=encoding utf8

=head1 NAME

Mail::Message::Field::Full - construct one smart line in a message header

=head1 INHERITANCE

 Mail::Message::Field::Full
   is a Mail::Message::Field
   is a Mail::Reporter

 Mail::Message::Field::Full is extended by
   Mail::Message::Field::Structured
   Mail::Message::Field::Unstructured

=head1 SYNOPSIS

  # Getting to understand the complexity of a header field ...

  my $fast = $msg->head->get('subject');
  my $full = Mail::Message::Field::Full->from($fast);

  my $full = $msg->head->get('subject')->study;  # same
  my $full = $msg->head->study('subject');       # same
  my $full = $msg->study('subject');             # same

  # ... or build a complex header field yourself

  my $f = Mail::Message::Field::Full->new('To');
  my $f = Mail::Message::Field::Full->new('Subject: hi!');
  my $f = Mail::Message::Field::Full->new(Subject => 'hi!');

=head1 DESCRIPTION

This is the I<full> implementation of a header field: it has I<full>
understanding of all predefined header fields.  These objects will be
quite slow, because header fields can be very complex.  Of course, this
class delivers the optimal result, but for a quite large penalty in
performance and memory consumption.  Are you willing to accept?

This class supports the common header description from RFC2822 (formerly
RFC822), the extensions with respect to character set encodings as specified
in RFC2047, and the extensions on language specification and long parameter
wrapping from RFC2231.  If you do not need the latter two, then the
L<Mail::Message::Field::Fast|Mail::Message::Field::Fast> and L<Mail::Message::Field::Flex|Mail::Message::Field::Flex>
are enough for your application.

RFC5322 (L<https://www.rfc-editor.org/rfc/rfc5322.html>) describes a
long list of obsolete syntax for structured header fields.  This mainly
refers to disallowing white-spaces and folding on many inconvenient
locations.  This matches MailBox's natural behavior.

Extends L<"DESCRIPTION" in Mail::Message::Field|Mail::Message::Field/"DESCRIPTION">.

=head1 OVERLOADED

Extends L<"OVERLOADED" in Mail::Message::Field|Mail::Message::Field/"OVERLOADED">.

=over 4

=item overload: B<""> stringification

In string context, the decoded body is returned, as if L<decodedBody()|Mail::Message::Field::Full/"Access to the body">
would have been called.
Improves base, see L<Mail::Message::Field/"OVERLOADED">

=item overload: B<0+> numification

Inherited, see L<Mail::Message::Field/"OVERLOADED">

=item overload: B<<=>> numeric comparison

Inherited, see L<Mail::Message::Field/"OVERLOADED">

=item overload: B<bool> boolean

Inherited, see L<Mail::Message::Field/"OVERLOADED">

=item overload: B<cmp> string comparison

Inherited, see L<Mail::Message::Field/"OVERLOADED">

=back

=head1 METHODS

Extends L<"METHODS" in Mail::Message::Field|Mail::Message::Field/"METHODS">.

=head2 Constructors

Extends L<"Constructors" in Mail::Message::Field|Mail::Message::Field/"Constructors">.

=over 4

=item $obj-E<gt>B<clone>()

Inherited, see L<Mail::Message::Field/"Constructors">

=item $class-E<gt>B<from>($field, %options)

Convert any C<$field> (a L<Mail::Message::Field|Mail::Message::Field> object) into a new
L<Mail::Message::Field::Full|Mail::Message::Field::Full> object.  This conversion is done the hard
way: the string which is produced by the original object is parsed
again.  Usually, the string which is parsed is exactly the line (or lines)
as found in the original input source, which is a good thing because Full
fields are much more careful with the actual content.

C<%options> are passed to the constructor (see L<new()|Mail::Message::Field::Full/"METHODS">).  In any case, some
extensions of this Full field class is returned.  It depends on which
field is created what kind of class we get.

