_DONT_HAVE_SENDMAIL) {
MIME::Lite->send('smtp', "smtp.myisp.net", Timeout=>60);
}

### Now this will do the right thing:
$msg->send; ### will now use Net::SMTP as shown above

PUBLIC INTERFACE
Global configuration

To alter the way the entire module behaves, you have the following
methods/options:

MIME::Lite->field_order()
When used as a classmethod, this changes the default order in which
headers are output for *all* messages. However, please consider
using the instance method variant instead, so you won't stomp on
other message senders in the same application.

MIME::Lite->quiet()
This classmethod can be used to suppress/unsuppress all warnings
coming from this module.

MIME::Lite->send()
When used as a classmethod, this can be used to specify a different
default mechanism for sending message. The initial default is:

MIME::Lite->send("sendmail", "/usr/lib/sendmail -t -oi -oem");

However, you should consider the similar but smarter and taint-safe
variant:

MIME::Lite->send("sendmail");

Or, for non-Unix users:

MIME::Lite->send("smtp");

$MIME::Lite::AUTO_CC
If true, automatically send to the Cc/Bcc addresses for
send_by_smtp(). Default is true.

$MIME::Lite::AUTO_CONTENT_TYPE
If true, try to automatically choose the content type from the file
name in `new()'/`build()'. In other words, setting this true changes
the default `Type' from `"TEXT"' to `"AUTO"'.

Default is false, since we must maintain backwards-compatibility
with prior behavior. Please consider keeping it false, and just
using Type 'AUTO' when you build() or attach().

$MIME::Lite::AUTO_ENCODE
If true, automatically choose the encoding from the content type.
Default is true.

$MIME::Lite::AUTO_VERIFY
If true, check paths to attachments right before printing, raising
an exception if any path is unreadable. Default is true.

$MIME::Lite::PARANOID
If true, we won't attempt to use MIME::Base64, MIME::QuotedPrint, or
MIME::Types, even if they're available. Default is false. Please
consider keeping it false, and trusting these other packages to do
the right thing.

Construction

new [PARAMHASH]
*Class method, constructor.* Create a new message object.

If any arguments are given, they are passed into `build()';
otherwise, just the empty object is created.

attach PART
attach PARAMHASH...
*Instance method.* Add a new part to this message, and return the
new part.

If you supply a single PART argument, it will be regarded as a
MIME::Lite object to be attached. Otherwise, this method assumes
that you are giving in the pairs of a PARAMHASH which will be sent
into `new()' to create the new part.

One of the possibly-quite-useful hacks thrown into this is the
"attach-to-singlepart" hack: if you attempt to attach a part (let's
call it "part 1") to a message that doesn't have a content-type of
"multipart" or "message", the following happens:

* A new part (call it "part 0") is made.

* The MIME attributes and data (but *not* the other headers) are cut
from the "self" message, and pasted into "part 0".

* The "self" is turned into a "multipart/mixed" message.

* The new "part 0" is added to the "self", and *then* "part 1" is
added.

One of the nice side-effects is that you can create a text message
and then add zero or more attachments to it, much in the same way
that a user agent like Netscape allows you to do.

build [PARAMHASH]
*Class/instance method, initializer.* Create (or initialize) a MIME
message object. Normally, you'll use the following keys in
PARAMHASH:

* Data, FH, or Path (either one of these, or none if multipart)
* Type (e.g., "image/jpeg")
* From, To, and Subject (if this is the "top level" of a message)

The PARAMHASH can contain the following keys:

(fieldname)
Any field you want placed in the message header, taken from the
standard list of header fields (you don't need to worry about
case):

Approved Encrypted Received Sender
Bcc From References Subject
Cc Keywords Reply-To To
Comments Message-ID Resent-* X-*
Content-* MIME-Version Return-Path
Date Organization

To give experienced users some veto power, these fields will be
set *after* the ones I set... so be careful: *don't set any MIME
fields* (like `Content-type') unless you know what you're doing!

To specify a fieldname that's *not* in the above list, even one
that's identical to an option below, just give it with a
trailing `":"', like `"My-field:"'. When in doubt, that *always*
signals a mail field (and it sort of looks like one too).

Data *Alternative to "Path" or "FH".* The actual message data. This may
be a scalar or a ref to an array of strings; if the latter, the
message consists of a simple concatenation of all the strings in
the array.

Datestamp
*Optional.* If given true (or omitted), we force the creation of
a `Date:' field stamped with the current date/time if this is a
top-level message. You may want this if using send_by_smtp(). If
you don't want this to be done, either provide your own Date or
explicitly set this to false.

Disposition
*Optional.* The content disposition, `"inline"' or
`"attachment"'. The default is `"inline"'.

Encoding
*Optional.* The content transfer encoding that should be used to
encode your data:

Use encoding: | If your message contains:
------------------------------------------------------------
7bit | Only 7-bit text, all lines <1000 characters
8bit | 8-bit text, all lines <1000 characters
quoted-printable | 8-bit text or long lines (more reliable than "8bit")
base64 | Largely non-textual data: a GIF, a tar file, etc.

The default is taken from the Type; generally it is "binary" (no
encoding) for text/*, message/*, and multipart/*, and "base64"
for everything else. A value of `"binary"' is generally *not*
suitable for sending anything but ASCII text files with lines
under 1000 characters, so consider using one of the other values
instead.

