1 # Copyright (C) 2014-2020 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # two-step linkification.
5 # intended usage is in the following order:
8 # <escape unsafe chars for HTML>
11 # Maybe this could be done more efficiently...
12 package PublicInbox::Linkify;
15 use Digest::SHA qw/sha1_hex/;
16 use PublicInbox::Hval qw(ascii_html mid_href);
17 use PublicInbox::MID qw($MID_EXTRACT);
20 my $LINK_RE = qr{([\('!])?\b((?:ftps?|https?|nntps?|gopher)://
22 (?:[a-z0-9\-\._~!\$\&\';\(\)\*\+,;=:@/%]*)
23 (?:\?[a-z0-9\-\._~!\$\&\';\(\)\*\+,;=:@/%]+)?
24 (?:\#[a-z0-9\-\._~!\$\&\';\(\)\*\+,;=:@/%\?]+)?
28 sub new { bless {}, $_[0] }
30 # try to distinguish paired punctuation chars from the URL itself
31 # Maybe other languages/formats can be supported here, too...
33 "(" => qr/(\)[\.,;\+]?)\z/, # Markdown (,), Ruby (+) (, for arrays)
34 "'" => qr/('[\.,;\+]?)\z/, # Perl / Ruby
35 "!" => qr/(![\.,;\+]?)\z/, # Perl / Ruby
44 # it's fairly common to end URLs in messages with
45 # '.', ',' or ';' to denote the end of a statement;
46 # assume the intent was to end the statement/sentence
48 if (defined(my $re = $pairs{$beg})) {
49 if ($url =~ s/$re//) {
52 } elsif ($url =~ s/(\))?([\.,;])\z//) {
54 # require ')' to be paired with '('
55 if (defined $1) { # ')'
56 if (index($url, '(') < 0) {
62 } elsif ($url !~ /\(/ && $url =~ s/\)\z//) {
66 $url = ascii_html($url); # for IDN
68 # salt this, as this could be exploited to show
69 # links in the HTML which don't show up in the raw mail.
70 my $key = sha1_hex($url . $SALT);
73 $beg . 'PI-LINK-'. $key . $end;
79 # Added "PI-LINK-" prefix to avoid false-positives on git commits
80 $_[1] =~ s!\bPI-LINK-([a-f0-9]{40})\b!
82 my $url = $_[0]->{$key};
84 "<a\nhref=\"$url\">$url</a>";
86 # false positive or somebody tried to mess with us
93 # single pass linkification of <Message-ID@example.com> within $str
94 # with $pfx being the URL prefix
96 my ($self, $pfx, $str, $raw) = @_;
97 $$str =~ s!$MID_EXTRACT!
99 my $html = ascii_html($mid);
100 my $href = mid_href($mid);
102 # salt this, as this could be exploited to show
103 # links in the HTML which don't show up in the raw mail.
104 my $key = sha1_hex($html . $SALT);
105 my $repl = qq(<<a\nhref="$pfx/$href/">$html</a>>);
106 $repl .= qq{ (<a\nhref="$pfx/$href/raw">raw</a>)} if $raw;
107 $self->{$key} = $repl;
110 $$str = ascii_html($$str);
111 $$str =~ s!\bPI-LINK-([a-f0-9]{40})\b!
113 my $repl = $_[0]->{$key};
117 # false positive or somebody tried to mess with us
123 sub to_html { linkify_2($_[0], ascii_html(linkify_1(@_))) }