#!/usr/local/bin/perl -w
#
# HTML::Defaultify-- Pre-fill default values into an existing HTML form.
#
# The main purpose of this module is the defaultify() routine, which
# takes a block of HTML and a hash of default values, and returns that HTML
# with all form fields set based on those default values. Default values
# (hash elements) may each be given in any of three forms: as a single
# scalar, as a list in "\0"-delimited form, or as a reference to an actual
# list. If the HTML contains more than one form, you can name which form to
# defaultify. Return values are the defaultified block of HTML and a hash of
# all unused default values (which may be useful as input to hidden_vars()).
# Multiple form fields with the same name are handled correctly. Besides
# the main defaultify() routine, this module includes several related
# routines which the programmer may find useful.
#
# This package prefers to have the HTML::Entities module available, but
# can improvise without it.
#
# Copyright (c) 1996, 1997, 2002 James Marshall (james@jmarshall.com).
# Adapted from the toolbox htmlutil.pl, which is (c) 1996, 1997 by same.
# All rights reserved.
#
# This program is free software; you can redistribute it and/or modify it
# under the same terms as Perl itself.
#
# Exported by default:
# $new_HTML= &defaultify($HTML, \%defaults [, $form_name]) ;
# ($new_HTML, $unused_defs)= &defaultify($HTML, \%defaults [, $form_name]) ;
#
# $my_subset_ref= &subhash(\%hash, @keys_to_include) ;
#
# Export is allowed:
# $hidden_vars= &hidden_vars($unused_defaults_ref) ;
# $hidden_vars= &hidden_vars(%unused_defaults) ;
# $hidden_tag= &hidden_var($name, $value) ;
#
# ($tag_name, $attr_ref)= &parse_tag($tag) ;
# $new_tag= &build_tag($tag_name, $attr_ref) ;
# $new_tag= &build_tag($tag_name, %attr) ;
#
#
# For better documentation, see "perldoc HTML::Defaultify" (or
# "perldoc -F this_file_name.pm").
#
# For the latest, see http://www.jmarshall.com/tools/defaultify/ .
#
#---- package-definition-related stuff -------------------------------------
package HTML::Defaultify ;
use strict ;
use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS) ;
require Exporter ;
@ISA= qw(Exporter) ;
$VERSION= '1.01' ;
@EXPORT= qw( defaultify subhash ) ;
@EXPORT_OK= qw( hidden_vars hidden_var parse_tag build_tag ) ;
%EXPORT_TAGS= (
parse => [qw( parse_tag build_tag )],
all => [@EXPORT, @EXPORT_OK],
) ;
#---- actual package code below --------------------------------------------
use Carp ;
use vars qw($HAS_HTML_ENTITIES) ;
# Load HTML::Entities if available; set $HAS_HTML_ENTITIES accordingly.
eval 'use HTML::Entities' ;
$HAS_HTML_ENTITIES= ($@ eq '') ;
# defaultify()-- takes a chunk of HTML that includes form input fields,
# and sets defaults according to the hash sent.
# Returns defaultified HTML block, and a reference to a hash of all defaults
# that were not used (possibly for use with hidden_vars()).
# In scalar context, only returns defaultified HTML block.
#
# ($new_HTML, $unused_defs)= &defaultify($HTML, $defaults [, $form_name]) ;
# $new_HTML= &defaultify($HTML, $defaults [, $form_name]) ;
#
# $defaults is a reference to a hash of default values. Each default (hash
# element) may be a scalar, a list in the form of a "\0"-delimited scalar,
# or a reference to a real list.
# As a special case, if $defaults is undefined, this routine clears all default
# settings from $HTML, even if they were set with tag attributes, etc.
# As another special case, if $defaults is a CGI object (from CGI.pm), this
# routine uses its existing parameters as the default set, by calling its
# Vars() method.
# If you have an existing hash instead of a reference, use e.g. \%my_hash .
# If $form_name is given, then only the form(s) with that name in $HTML will
# be defaultified. Otherwise, all of $HTML will be defaultified.
# Tags inside comments or blocks,
# and blocks from $HTML, to avoid matching tags
# inside them. Replace these extractions with markers, so they can be
# restored after defaultification is complete. Somewhat hacky approach,
# but works.
# Extractions are stored in @extracts, and the markers consist of: a random
# string not otherwise in $HTML, plus each extraction's location in
# @extracts, plus "\0".
# All four kinds of extractions (two comment formats, scripts, and styles)
# are handled simultaneously. This correctly handles cases of when
# "
# or tags. This is most likely what the HTML author expects
# anyway, though it violates the HTML spec. Worse, browsers vary on
# whether they'll end a "
# inside the script code. Balancing all this, for here it's a reasonable
# policy to end those blocks on "" and "".
# There's a potential problem with the marker: Even if it's not in
# $HTML, certain sequences could cause problems. Consider a marker of
# "xy1xy", and a comment preceded by "xy1". After the comment->marker
# replacement, the string is "xy1xy1xy" and will match too early. But
# since we know \d+\0 will always follow the marker, then excluding
# digits and \0 from the marker will prevent a wrong match like this.
# I'm pretty sure this solves it, but please tell me if you think of
# any combinations that could break this.
# Generate a random 5-character string. Exclude digits, \0, and
# what the hell, "<" and ">".
# srand is automatically called in Perl 5.004 and later.
do {
$marker= pack("C5", map {rand(193)+63} 1..5) ; # start after ">"
} while $HTML=~ /\Q$marker/ ;
# Extract comments,