headers.pm

美国CMU大学开发的操作系统健壮性评测软件

#
# $Id: Headers.pm,v 1.1 1999/07/21 19:12:28 kraven Exp $

package HTTP::Headers;

=head1 NAME

HTTP::Headers - Class encapsulating HTTP Message headers

=head1 SYNOPSIS

 require HTTP::Headers;
 $h = new HTTP::Headers;

=head1 DESCRIPTION

The C<HTTP::Headers> class encapsulates HTTP-style message headers.
The headers consist of attribute-value pairs, which may be repeated,
and which are printed in a particular order.

Instances of this class are usually created as member variables of the
C<HTTP::Request> and C<HTTP::Response> classes, internal to the
library.

The following methods are available:

=over 4

=cut

use strict;
use vars qw($VERSION $TRANSLATE_UNDERSCORE);
$VERSION = sprintf("%d.%02d", q$Revision: 1.1 $ =~ /(\d+)\.(\d+)/);

use Carp ();

# Could not use the AutoLoader becase several of the method names are
# not unique in the first 8 characters.
#use SelfLoader;


# "Good Practice" order of HTTP message headers:
#    - General-Headers
#    - Request-Headers
#    - Response-Headers
#    - Entity-Headers
# (From draft-ietf-http-v11-spec-rev-01, Nov 21, 1997)

my @header_order = qw(
   Cache-Control Connection Date Pragma Transfer-Encoding Upgrade Trailer Via

   Accept Accept-Charset Accept-Encoding Accept-Language
   Authorization Expect From Host
   If-Modified-Since If-Match If-None-Match If-Range If-Unmodified-Since
   Max-Forwards Proxy-Authorization Range Referer TE User-Agent

   Accept-Ranges Age Location Proxy-Authenticate Retry-After Server Vary
   Warning WWW-Authenticate

   Allow Content-Base Content-Encoding Content-Language Content-Length
   Content-Location Content-MD5 Content-Range Content-Type
   ETag Expires Last-Modified
);

# Make alternative representations of @header_order.  This is used
# for sorting and case matching.
my $i = 0;
my %header_order;
my %standard_case;
for (@header_order) {
    my $lc = lc $_;
    $header_order{$lc} = $i++;
    $standard_case{$lc} = $_;
}

$TRANSLATE_UNDERSCORE = 1 unless defined $TRANSLATE_UNDERSCORE;



=item $h = new HTTP::Headers

Constructs a new C<HTTP::Headers> object.  You might pass some initial
attribute-value pairs as parameters to the constructor.  I<E.g.>:

 $h = new HTTP::Headers
     Date         => 'Thu, 03 Feb 1994 00:00:00 GMT',
     Content_Type => 'text/html; version=3.2',
     Content_Base => 'http://www.sn.no/';

=cut

sub new
{
    my($class) = shift;
    my $self = bless {}, $class;
    $self->header(@_); # set up initial headers
    $self;
}


=item $h->header($field [=> $value],...)

Get or set the value of a header.  The header field name is not case
sensitive.  To make the life easier for perl users who wants to avoid
quoting before the => operator, you can use '_' as a synonym for '-'
in header names (this behaviour can be suppressed by setting
$HTTP::Headers::TRANSLATE_UNDERSCORE to a FALSE value).

The header() method accepts multiple ($field => $value) pairs, so you
can update several fields with a single invocation.

The optional $value argument may be a scalar or a reference to a list
of scalars. If the $value argument is undefined or not given, then the
header is not modified.

The old value of the last of the $field values is returned.
Multi-valued fields will be concatenated with "," as separator in
scalar context.