In the case of "7bit"/"8bit", long lines are automatically
chopped to legal length; in the case of "7bit", all 8-bit
characters are automatically *removed*. This may not be what you
want, so pick your encoding well! For more info, see the section
on "A MIME PRIMER".

FH *Alternative to "Data" or "Path".* Filehandle containing the data,
opened for reading. See "ReadNow" also.

Filename
*Optional.* The name of the attachment. You can use this to
supply a recommended filename for the end-user who is saving the
attachment to disk. You only need this if the filename at the
end of the "Path" is inadequate, or if you're using "Data"
instead of "Path". You should *not* put path information in here
(e.g., no "/" or "\" or ":" characters should be used).

Id *Optional.* Same as setting "content-id".

Length *Optional.* Set the content length explicitly. Normally, this header
is automatically computed, but only under certain circumstances
(see the section on "Limitations").

Path *Alternative to "Data" or "FH".* Path to a file containing the
data... actually, it can be any open()able expression. If it
looks like a path, the last element will automatically be
treated as the filename. See "ReadNow" also.

ReadNow *Optional, for use with "Path".* If true, will open the path and
slurp the contents into core now. This is useful if the Path
points to a command and you don't want to run the command over
and over if outputting the message several times. Fatal
exception raised if the open fails.

Top *Optional.* If defined, indicates whether or not this is a "top-
level" MIME message. The parts of a multipart message are *not*
top-level. Default is true.

Type *Optional.* The MIME content type, or one of these special values
(case-sensitive):

"TEXT" means "text/plain"
"BINARY" means "application/octet-stream"
"AUTO" means attempt to guess from the filename, falling back
to 'application/octet-stream'. This is good if you have
MIME::Types on your system and you have no idea what
file might be used for the attachment.

The default is `"TEXT"', but it will be `"AUTO"' if you set
$AUTO_CONTENT_TYPE to true (sorry, but you have to enable it
explicitly, since we don't want to break code which depends on
the old behavior).

A picture being worth 1000 words (which is of course 2000 bytes, so
it's probably more of an "icon" than a "picture", but I digress...),
here are some examples:

$msg = MIME::Lite->build(
From => '[email protected]',
To => '[email protected]',
Subject => "Hi there!",
Type => 'TEXT',
Encoding => '7bit',
Data => "Just a quick note to say hi!");

$msg = MIME::Lite->build(
From => '[email protected]',
To => '[email protected]',
Subject => "A gif for U"
Type => 'image/gif',
Path => "/home/httpd/logo.gif");

$msg = MIME::Lite->build(
From => '[email protected]',
To => '[email protected]',
Subject => "A gzipp'ed tar file",
Type => 'x-gzip',
Path => "gzip < /usr/inc/somefile.tar |",
ReadNow => 1,
Filename => "somefile.tgz");

To show you what's really going on, that last example could also
have been written:

$msg = new MIME::Lite;
$msg->build(Type => 'x-gzip',
Path => "gzip < /usr/inc/somefile.tar |",
ReadNow => 1,
Filename => "somefile.tgz");
$msg->add(From => "[email protected]");
$msg->add(To => "[email protected]");
$msg->add(Subject => "A gzipp'ed tar file");

Setting/getting headers and attributes

add TAG,VALUE
*Instance method.* Add field TAG with the given VALUE to the end of
the header. The TAG will be converted to all-lowercase, and the
VALUE will be made "safe" (returns will be given a trailing space).

Beware: any MIME fields you "add" will override any MIME attributes
I have when it comes time to output those fields. Normally, you will
use this method to add *non-MIME* fields:

$msg->add("Subject" => "Hi there!");

Giving VALUE as an arrayref will cause all those values to be added.
This is only useful for special multiple-valued fields like
"Received":

$msg->add("Received" => ["here", "there", "everywhere"]

Giving VALUE as the empty string adds an invisible placeholder to
the header, which can be used to suppress the output of the
"Content-*" fields or the special "MIME-Version" field. When
suppressing fields, you should use replace() instead of add():

$msg->replace("Content-disposition" => "");

*Note:* add() is probably going to be more efficient than
`replace()', so you're better off using it for most applications if
you are certain that you don't need to delete() the field first.

*Note:* the name comes from Mail::Header.

attr ATTR,[VALUE]
*Instance method.* Set MIME attribute ATTR to the string VALUE. ATTR
is converted to all-lowercase. This method is normally used to
set/get MIME attributes:

$msg->attr("content-type" => "text/html");
$msg->attr("content-type.charset" => "US-ASCII");
$msg->attr("content-type.name" => "homepage.html");

This would cause the final output to look something like this:

Content-type: text/html; charset=US-ASCII; name="homepage.html"

Note that the special empty sub-field tag indicates the anonymous
first sub-field.

Giving VALUE as undefined will cause the contents of the named
subfield to be deleted.

Supplying no VALUE argument just returns the attribute's value:

$type = $msg->attr("content-type"); ### returns "text/html"
$name = $msg->attr("content-type.name"); ### returns "homepage.html"

delete TAG
*Instance method.* Delete field TAG with the given VALUE to the end
of the header. The TAG will be converted to all-lowercase.