» example: 

  my $fast = $msg->head->get('subject');
  my $full = Mail::Message::Field::Full->from($fast);

  my $full = $msg->head->get('subject')->study;  # same
  my $full = $msg->head->study('subject');       # same
  my $full = $msg->get('subject');               # same

=item $class-E<gt>B<new>(($line | $name, [$body], @attributes), %options)

Creating a new field object the correct way is a lot of work, because
there is so much freedom in the RFCs, but at the same time so many
restrictions.  Most fields are implemented, but if you have your own
field (and do no want to contribute it to MailBox), then simply call
new on your own package.

You have the choice to instantiate the object as string or in prepared
parts:

=over 4

=item * B<new>(C<$line>, C<%options>)

Pass a C<$line> as it could be found in a file: a (possibly folded) line
which is terminated by a new-line.

=item * B<new>(C<$name>, [C<$body>], C<@attributes>, C<%options>)

A set of values which shape the line.

=back

The C<$name> is a wellformed header name (you may use L<wellformedName()|Mail::Message::Field/"Access to the name">) to
be sure about the casing.  The C<$body> is a string, one object, or an
ARRAY of objects.  In case of objects, they must fit to the
constructor of the field: the types which are accepted may differ.
The optional ATTRIBUTE list contains L<Mail::Message::Field::Attribute|Mail::Message::Field::Attribute>
objects.  Finally, there are some OPTIONS.
Improves base, see L<Mail::Message::Field/"Constructors">

 -Option  --Default
  charset   undef
  encoding  'q'
  force     false
  language  undef

=over 2

=item charset => $charset

The body is specified in UTF-8, and must become 7-bits ascii to be
transmited.  Specify a C<$charset> to which the multi-byte utf8 is converted
before it gets encoded.  See L<encode()|Mail::Message::Field::Full/"Internals">, which does the job.

=item encoding => 'q'|'Q'|'b'|'B'

Non-ascii characters are encoded using Quoted-Printable ('q' or 'Q') or
Base64 ('b' or 'B') encoding.

=item force => BOOLEAN

Enforce encoding in the specified charset, even when it is not needed
because the body does not contain any non-ascii characters.

=item language => $language

The C<$language> used can be specified, however is rarely used my mail clients.

=back

» example: 

  my $s = Mail::Message::Field::Full->new('Subject: Hello World');
  my $s = Mail::Message::Field::Full->new('Subject', 'Hello World');

  my @attrs   = (Mail::Message::Field::Attribute->new(...), ...);
  my @options = (extra => 'the color blue');
  my $t = Mail::Message::Field::Full->new(To => \@addrs, @attrs, @options);

=back

=head2 Attributes

Extends L<"Attributes" in Mail::Message::Field|Mail::Message::Field/"Attributes">.

=head2 The field

Extends L<"The field" in Mail::Message::Field|Mail::Message::Field/"The field">.

=over 4

=item $any-E<gt>B<isStructured>( [$name] )

Inherited, see L<Mail::Message::Field/"The field">

=item $obj-E<gt>B<length>()

Inherited, see L<Mail::Message::Field/"The field">

=item $obj-E<gt>B<nrLines>()

Inherited, see L<Mail::Message::Field/"The field">

=item $obj-E<gt>B<print>( [$fh] )

Inherited, see L<Mail::Message::Field/"The field">

=item $obj-E<gt>B<size>()

Inherited, see L<Mail::Message::Field/"The field">

=item $obj-E<gt>B<string>( [$wrap] )

Inherited, see L<Mail::Message::Field/"The field">

=item $obj-E<gt>B<toDisclose>()

Inherited, see L<Mail::Message::Field/"The field">

=back

=head2 Access to the name

Extends L<"Access to the name" in Mail::Message::Field|Mail::Message::Field/"Access to the name">.

