1 # Copyright (C) 2014-2016 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/;
18 my $LINK_RE = qr{\b((?:ftps?|https?|nntps?|gopher)://
20 (?:[a-z0-9\-\._~!\$\&\';\(\)\*\+,;=:@/%]*)
21 (?:\?[a-z0-9\-\._~!\$\&\';\(\)\*\+,;=:@/%]+)?
22 (?:\#[a-z0-9\-\._~!\$\&\';\(\)\*\+,;=:@/%\?]+)?
25 sub new { bless {}, shift }
33 # it's fairly common to end URLs in messages with
34 # '.', ',' or ';' to denote the end of a statement;
35 # assume the intent was to end the statement/sentence
37 if ($url =~ s/([\.,;])\z//) {
41 # salt this, as this could be exploited to show
42 # links in the HTML which don't show up in the raw mail.
43 my $key = sha1_hex($url . $SALT);
45 # only escape ampersands, others do not match LINK_RE
48 'PI-LINK-'. $key . $end;
56 # Added "PI-LINK-" prefix to avoid false-positives on git commits
57 $s =~ s!\bPI-LINK-([a-f0-9]{40})\b!
59 my $url = $self->{$key};
61 "<a\nhref=\"$url\">$url</a>";
63 # false positive or somebody tried to mess with us