$msg->delete("Subject");

*Note:* the name comes from Mail::Header.

field_order FIELD,...FIELD
*Class/instance method.* Change the order in which header fields are
output for this object:

$msg->field_order('from', 'to', 'content-type', 'subject');

When used as a class method, changes the default settings for all
objects:

MIME::Lite->field_order('from', 'to', 'content-type', 'subject');

Case does not matter: all field names will be coerced to lowercase.
In either case, supply the empty array to restore the default
ordering.

fields
*Instance method.* Return the full header for the object, as a ref
to an array of `[TAG, VALUE]' pairs, where each TAG is all-
lowercase. Note that any fields the user has explicitly set will
override the corresponding MIME fields that we would otherwise
generate. So, don't say...

$msg->set("Content-type" => "text/html; charset=US-ASCII");

unless you want the above value to override the "Content-type" MIME
field that we would normally generate.

*Note:* I called this "fields" because the header() method of
Mail::Header returns something different, but similar enough to be
confusing.

You can change the order of the fields: see the field_order entry
elsewhere in this document . You really shouldn't need to do this,
but some people have to deal with broken mailers.

filename [FILENAME]
*Instance method.* Set the filename which this data will be reported
as. This actually sets both "standard" attributes.

With no argument, returns the filename as dictated by the content-
disposition.

get TAG,[INDEX]
*Instance method.* Get the contents of field TAG, which might have
been set with set() or replace(). Returns the text of the field.

$ml->get('Subject', 0);

If the optional 0-based INDEX is given, then we return the INDEX'th
occurence of field TAG. Otherwise, we look at the context: In a
scalar context, only the first (0th) occurence of the field is
returned; in an array context, *all* occurences are returned.

*Warning:* this should only be used with non-MIME fields. Behavior
with MIME fields is TBD, and will raise an exception for now.

get_length
*Instance method.* Recompute the content length for the message *if
the process is trivial*, setting the "content-length" attribute as a
side-effect:

$msg->get_length;

Returns the length, or undefined if not set.

*Note:* the content length can be difficult to compute, since it
involves assembling the entire encoded body and taking the length of
it (which, in the case of multipart messages, means freezing all the
sub-parts, etc.).

This method only sets the content length to a defined value if the
message is a singlepart with `"binary"' encoding, *and* the body is
available either in-core or as a simple file. Otherwise, the content
length is set to the undefined value.

Since content-length is not a standard MIME field anyway (that's
right, kids: it's not in the MIME RFCs, it's an HTTP thing), this
seems pretty fair.

parts
*Instance method.* Return the parts of this entity, and this entity
only. Returns empty array if this entity has no parts.

This is not recursive! Parts can have sub-parts; use parts_DFS() to
get everything.

parts_DFS
*Instance method.* Return the list of all MIME::Lite objects
included in the entity, starting with the entity itself, in depth-
first-search order. If this object has no parts, it alone will be
returned.

preamble [TEXT]
*Instance method.* Get/set the preamble string, assuming that this
object has subparts. Set it to undef for the default string.

replace TAG,VALUE
*Instance method.* Delete all occurences of fields named TAG, and
add a new field with the given VALUE. TAG is converted to all-
lowercase.

Beware the special MIME fields (MIME-version, Content-*): if you
"replace" a MIME field, the replacement text will override the
*actual* MIME attributes when it comes time to output that field. So
normally you use attr() to change MIME fields and add()/replace() to
change *non-MIME* fields:

$msg->replace("Subject" => "Hi there!");

Giving VALUE as the *empty string* will effectively *prevent* that
field from being output. This is the correct way to suppress the
special MIME fields:

$msg->replace("Content-disposition" => "");

Giving VALUE as *undefined* will just cause all explicit values for
TAG to be deleted, without having any new values added.

*Note:* the name of this method comes from Mail::Header.

scrub
*Instance method.* This is Alpha code. If you use it, please let me
know how it goes. Recursively goes through the "parts" tree of this
message and tries to find MIME attributes that can be removed. With
an array argument, removes exactly those attributes; e.g.:

$msg->scrub(['content-disposition', 'content-length']);

Is the same as recursively doing:

$msg->replace('Content-disposition' => '');
$msg->replace('Content-length' => '');

Setting/getting message data

binmode [OVERRIDE]
*Instance method.* With no argument, returns whether or not it
thinks that the data (as given by the "Path" argument of `build()')
should be read using binmode() (for example, when `read_now()' is
invoked).

The default behavior is that any content type other than `text/*' or
`message/*' is binmode'd; this should in general work fine.

With a defined argument, this method sets an explicit "override"
value. An undefined argument unsets the override. The new current
value is returned.

data [DATA]
*Instance method.* Get/set the literal DATA of the message. The DATA
may be either a scalar, or a reference to an array of scalars (which
will simply be joined).

*Warning:* setting the data causes the "content-length" attribute to
be recomputed (possibly to nothing).

fh [FILEHANDLE]
*Instance method.* Get/set the FILEHANDLE which contains the message
data.

Takes a filehandle as an input and stores it in the object. This
routine is similar to path(); one important difference is that no
attempt is made to set the content length.

path [PATH]
*Instance method.* Get/set the PATH to the message data.

*Warning:* setting the path recomputes any existing "content-length"
field, and re-sets the "filename" (to the last element of the path
if it looks like a simple path, and to nothing if not).

resetfh [FILEHANDLE]
*Instance method.* Set the current position of the filehandle back
to the beginning. Only applies if you used "FH" in build() or
attach() for this message.