=over 4

=item $obj-E<gt>B<Name>()

Inherited, see L<Mail::Message::Field/"Access to the name">

=item $obj-E<gt>B<name>()

Inherited, see L<Mail::Message::Field/"Access to the name">

=item $obj-E<gt>B<wellformedName>( [STRING] )

Inherited, see L<Mail::Message::Field/"Access to the name">

=back

=head2 Access to the body

Extends L<"Access to the body" in Mail::Message::Field|Mail::Message::Field/"Access to the body">.

=over 4

=item $obj-E<gt>B<body>()

Inherited, see L<Mail::Message::Field/"Access to the body">

=item $obj-E<gt>B<decodedBody>(%options)

Returns the unfolded body of the field, where encodings are resolved.  The
returned line will still contain comments and such.  The C<%options> are passed
to the decoder, see L<decode()|Mail::Message::Field::Full/"Internals">.

BE WARNED: if the field is a structured field, the content may change syntax,
because of encapsulated special characters.  By default, the body is decoded
as text, which results in a small difference within comments as well (read the RFC).

=item $obj-E<gt>B<folded>()

Inherited, see L<Mail::Message::Field/"Access to the body">

=item $obj-E<gt>B<foldedBody>( [$body] )

Inherited, see L<Mail::Message::Field/"Access to the body">

=item $any-E<gt>B<stripCFWS>( [STRING] )

Inherited, see L<Mail::Message::Field/"Access to the body">

=item $obj-E<gt>B<unfoldedBody>( [$body, [$wrap]] )

Inherited, see L<Mail::Message::Field/"Access to the body">

=back

=head2 Access to the content

Extends L<"Access to the content" in Mail::Message::Field|Mail::Message::Field/"Access to the content">.

=over 4

=item $obj-E<gt>B<addresses>()

Inherited, see L<Mail::Message::Field/"Access to the content">

=item $obj-E<gt>B<attribute>( $name, [$value] )

Inherited, see L<Mail::Message::Field/"Access to the content">

=item $obj-E<gt>B<attributes>()

Inherited, see L<Mail::Message::Field/"Access to the content">

=item $obj-E<gt>B<beautify>()

For structured header fields, this removes the original encoding of the
field's body (the format as it was offered to L<parse()|Mail::Message::Field::Full/"Parsing">), therefore the
next request for the field will have to re-produce the read data clean
and nice.  For unstructured bodies, this method doesn't do a thing.

=item $obj-E<gt>B<comment>( [STRING] )

Inherited, see L<Mail::Message::Field/"Access to the content">

=item $any-E<gt>B<createComment>(STRING, %options)

Create a comment to become part in a field.  Comments are automatically
included within parenthesis.  Matching pairs of parenthesis are
permitted within the C<STRING>.  When a non-matching parenthesis are used,
it is only permitted with an escape (a backslash) in front of them.
These backslashes will be added automatically if needed (don't worry!).
Backslashes will stay, except at the end, where it will be doubled.

The C<%options> are C<charset>, C<language>, and C<encoding> as always.
The created comment is returned.

=item $any-E<gt>B<createPhrase>(STRING, %options)

A phrase is a text which plays a well defined role.  This is the main
difference with comments, which have do specified meaning.  Some special
characters in the phrase will cause it to be surrounded with double
quotes: do not specify them yourself.

The C<%options> are C<charset>, C<language>, and C<encoding>, as always.

=item $obj-E<gt>B<study>()

Inherited, see L<Mail::Message::Field/"Access to the content">

=item $any-E<gt>B<toDate>( [$time] )

Inherited, see L<Mail::Message::Field/"Access to the content">

=item $obj-E<gt>B<toInt>()

Inherited, see L<Mail::Message::Field/"Access to the content">

=back

=head2 Other methods

Extends L<"Other methods" in Mail::Message::Field|Mail::Message::Field/"Other methods">.

=over 4

=item $any-E<gt>B<dateToTimestamp>(STRING)

Inherited, see L<Mail::Message::Field/"Other methods">

=back

=head2 Internals

Extends L<"Internals" in Mail::Message::Field|Mail::Message::Field/"Internals">.

