| Server IP : 14.207.165.8 / Your IP : 216.73.216.102 Web Server : Apache/2.4.18 (Ubuntu) System : Linux 246 4.4.0-210-generic #242-Ubuntu SMP Fri Apr 16 09:57:56 UTC 2021 x86_64 User : root ( 0) PHP Version : 7.0.33-0ubuntu0.16.04.16 Disable Function : exec,passthru,shell_exec,system,proc_open,popen,pcntl_exec MySQL : OFF | cURL : ON | WGET : ON | Perl : ON | Python : ON | Sudo : ON | Pkexec : ON Directory : /proc/thread-self/root/usr/share/doc/libhtml-parser-perl/examples/ |
Upload File : |
#!/usr/bin/perl -w
# This script cleans up an HTML document.
# Specifically it removes deprecated styling and scripting tags.
use strict;
use HTML::Parser ();
# configure these values
my @ignore_attr =
qw(bgcolor background color face style link alink vlink text
onblur onchange onclick ondblclick onfocus onkeydown onkeyup onload
onmousedown onmousemove onmouseout onmouseover onmouseup
onreset onselect onunload
);
my @ignore_tags = qw(font big small b i);
my @ignore_elements = qw(script style);
# make it easier to look up attributes
my %ignore_attr = map { $_ => 1} @ignore_attr;
sub tag
{
my($pos, $text) = @_;
if (@$pos >= 4) {
# kill some attributes
my($k_offset, $k_len, $v_offset, $v_len) = @{$pos}[-4 .. -1];
my $next_attr = $v_offset ? $v_offset + $v_len : $k_offset + $k_len;
my $edited;
while (@$pos >= 4) {
($k_offset, $k_len, $v_offset, $v_len) = splice @$pos, -4;
if ($ignore_attr{lc substr($text, $k_offset, $k_len)}) {
substr($text, $k_offset, $next_attr - $k_offset) = "";
$edited++;
}
$next_attr = $k_offset;
}
# if we killed all attributed, kill any extra whitespace too
$text =~ s/^(<\w+)\s+>$/$1>/ if $edited;
}
print $text;
}
sub decl
{
my $type = shift;
print shift if $type eq "doctype";
}
sub text
{
print shift;
}
HTML::Parser->new(api_version => 3,
start_h => [\&tag, "tokenpos, text"],
process_h => ["", ""],
comment_h => ["", ""],
declaration_h => [\&decl, "tagname, text"],
default_h => [\&text, "text"],
ignore_tags => \@ignore_tags,
ignore_elements => \@ignore_elements,
)
->parse_file(shift) || die "Can't open file: $!\n";