Returns false if unable to reset the filehandle (since not all
filehandles are seekable).

read_now
*Instance method.* Forces data from the path/filehandle (as
specified by `build()') to be read into core immediately, just as
though you had given it literally with the `Data' keyword.

Note that the in-core data will always be used if available.

Be aware that everything is slurped into a giant scalar: you may not
want to use this if sending tar files! The benefit of *not* reading
in the data is that very large files can be handled by this module
if left on disk until the message is output via `print()' or
`print_body()'.

sign PARAMHASH
*Instance method.* Sign the message. This forces the message to be
read into core, after which the signature is appended to it.

Data As in `build()': the literal signature data. Can be either a scalar
or a ref to an array of scalars.

Path As in `build()': the path to the file.

If no arguments are given, the default is:

Path => "$ENV{HOME}/.signature"

The content-length is recomputed.

verify_data
*Instance method.* Verify that all "paths" to attached data exist,
recursively. It might be a good idea for you to do this before a
print(), to prevent accidental partial output if a file might be
missing. Raises exception if any path is not readable.

Output

print [OUTHANDLE]
*Instance method.* Print the message to the given output handle, or
to the currently-selected filehandle if none was given.

All OUTHANDLE has to be is a filehandle (possibly a glob ref), or
any object that responds to a print() message.

print_body [OUTHANDLE]
*Instance method.* Print the body of a message to the given output
handle, or to the currently-selected filehandle if none was given.

All OUTHANDLE has to be is a filehandle (possibly a glob ref), or
any object that responds to a print() message.

Fatal exception raised if unable to open any of the input files, or
if a part contains no data, or if an unsupported encoding is
encountered.

print_header [OUTHANDLE]
*Instance method.* Print the header of the message to the given
output handle, or to the currently-selected filehandle if none was
given.

All OUTHANDLE has to be is a filehandle (possibly a glob ref), or
any object that responds to a print() message.

as_string
*Instance method.* Return the entire message as a string, with a
header and an encoded body.

body_as_string
*Instance method.* Return the encoded body as a string. This is the
portion after the header and the blank line.

*Note:* actually prepares the body by "printing" to a scalar. Proof
that you can hand the `print*()' methods any blessed object that
responds to a `print()' message.

header_as_string
*Instance method.* Return the header as a string.

Sending

send
send HOW, HOWARGS...
*Class/instance method.* This is the principal method for sending
mail, and for configuring how mail will be sent.

*As a class method* with a HOW argument and optional HOWARGS, it
sets the default sending mechanism that the no-argument instance
method will use. The HOW is a facility name (see below), and the
HOWARGS is interpreted by the facilty. The class method returns the
previous HOW and HOWARGS as an array.

MIME::Lite->send('sendmail', "d:\\programs\\sendmail.exe");
...
$msg = MIME::Lite->new(...);
$msg->send;

*As an instance method with arguments* (a HOW argument and optional
HOWARGS), sends the message in the requested manner; e.g.:

$msg->send('sendmail', "d:\\programs\\sendmail.exe");

*As an instance method with no arguments,* sends the message by the
default mechanism set up by the class method. Returns whatever the
mail-handling routine returns: this should be true on success,
false/exception on error:

$msg = MIME::Lite->new(From=>...);
$msg->send || die "you DON'T have mail!";

On Unix systems (at least), the default setting is equivalent to:

MIME::Lite->send("sendmail", "/usr/lib/sendmail -t -oi -oem");

There are three facilities:

"sendmail", ARGS...
Send a message by piping it into the "sendmail" command. Uses
the send_by_sendmail() method, giving it the ARGS. This usage
implements (and deprecates) the `sendmail()' method.

"smtp", [HOSTNAME]
Send a message by SMTP, using optional HOSTNAME as SMTP-sending
host. Uses the send_by_smtp() method.

"sub", \&SUBREF, ARGS...
Sends a message MSG by invoking the subroutine SUBREF of your
choosing, with MSG as the first argument, and ARGS following.

*For example:* let's say you're on an OS which lacks the usual Unix
"sendmail" facility, but you've installed something a lot like it,
and you need to configure your Perl script to use this
"sendmail.exe" program. Do this following in your script's setup:

MIME::Lite->send('sendmail', "d:\\programs\\sendmail.exe");

Then, whenever you need to send a message $msg, just say:

$msg->send;

That's it. Now, if you ever move your script to a Unix box, all you
need to do is change that line in the setup and you're done. All of
your $msg->send invocations will work as expected.

send_by_sendmail SENDMAILCMD
send_by_sendmail PARAM=>VALUE, ...
*Instance method.* Send message via an external "sendmail" program
(this will probably only work out-of-the-box on Unix systems).

Returns true on success, false or exception on error.

You can specify the program and all its arguments by giving a single
string, SENDMAILCMD. Nothing fancy is done; the message is simply
piped in.

However, if your needs are a little more advanced, you can specify
zero or more of the following PARAM/VALUE pairs; a Unix-style,
taint-safe "sendmail" command will be constructed for you:

Sendmail
Full path to the program to use. Default is "/usr/lib/sendmail".

BaseArgs
Ref to the basic array of arguments we start with. Default is
`["-t", "-oi", "-oem"]'.

SetSender
Unless this is *explicitly* given as false, we attempt to
automatically set the `-f' argument to the first address that
can be extracted from the "From:" field of the message (if there
is one).

