2 # Copyright (C) 2014, Eric Wong <normalperson@yhbt.net> and all contributors
3 # License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
5 # We focus on the lowest common denominators here:
6 # - targeted at text-only console browsers (lynx, w3m, etc..)
7 # - Only basic HTML, CSS only for line-wrapping <pre> text content for GUIs
8 # - No JavaScript, graphics or icons allowed.
9 # - Must not rely on static content
10 # - UTF-8 is only for user-content, 7-bit US-ASCII for us
15 use CGI qw(:cgi -nosticky); # PSGI/FastCGI/mod_perl compat
16 use Encode qw(find_encoding);
17 use PublicInbox::Config;
18 use URI::Escape qw(uri_escape uri_unescape);
19 my $enc_utf8 = find_encoding('UTF-8');
20 our $LISTNAME_RE = qr!\A/([\w\.\-]+)!;
23 $pi_config = PublicInbox::Config->new;
24 # TODO: detect and reload config as needed
32 my ($status, $headers, $body) = @$ret;
33 set_binmode($headers);
34 if (@ARGV && $ARGV[0] eq 'static') {
37 print "Status: $status\r\n";
38 while (my ($k, $v) = each %$headers) {
46 # private functions below
49 # some servers (Ruby webrick) include scheme://host[:port] here,
50 # which confuses CGI.pm when generating self_url.
51 # RFC 3875 does not mention REQUEST_URI at all,
52 # so nuke it since CGI.pm functions without it.
53 delete $ENV{REQUEST_URI};
57 if ($cgi->request_method !~ /\AGET|HEAD\z/) {
58 return r("405 Method Not Allowed");
60 my $path_info = $enc_utf8->decode($cgi->path_info);
62 # top-level indices and feeds
63 if ($path_info eq "/") {
65 } elsif ($path_info =~ m!$LISTNAME_RE\z!o) {
66 invalid_list(\%ctx, $1) || redirect_list_index(\%ctx, $cgi);
67 } elsif ($path_info =~ m!$LISTNAME_RE(?:/|/index\.html)?\z!o) {
68 invalid_list(\%ctx, $1) || get_index(\%ctx, $cgi, 0);
69 } elsif ($path_info =~ m!$LISTNAME_RE/atom\.xml\z!o) {
70 invalid_list(\%ctx, $1) || get_atom(\%ctx, $cgi, 0);
72 # single-message pages
73 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)\.txt\z!o) {
74 invalid_list_mid(\%ctx, $1, $2) || get_mid_txt(\%ctx, $cgi);
75 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)\.html\z!o) {
76 invalid_list_mid(\%ctx, $1, $2) || get_mid_html(\%ctx, $cgi);
79 } elsif ($path_info =~ m!$LISTNAME_RE/f/(\S+)\.html\z!o) {
80 invalid_list_mid(\%ctx, $1, $2) || get_full_html(\%ctx, $cgi);
82 # convenience redirects, order matters
83 } elsif ($path_info =~ m!$LISTNAME_RE/(?:m|f)/(\S+)\z!o) {
84 invalid_list_mid(\%ctx, $1, $2) || redirect_mid(\%ctx, $cgi);
91 sub r404 { r("404 Not Found") }
93 # simple response for errors
94 sub r { [ $_[0], { 'Content-Type' => 'text/plain' }, $_[0]."\n" ] }
96 # returns undef if valid, array ref response if invalid
98 my ($ctx, $listname) = @_;
99 my $git_dir = $pi_config->get($listname, "mainrepo");
100 if (defined $git_dir) {
101 $ctx->{git_dir} = $git_dir;
102 $ctx->{listname} = $listname;
108 # returns undef if valid, array ref response if invalid
109 sub invalid_list_mid {
110 my ($ctx, $listname, $mid) = @_;
111 my $ret = invalid_list($ctx, $listname, $mid) and return $ret;
112 $ctx->{mid} = uri_unescape($mid);
116 # /$LISTNAME/atom.xml -> Atom feed, includes replies
118 my ($ctx, $cgi, $top) = @_;
119 require PublicInbox::Feed;
120 [ '200 OK', { 'Content-Type' => 'application/xml' },
121 PublicInbox::Feed->generate({
122 git_dir => $ctx->{git_dir},
123 listname => $ctx->{listname},
124 pi_config => $pi_config,
131 # /$LISTNAME/?r=$GIT_COMMIT -> HTML only
133 my ($ctx, $cgi, $top) = @_;
134 require PublicInbox::Feed;
135 [ '200 OK', { 'Content-Type' => 'text/html' },
136 PublicInbox::Feed->generate_html_index({
137 git_dir => $ctx->{git_dir},
138 listname => $ctx->{listname},
139 pi_config => $pi_config,
146 # just returns a string ref for the blob in the current ctx
149 local $ENV{GIT_DIR} = $ctx->{git_dir};
151 my $hex = Digest::SHA::sha1_hex($ctx->{mid});
152 $hex =~ /\A([a-f0-9]{2})([a-f0-9]{38})\z/i or
153 die "BUG: not a SHA-1 hex: $hex";
154 my $blob = `git cat-file blob HEAD:$1/$2 2>/dev/null`;
155 $? == 0 ? \$blob : undef;
158 # /$LISTNAME/m/$MESSAGE_ID.txt -> raw original
160 my ($ctx, $cgi) = @_;
161 my $x = mid2blob($ctx);
162 $x ? [ "200 OK", {'Content-Type' => 'text/plain'}, $$x ] : r404();
165 # /$LISTNAME/m/$MESSAGE_ID.html -> HTML content (short quotes)
167 my ($ctx, $cgi) = @_;
168 my $x = mid2blob($ctx);
169 return r404() unless $x;
171 require PublicInbox::View;
172 my $mid_href = PublicInbox::Hval::ascii_html(uri_escape($ctx->{mid}));
173 my $pfx = "../f/$mid_href.html";
175 [ "200 OK", {'Content-Type' => 'text/html'},
176 PublicInbox::View->as_html(Email::MIME->new($$x), $pfx)];
179 # /$LISTNAME/f/$MESSAGE_ID.html -> HTML content (fullquotes)
181 my ($ctx, $cgi) = @_;
182 my $x = mid2blob($ctx);
183 return r404() unless $x;
184 require PublicInbox::View;
186 [ "200 OK", {'Content-Type' => 'text/html'},
187 PublicInbox::View->as_html(Email::MIME->new($$x))];
190 sub redirect_list_index {
191 my ($ctx, $cgi) = @_;
192 do_redirect($cgi->self_url . "/");
196 my ($ctx, $cgi) = @_;
197 my $url = $cgi->self_url;
199 do_redirect($url . '.html');
204 [ '301 Moved Permanently',
205 { Location => $url, 'Content-Type' => 'text/plain' },
206 "Redirecting to $url\n"
210 # only used for CGI and static file generation modes
213 if ($headers->{'Content-Type'} eq 'text/plain') {
214 # no way to validate raw messages, mixed encoding is possible.
216 } else { # strict encoding for HTML and XML
217 binmode STDOUT, ':encoding(us-ascii)';