=over 4

=item $obj-E<gt>B<consume>( $line | <$name,<$body|$objects>> )

Inherited, see L<Mail::Message::Field/"Internals">

=item $any-E<gt>B<decode>(STRING, %options)

Decode field encoded C<STRING> to an utf8 string.  The input C<STRING> is part of
a header field, and as such, may contain encoded words in C<=?...?.?...?=>
format defined by RFC2047.  The C<STRING> may contain multiple encoded parts,
maybe using different character sets.

Be warned:  you MUST first interpret the field into parts, like phrases and
comments, and then decode each part separately, otherwise the decoded text
may interfere with your markup characters.

Be warned: language information, which is defined in RFC2231, is ignored.

Encodings with unknown charsets are left untouched [requires v2.085,
otherwise croaked].  Unknown characters within an charset are replaced by
a '?'.

 -Option --Default
  is_text  true

=over 2

=item is_text => BOOLEAN

Encoding on text is slightly more complicated than encoding structured data,
because it contains blanks.  Visible blanks have to be ignored between two
encoded words in the text, but not when an encoded word follows or precedes
an unencoded word.  Phrases and comments are texts.

=back

» example: 

  print Mail::Message::Field::Full->decode('=?iso-8859-1?Q?J=F8rgen?=');
     # prints   JE<0slash>rgen

=item $obj-E<gt>B<defaultWrapLength>( [$length] )

Inherited, see L<Mail::Message::Field/"Internals">

=item $obj-E<gt>B<encode>(STRING, %options)

Encode the (possibly utf8 encoded) C<STRING> to a string which is acceptable
to the RFC2047 definition of a header: only containing us-ascii characters.

 -Option  --Default
  charset   'us-ascii'
  encoding  'q'
  force     false
  language  undef
  name      undef

=over 2

=item charset => STRING

C<STRING> is an utf8 string which has to be translated into any byte-wise
character set for transport, because MIME-headers can only contain ascii
characters.

=item encoding => 'q'|'Q'|'b'|'B'

The character encoding to be used.  With C<q> or C<Q>, quoted-printable
encoding will be used.  With C<b > or C<B >, base64 encoding will be taken.

=item force => BOOLEAN

Encode the string, even when it only contains us-ascii characters.  By
default, this is off because it decreases readibility of the produced
header fields.

=item language => STRING

RFC2231 defines how to specify language encodings in encoded words.  The
C<STRING> is a strandard iso language name.

=item name => STRING

[3.002] When the name of the field is given, the first encoded line will
be shorter.

=back

=item $any-E<gt>B<fold>( $name, $body, [$maxchars] )

Inherited, see L<Mail::Message::Field/"Internals">

=item $obj-E<gt>B<setWrapLength>( [$length] )

Inherited, see L<Mail::Message::Field/"Internals">

=item $obj-E<gt>B<stringifyData>(STRING|ARRAY|$objects)

Inherited, see L<Mail::Message::Field/"Internals">

=item $obj-E<gt>B<unfold>(STRING)

Inherited, see L<Mail::Message::Field/"Internals">

=back

=head2 Parsing

You probably do not want to call these parsing methods yourself: use
the standard constructors (L<new()|Mail::Message::Field::Full/"METHODS">) and it will be done for you.

=over 4

=item $any-E<gt>B<consumeComment>(STRING)

Try to read a comment from the C<STRING>.  When successful, the comment
without encapsulating parenthesis is returned, together with the rest
of the string.

=item $obj-E<gt>B<consumeDotAtom>(STRING)

Returns three elemens: the atom-text, the rest string, and the
concatenated comments.  Both atom and comments can be C<undef>.