*What is the -f, and why do we use it?* Suppose we did *not* use
`-f', and you gave an explicit "From:" field in your message: in
this case, the sendmail "envelope" would indicate the *real*
user your process was running under, as a way of preventing mail
forgery. Using the `-f' switch causes the sender to be set in
the envelope as well.

*So when would I NOT want to use it?* If sendmail doesn't regard
you as a "trusted" user, it will permit the `-f' but also add an
"X-Authentication-Warning" header to the message to indicate a
forged envelope. To avoid this, you can either (1) have
SetSender be false, or (2) make yourself a trusted user by
adding a `T' configuration command to your *sendmail.cf* file
(e.g.: `Teryq' if the script is running as user "eryq").

FromSender
If defined, this is identical to setting SetSender to true,
except that instead of looking at the "From:" field we use the
address given by this option. Thus:

FromSender => '[email protected]'

send_by_smtp ARGS...
*Instance method.* Send message via SMTP, using Net::SMTP. The
optional ARGS are sent into Net::SMTP::new(): usually, these are

MAILHOST, OPTION=>VALUE, ...

Note that the list of recipients is taken from the "To", "Cc" and
"Bcc" fields.

Returns true on success, false or exception on error.

sendmail COMMAND...
*Class method, DEPRECATED.* Declare the sender to be "sendmail", and
set up the "sendmail" command. *You should use send() instead.*

Miscellaneous

quiet ONOFF
*Class method.* Suppress/unsuppress all warnings coming from this
module.

MIME::Lite->quiet(1); ### I know what I'm doing

I recommend that you include that comment as well. And while you
type it, say it out loud: if it doesn't feel right, then maybe you
should reconsider the whole line. `;-)'

NOTES
How do I prevent "Content" headers from showing up in my mail reader?

Apparently, some people are using mail readers which display the MIME
headers like "Content-disposition", and they want MIME::Lite not to
generate them "because they look ugly".

Sigh.

Y'know, kids, those headers aren't just there for cosmetic purposes.
They help ensure that the message is *understood* correctly by mail
readers. But okay, you asked for it, you got it... here's how you can
suppress the standard MIME headers. Before you send the message, do
this:

$msg->scrub;

You can scrub() any part of a multipart message independently; just be
aware that it works recursively. Before you scrub, note the rules that I
follow:

Content-type
You can safely scrub the "content-type" attribute if, and only if,
the part is of type "text/plain" with charset "us-ascii".

Content-transfer-encoding
You can safely scrub the "content-transfer-encoding" attribute if,
and only if, the part uses "7bit", "8bit", or "binary" encoding. You
are far better off doing this if your lines are under 1000
characters. Generally, that means you *can* scrub it for plain text,
and you can *not* scrub this for images, etc.

Content-disposition
You can safely scrub the "content-disposition" attribute if you
trust the mail reader to do the right thing when it decides whether
to show an attachment inline or as a link. Be aware that scrubbing
both the content-disposition and the content-type means that there
is no way to "recommend" a filename for the attachment!

Note: there are reports of brain-dead MUAs out there that do the
wrong thing if you *provide* the content-disposition. If your
attachments keep showing up inline or vice-versa, try scrubbing this
attribute.

Content-length
You can always scrub "content-length" safely.

How do I give my attachment a [different] recommended filename?

By using the Filename option (which is different from Path!):

$msg->attach(Type => "image/gif",
Path => "/here/is/the/real/file.GIF",
Filename => "logo.gif");

You should *not* put path information in the Filename.

Benign limitations

This is "lite", after all...

* There's no parsing. Get MIME-tools if you need to parse MIME messages.

* MIME::Lite messages are currently *not* interchangeable with either
Mail::Internet or MIME::Entity objects. This is a completely
separate module.

* A content-length field is only inserted if the encoding is binary, the
message is a singlepart, and all the document data is available at
`build()' time by virtue of residing in a simple path, or in-core.
Since content-length is not a standard MIME field anyway (that's
right, kids: it's not in the MIME RFCs, it's an HTTP thing), this
seems pretty fair.

* MIME::Lite alone cannot help you lose weight. You must supplement your
use of MIME::Lite with a healthy diet and exercise.

Cheap and easy mailing

I thought putting in a default "sendmail" invocation wasn't too bad an
idea, since a lot of Perlers are on UNIX systems. The out-of-the-box
configuration is:

MIME::Lite->send('sendmail', "/usr/lib/sendmail -t -oi -oem");

By the way, these arguments to sendmail are:

-t Scan message for To:, Cc:, Bcc:, etc.

-oi Do NOT treat a single "." on a line as a message terminator.
As in, "-oi vey, it truncated my message... why?!"

-oem On error, mail back the message (I assume to the
appropriate address, given in the header).
When mail returns, circle is complete. Jai Guru Deva -oem.

Note that these are the same arguments you get if you configure to use
the smarter, taint-safe mailing:

MIME::Lite->send('sendmail');

If you get "X-Authentication-Warning" headers from this, you can forgo
diddling with the envelope by instead specifying:

MIME::Lite->send('sendmail', SetSender=>0);

And, if you're not on a Unix system, or if you'd just rather send mail
some other way, there's always:

MIME::Lite->send('smtp', "smtp.myisp.net");

Or you can set up your own subroutine to call. In any case, check out
the send() method.