 $header->header(MIME_Version => '1.0',
		 User_Agent   => 'My-Web-Client/0.01');
 $header->header(Accept => "text/html, text/plain, image/*");
 $header->header(Accept => [qw(text/html text/plain image/*)]);
 @accepts = $header->header('Accept');

=cut

sub header
{
    my $self = shift;
    my($field, $val, @old);
    while (($field, $val) = splice(@_, 0, 2)) {
	@old = $self->_header($field, $val);
    }
    return @old if wantarray;
    return $old[0] if @old <= 1;
    join(", ", @old);
}

sub _header
{
    my($self, $field, $val, $push) = @_;
    $field =~ tr/_/-/ if $TRANSLATE_UNDERSCORE;

    # $push is only used interally sub push_header
    Carp::croak('Need a field name') unless length($field);

    my $lc_field = lc $field;
    unless(defined $standard_case{$lc_field}) {
	# generate a %stadard_case entry for this field
	$field =~ s/\b(\w)/\u$1/g;
	$standard_case{$lc_field} = $field;
    }

    my $h = $self->{$lc_field};
    my @old = ref($h) ? @$h : (defined($h) ? ($h) : ());

    if (defined $val) {
	my @new = $push ? @old : ();
	if (!ref($val)) {
	    push(@new, $val);
	} elsif (ref($val) eq 'ARRAY') {
	    push(@new, @$val);
	} else {
	    Carp::croak("Unexpected field value $val");
	}
	$self->{$lc_field} = @new > 1 ? \@new : $new[0];
    }
    @old;
}


# Compare function which makes it easy to sort headers in the
# recommended "Good Practice" order.
sub _header_cmp
{
    # Unknown headers are assign a large value so that they are
    # sorted last.  This also helps avoiding a warning from -w
    # about comparing undefined values.
    $header_order{$a} = 999 unless defined $header_order{$a};
    $header_order{$b} = 999 unless defined $header_order{$b};

    $header_order{$a} <=> $header_order{$b} || $a cmp $b;
}


=item $h->scan(\&doit)

Apply a subroutine to each header in turn.  The callback routine is
called with two parameters; the name of the field and a single value.
If the header has more than one value, then the routine is called once
for each value.  The field name passed to the callback routine has
case as suggested by HTTP Spec, and the headers will be visited in the
recommended "Good Practice" order.

=cut

sub scan
{
    my($self, $sub) = @_;
    my $key;
    foreach $key (sort _header_cmp keys %$self) {
        next if $key =~ /^_/;
	my $vals = $self->{$key};
	if (ref($vals)) {
	    my $val;
	    for $val (@$vals) {
		&$sub($standard_case{$key} || $key, $val);
	    }
	} else {
	    &$sub($standard_case{$key} || $key, $vals);
	}
    }
}


=item $h->as_string([$endl])

Return the header fields as a formatted MIME header.  Since it
internally uses the C<scan()> method to build the string, the result
will use case as suggested by HTTP Spec, and it will follow
recommended "Good Practice" of ordering the header fieds.  Long header
values are not folded. 

The optional parameter specifies the line ending sequence to use.  The
default is C<"\n">.  Embedded "\n" characters in the header will be
substitued with this line ending sequence.

=cut

sub as_string
{
    my($self, $endl) = @_;
    $endl = "\n" unless defined $endl;

    my @result = ();
    $self->scan(sub {
	my($field, $val) = @_;
	if ($val =~ /\n/) {
	    # must handle header values with embedded newlines with care
	    $val =~ s/\s+$//;          # trailing newlines and space must go
	    $val =~ s/\n\n+/\n/g;      # no empty lines
	    $val =~ s/\n([^\040\t])/\n $1/g;  # intial space for continuation
	    $val =~ s/\n/$endl/g;      # substitute with requested line ending
	}
	push(@result, "$field: $val");
    });

    join($endl, @result, '');
}


# The remaining functions should autoloaded only when needed

# A bug in 5.002gamma makes it risky to have POD text inside the
# autoloaded section of the code, so we keep the documentation before
# the __DATA__ token.

=item $h->push_header($field, $val)

Add a new field value of the specified header.  The header field name
is not case sensitive.  The field need not already have a
value. Previous values for the same field are retained.  The argument
may be a scalar or a reference to a list of scalars.

 $header->push_header(Accept => 'image/jpeg');

=item $h->remove_header($field,...)

This function removes the headers with the specified names.