=item $any-E<gt>B<consumePhrase>(STRING)

Take the C<STRING>, and try to strip-off a valid phrase.  In the obsolete
phrase syntax, any sequence of words is accepted as phrase (as long as
certain special characters are not used).  RFC2822 is stricter: only
one word or a quoted string is allowed.  As always, the obsolete
syntax is accepted, and the new syntax is produced.

This method returns two elements: the phrase (or C<undef>) followed
by the resulting string.  The phrase will be removed from the optional
quotes.  Be warned that C<""> will return an empty, valid phrase.

» example: 

  my ($phrase, $rest) = $field->consumePhrase( q["hi!" <sales@example.com>] );

=item $obj-E<gt>B<parse>(STRING)

Get the detailed information from the C<STRING>, and store the data found
in the field object.  The accepted input is very field type dependent.
Unstructured fields do no parsing whatsoever.

=item $obj-E<gt>B<produceBody>()

Produce the text for the field, based on the information stored within the
field object.

Usually, you wish the exact same line as was found in the input source
of a message.  But when you have created a field yourself, it should get
formatted.  You may call L<beautify()|Mail::Message::Field::Full/"Access to the content"> on a preformatted field to enforce
a call to this method when the field is needed later.

=back

=head2 Error handling

Extends L<"Error handling" in Mail::Message::Field|Mail::Message::Field/"Error handling">.

=over 4

=item $obj-E<gt>B<AUTOLOAD>()

Inherited, see L<Mail::Reporter/"Error handling">

=item $obj-E<gt>B<notImplemented>()

Inherited, see L<Mail::Reporter/"Error handling">

=back

=head2 Cleanup

Extends L<"Cleanup" in Mail::Message::Field|Mail::Message::Field/"Cleanup">.

=over 4

=item $obj-E<gt>B<DESTROY>()

Inherited, see L<Mail::Reporter/"Cleanup">

=back

=head1 DETAILS

Extends L<"DETAILS" in Mail::Message::Field|Mail::Message::Field/"DETAILS">.

=head1 DIAGNOSTICS

=over 4

=item Error: class $package does not implement method $method.

Fatal error: the specific C<$package> (or one of its superclasses) does not
implement this method where it should. This message means that some other
related classes do implement this method however the class at hand does
not.  Probably you should investigate this and probably inform the author
of the package.
Cast by C<notImplemented()>

=item Warning: field content is not numerical: $content

The numeric value of a field is requested (for instance the C<Lines> or
C<Content-Length> fields should be numerical), however the data contains
weird characters.
Cast by C<toInt()>

=item Error: field name too long (max $count), in '$name'.

It is not specified in the RFCs how long a field name can be, but
at least it should be a few characters shorter than the line wrap.
Cast by C<fold()>

=item Warning: illegal character in charset '$name'.

The field is created with an utf8 string which only contains data from the
specified character set.  However, that character set can never be a valid
name because it contains characters which are not permitted.
Cast by C<encode()>

=item Warning: illegal character in field name $name.

A new field is being created which does contain characters not permitted
by the RFCs.  Using this field in messages may break other e-mail clients
or transfer agents, and therefore mutulate or extinguish your message.
Cast by C<consume()>

=item Warning: illegal character in language '$name'.

The field is created with data which is specified to be in a certain language,
however, the name of the language cannot be valid: it contains characters
which are not permitted by the RFCs.
Cast by C<encode()>

=item Warning: illegal encoding '$name', using 'q'.

The RFCs only permit base64 (C<b > or C<B >) or quoted-printable
(C<q> or C<Q>) encoding.  Other than these four options are illegal.
Cast by C<encode()>

=back

=head1 SEE ALSO

This module is part of Mail-Message version 4.02,
built on January 22, 2026. Website: F<http://perl.overmeer.net/CPAN/>

=head1 LICENSE

For contributors see file ChangeLog.

This software is copyright (c) 2001-2026 by Mark Overmeer.

This is free software; you can redistribute it and/or modify it under
the same terms as the Perl 5 programming language system itself.