WARNINGS
Good-vs-bad email addresses with send_by_smtp()

If using send_by_smtp(), be aware that you are forcing MIME::Lite to
extract email addresses out of a possible list provided in the `To:',
`Cc:', and `Bcc:' fields. This is tricky stuff, and as such only the
following sorts of addresses will work reliably:

username
[email protected]
"Name, Full" <[email protected]>

This last form is discouraged because SMTP must be able to get at the
*name* or *name@domain* portion.

Disclaimer: MIME::Lite was never intended to be a Mail User Agent, so
please don't expect a full implementation of RFC-822. Restrict yourself
to the common forms of Internet addresses described herein, and you
should be fine. If this is not feasible, then consider using MIME::Lite
to *prepare* your message only, and using Net::SMTP explicitly to *send*
your message.

Formatting of headers delayed until print()

This class treats a MIME header in the most abstract sense, as being a
collection of high-level attributes. The actual RFC-822-style header
fields are not constructed until it's time to actually print the darn
thing.

Encoding of data delayed until print()

When you specify message bodies (in build() or attach()) -- whether by
FH, Data, or Path -- be warned that we don't attempt to open files, read
filehandles, or encode the data until print() is invoked.

In the past, this created some confusion for users of sendmail who gave
the wrong path to an attachment body, since enough of the print() would
succeed to get the initial part of the message out. Nowadays,
$AUTO_VERIFY is used to spot-check the Paths given before the mail
facility is employed. A whisker slower, but tons safer.

Note that if you give a message body via FH, and try to print() a
message twice, the second print() will not do the right thing unless you
explicitly rewind the filehandle.

You can get past these difficulties by using the ReadNow option,
provided that you have enough memory to handle your messages.

MIME attributes are separate from header fields!

Important: the MIME attributes are stored and manipulated separately
from the message header fields; when it comes time to print the header
out, *any explicitly-given header fields override the ones that would be
created from the MIME attributes.* That means that this:

### DANGER ### DANGER ### DANGER ### DANGER ### DANGER ###
$msg->add("Content-type", "text/html; charset=US-ASCII");

will set the exact `"Content-type"' field in the header I write,
*regardless of what the actual MIME attributes are.*

*This feature is for experienced users only,* as an escape hatch in case
the code that normally formats MIME header fields isn't doing what you
need. And, like any escape hatch, it's got an alarm on it: MIME::Lite
will warn you if you attempt to `set()' or `replace()' any MIME header
field. Use `attr()' instead.

Beware of lines consisting of a single dot

Julian Haight noted that MIME::Lite allows you to compose messages with
lines in the body consisting of a single ".". This is true: it should be
completely harmless so long as "sendmail" is used with the -oi option
(see the section on "Cheap and easy mailing").

However, I don't know if using Net::SMTP to transfer such a message is
equally safe. Feedback is welcomed.

My perspective: I don't want to magically diddle with a user's message
unless absolutely positively necessary. Some users may want to send
files with "." alone on a line; my well-meaning tinkering could
seriously harm them.

Infinite loops may mean tainted data!

Stefan Sautter noticed a bug in 2.106 where a m//gc match was failing
due to tainted data, leading to an infinite loop inside MIME::Lite.

I am attempting to correct for this, but be advised that my fix will
silently untaint the data (given the context in which the problem
occurs, this should be benign: I've labelled the source code with
UNTAINT comments for the curious).

So: don't depend on taint-checking to save you from outputting tainted
data in a message.

Don't tweak the global configuration

Global configuration variables are bad, and should go away. Until they
do, please follow the hints with each setting on how *not* to change it.

A MIME PRIMER
Content types

The "Type" parameter of `build()' is a *content type*. This is the
actual type of data you are sending. Generally this is a string of the
form `"majortype/minortype"'.

Here are the major MIME types. A more-comprehensive listing may be found
in RFC-2046.

application
Data which does not fit in any of the other categories, particularly
data to be processed by some type of application program.
`application/octet-stream', `application/gzip',
`application/postscript'...

audio
Audio data. `audio/basic'...

image
Graphics data. `image/gif', `image/jpeg'...

message
A message, usually another mail or MIME message. `message/rfc822'...

multipart
A message containing other messages. `multipart/mixed',
`multipart/alternative'...

text
Textual data, meant for humans to read. `text/plain', `text/html'...

video
Video or video+audio data. `video/mpeg'...

Content transfer encodings

The "Encoding" parameter of `build()'. This is how the message body is
packaged up for safe transit.

Here are the 5 major MIME encodings. A more-comprehensive listing may be
found in RFC-2045.

7bit
Basically, no *real* encoding is done. However, this label
guarantees that no 8-bit characters are present, and that lines do
not exceed 1000 characters in length.

8bit
Basically, no *real* encoding is done. The message might contain 8-
bit characters, but this encoding guarantees that lines do not
exceed 1000 characters in length.

binary
No encoding is done at all. Message might contain 8-bit characters,
and lines might be longer than 1000 characters long.

The most liberal, and the least likely to get through mail gateways.
Use sparingly, or (better yet) not at all.

base64
Like "uuencode", but very well-defined. This is how you should send
essentially binary information (tar files, GIFs, JPEGs, etc.).

quoted-printable
Useful for encoding messages which are textual in nature, yet which
contain non-ASCII characters (e.g., Latin-1, Latin-2, or any other
8-bit alphabet).

