#!/usr/bin/perl
use Encode qw(decode_utf8);
use File::Find;
use HTML::Escape qw(escape_html);
my $search_text = '';
if ($ENV{QUERY_STRING} =~ /^q=([^&]*)/) {
$search_text = decode_utf8($1 // "");
$search_text =~ s/\P{Print}//g; # toss any non-printable utf-8 characters
$search_text = substr($search_text, 0, 64);
}
my $directory = '../_site/log/';
my @results;
my %excluded_files = (
'index.html' => 1, # /log/index.html
);
if ($search_text =~ /\S/) {
find({
wanted => sub {
# Ignore directories and only process index.html
return unless -f $_ && $_ eq 'index.html';
# Calculate the relative path for the URL (prevents leaking server file structure)
my $rel_path = $File::Find::name;
$rel_path =~ s|^\Q$directory\E/?||;
return if $excluded_files{$rel_path};
if (open my $fh, '<', $_) {
my $content = do { local $/; <$fh> };
close $fh;
if ($content =~ /\Q$search_text\E/i) {
# Extract Title
my ($title) = $content =~ /(.*?)<\/title>/is;
$title = $title ? escape_html($title) : $rel_path;
# Extract the first <p> tag content
my ($p_content) = $content =~ /<p[^>]*>(.*?)<\/p>/is;
# Process the snippet
my $snippet = $p_content || "";
$snippet =~ s/<[^>]*>//g; # Remove internal tags
$snippet =~ s/\s+/ /g; # Collapse whitespace
# Escape HTML entities AFTER stripping tags
# but BEFORE sending to the user to prevent XSS.
$snippet = escape_html(substr($snippet, 0, 50));
$snippet .= "..." if length($p_content || "") > 50;
push @results, {
path => $File::Find::name,
title => $title,
snippet => $snippet
};
}
}
},
no_chdir => 0,
follow => 0,
}, $directory);
}
print "Content-Type: text/html\n\n";
my $list;
if ($search_text eq '') {
$list = "<p>Please enter a search term above.</p>";
} elsif (@results == 0) {
$list = "<p>No results found for \"<b>$search_text</b>\".</p>";
} else {
$list = "<ul>";
foreach my $res (@results) {
my $url = $res->{path};
$list .= "<li><a href=\"/$url\">$res->{title}</a><br><small>$res->{snippet}</small></li>";
}
$list .= "</ul>";
}
my $safe_search_text = escape_html($search_text);
my $year = (localtime)[5] + 1900;
print <<"HTML";
<!DOCTYPE html>
<html lang="en-us">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Search
Search
$list