VERSION
$Id: Lite.pm,v 2.117 2001/08/20 20:40:39 eryq Exp $

CHANGE LOG
Version 2.117 (2001/08/20)
The terms-of-use have been placed in the distribution file
"COPYING". Also, small documentation tweaks were made.

Version 2.116 (2001/08/17)
Added long-overdue patch which makes the instance method form of
send() do the right thing when given HOW... arguments. *Thanks to
Casey West for the patch.*

Version 2.114 (2001/08/16)
New special 'AUTO' content type in new()/build() tells MIME::Lite to
try and guess the type from file extension. To make use of this,
you'll want to install MIME::Types. The "AUTO" setting can be made
the default default (instead of "TEXT") if you set
`$AUTO_CONTENT_TYPE = 1, $PARANOID = 0'. *Thanks to* Ville
SkyttE<#228> *for these patches.*

File::Basename is used if it is available. *Thanks to* Ville
SkyttE<#228> *for this patch.*

SMTP failures (in send_by_smtp) now add the $smtp->message to the
croak'ed exception, so if things go wrong, you get a better idea of
what and why. *Thanks to Thomas R. Wyant III for the patch.*

Made a subtle change to `as_string' which supposedly fixes a failed
MIME data.t test with Perl 5.004_04 on NT 4 sp6. The problem might
only exist in this old perl, but as the patch author says, not
everyone has climbed higher on the Perl ladder. *Thanks to John
Gotts for the patch.*

Added `contrib' directory, with MailTool.pm. *Thanks to Tom Wyant
for this contribution.*

Improved HTML documentation (notice the links to the individual
methods in the top menu).

Corrected some mis-docs.

Version 2.111 (2001/04/03)
Added long-overdue `parts()' and `parts_DFS()' methods.

No instance method
For accessing the subparts?
That can't be right. D'OH!

Added long-overdue auto-verify logic to `print()' method.

Added long-overdue `preamble()' method for getting/setting the
preamble text. *Thanks to Jim Daigle for inspiring this.*

Version 2.108 (2001/03/30)
New `field_order()' allows you to set the header order, both on a
per-message basis, and package-wide. *Thanks to Thomas Stromberg for
suggesting this.*

Added code to try and divine "sendmail" path more intelligently.
*Thanks to Slaven Rezic for the suggestion.*

Version 2.107 (2001/03/27)
Fixed serious bug where tainted data with quoted-printable encoding
was causing infinite loops. The "fix" untaints the data in question,
which is not optimal, but it's probably benign in this case. *Thanks
to Stefan Sautter for tracking this nasty little beast down.*
*Thanks to Larry Geralds for a related patch.*

"Doctor, O doctor:
it's painful when I do *this* --"
"Simple: don't *do* that."

Fixed bugs where a non-local `$_' was being modified... again! Will
I never learn? *Thanks to Maarten Koskamp for reporting this.*

Dollar-underscore
can poison distant waters;
'local' must it be.

Fixed buglet in `add()' where all value references were being
treated as arrayrefs, instead of as possibly-self-stringifying
object refs. Now you can send in an object ref as the 2nd argument.
*Thanks to dLux for the bug report.*

That ref is a string?
Operator overload
has ruined my day.

Added "Approved" as an acceptable header field for `new()', as per
RFC1036. *Thanks to Thomax for the suggestion regarding MIME-tools.*

Small improvements to docs to make different uses of attach() and
various arguments clearer. *Thanks to Sven Rassman and Roland Walter
for the suggestions.*

Version 2.106 (2000/11/21)
Added Alpha version of scrub() to make it easy for people to
suppress the printing of unwanted MIME attributes (like Content-
length). *Thanks to the many people who asked for this.*

Headers with empty-strings for their values are no longer printed.
This seems sensible, and helps us implement scrub().

Version 2.105 (2000/10/14)
The regression-test failure was identified, and it was my fault.
Apparently some of the \-quoting in my "autoloaded" code was making
Perl 5.6 unhappy. For this nesting-related idiocy, a nesting kaiku.
*Thanks to Scott Schwartz for identifying the problem.*

In a pattern, my
backslash-s dwells peacefully,
unambiguous --

but I embed it
in a double-quoted string
doubling the backslash --

interpolating
that same double-quoted string
in other patterns --

and, worlds within worlds,
I single-quote the function
to autoload it --

changing the meaning
of the backslash and the 's';
and Five-Point-Six growls.

Version 2.104 (2000/09/28)
Now attempts to load and use Mail::Address for parsing email
addresses *before* falling back to our own method. *Thanks to
numerous people for suggesting this.*

Parsing addresses
is too damn hard. One last hope:
Let Graham Barr do it!

For the curious, the version of Mail::Address appears as the "A"
number in the X-Mailer:

X-Mailer: MIME::Lite 2.104 (A1.15; B2.09; Q2.03)

Added FromSender option to send_by_sendmail(). *Thanks to Bill
Moseley for suggesting this feature.*

Version 2.101 (2000/06/06)
Major revision to print_body() and body_as_string() so that "body"
really means "the part after the header", which is what most people
would want in this context. This is not how it was used 1.x, where
"body" only meant "the body of a simple singlepart". Hopefully, this
change will solve many problems and create very few ones.

Added support for attaching a part to a "message/rfc822", treating
the "message" type as a multipart-like container.

Now takes care not to include "Bcc:" in header when using
send_by_smtp, as a safety precaution against qmail's behavior.
*Thanks to Tatsuhiko Miyagawa for identifying this problem.*

Improved efficiency of many stringifying operations by using string-
arrays which are joined, instead of doing multiple appends to a
scalar.

Cleaned up the "examples" directory.

Version 1.147 (2000/06/02)
Fixed buglet where lack of Cc:/Bcc: was causing extract_addrs to
emit "undefined variable" warnings. Also, lack of a "To:" field now
causes a croak. *Thanks to David Mitchell for the bug report and
suggested patch.*

Version 1.146 (2000/05/18)
Fixed bug in parsing of addresses; please read the WARNINGS section
which describes recommended address formats for "To:", "Cc:", etc.
Also added automatic inclusion of a UT "Date:" at top level unless
explicitly told not to. *Thanks to Andy Jacobs for the bug report
and the suggestion.*

Version 1.145 (2000/05/06)
Fixed bug in encode_7bit(): a lingering `/e' modifier was removed.
*Thanks to Michael A. Chase for the patch.*

Version 1.142 (2000/05/02)
Added new, taint-safe invocation of "sendmail", one which also sets
up the `-f' option. Unfortunately, I couldn't make this automatic:
the change could have broken a lot of code out there which used
send_by_sendmail() with unusual "sendmail" variants. So you'll have
to configure "send" to use the new mechanism:

MIME::Lite->send('sendmail'); ### no args!

*Thanks to Jeremy Howard for suggesting these features.*

Version 1.140 (2000/04/27)
Fixed bug in support for "To", "Cc", and "Bcc" in send_by_smtp():
multiple (comma-separated) addresses should now work fine. We try
real hard to extract addresses from the flat text strings. *Thanks
to John Mason for motivating this change.*

Added automatic verification that attached data files exist, done
immediately before the "send" action is invoked. To turn this off,
set $MIME::Lite::AUTO_VERIFY to false.

Version 1.137 (2000/03/22)
Added support for "Cc" and "Bcc" in send_by_smtp(). To turn this
off, set $MIME::Lite::AUTO_CC to false. *Thanks to Lucas Maneos for
the patch, and tons of others for the suggestion.*

Chooses a better default content-transfer-encoding if the content-
type is "image/*", "audio/*", etc. To turn this off, set
$MIME::Lite::AUTO_ENCODE to false. *Thanks to many folks for the
suggestion.*

Fixed bug in QP-encoding where a non-local `$_' was being modified.
*Thanks to Jochen Stenzel for finding this very obscure bug!*

Removed references to `$`', `$'', and `$&' (bad variables which slow
things down).

Added an example of how to send HTML files with enclosed in-line
images, per popular demand.

Version 1.133 (1999/04/17)
Fixed bug in "Data" handling: arrayrefs were not being handled
properly.

Version 1.130 (1998/12/14)
Added much larger and more-flexible send() facility. *Thanks to
Andrew McRae (and Optimation New Zealand Ltd) for the Net::SMTP
interface. Additional thanks to the many folks who requested this
feature.*

Added get() method for extracting basic attributes.

New... "t" tests!

Version 1.124 (1998/11/13)
Folded in filehandle (FH) support in build/attach. *Thanks to Miko
O'Sullivan for the code.*

Version 1.122 (1998/01/19)
MIME::Base64 and MIME::QuotedPrint are used if available.

The 7bit encoding no longer does "escapes"; it merely strips 8-bit
characters.

Version 1.121 (1997/04/08)
Filename attribute is now no longer ignored by build(). *Thanks to
Ian Smith for finding and patching this bug.*

Version 1.120 (1997/03/29)
Efficiency hack to speed up MIME::Lite::IO_Scalar. *Thanks to David
Aspinwall for the patch.*

Version 1.116 (1997/03/19)
Small bug in our private copy of encode_base64() was patched.
*Thanks to Andreas Koenig for pointing this out.*

New, prettier way of specifying mail message headers in `build()'.

New quiet method to turn off warnings.

Changed "stringify" methods to more-standard "as_string" methods.

Version 1.112 (1997/03/06)
Added `read_now()', and `binmode()' method for our non-Unix-using
brethren: file data is now read using binmode() if appropriate.
*Thanks to Xiangzhou Wang for pointing out this bug.*

Version 1.110 (1997/03/06)
Fixed bug in opening the data filehandle.

Version 1.102 (1997/03/01)
Initial release.

Version 1.101 (1997/03/01)
Baseline code. Originally created: 11 December 1996. Ho ho ho.

TERMS AND CONDITIONS
Copyright (c) 1997 by Eryq. Copyright (c) 1998 by ZeeGee Software Inc.
All rights reserved. This program is free software; you can redistribute
it and/or modify it under the same terms as Perl itself.

This software comes with NO WARRANTY of any kind. See the COPYING file
in the distribution for details.

NUTRITIONAL INFORMATION
For some reason, the US FDA says that this is now required by law on any
products that bear the name "Lite"...

MIME::Lite |
------------------------------------------------------------
Serving size: | 1 module
Servings per container: | 1
Calories: | 0
Fat: | 0g
Saturated Fat: | 0g

Warning: for consumption by hardware only! May produce indigestion in
humans if taken internally.

AUTHOR
Eryq ([email protected]). President, ZeeGee Software Inc.
(http://www.zeegee.com).

Go to http://www.zeegee.com for the latest downloads and on-line
documentation for this module. Enjoy.