1 # Copyright (C) 2014, Eric Wong <normalperson@yhbt.net> and all contributors
2 # License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
4 # We focus on the lowest common denominators here:
5 # - targeted at text-only console browsers (w3m, links, etc..)
6 # - Only basic HTML, CSS only for line-wrapping <pre> text content for GUIs
7 # - No JavaScript, graphics or icons allowed.
8 # - Must not rely on static content
9 # - UTF-8 is only for user-content, 7-bit US-ASCII for us
10 package PublicInbox::WWW;
14 use PublicInbox::Config;
15 use URI::Escape qw(uri_escape_utf8 uri_unescape);
16 use constant SSOMA_URL => 'http://ssoma.public-inbox.org/';
17 use constant PI_URL => 'http://public-inbox.org/';
18 our $LISTNAME_RE = qr!\A/([\w\.\-]+)!;
21 $pi_config = PublicInbox::Config->new;
25 my ($cgi, $method) = @_;
26 my %ctx = (cgi => $cgi, pi_config => $pi_config);
27 if ($method !~ /\AGET|HEAD\z/) {
28 return r(405, 'Method Not Allowed');
30 my $path_info = $cgi->path_info;
32 # top-level indices and feeds
33 if ($path_info eq '/') {
35 } elsif ($path_info =~ m!$LISTNAME_RE\z!o) {
36 invalid_list(\%ctx, $1) || redirect_list_index($cgi);
37 } elsif ($path_info =~ m!$LISTNAME_RE(?:/|/index\.html)?\z!o) {
38 invalid_list(\%ctx, $1) || get_index(\%ctx);
39 } elsif ($path_info =~ m!$LISTNAME_RE/atom\.xml\z!o) {
40 invalid_list(\%ctx, $1) || get_atom(\%ctx);
42 # single-message pages
43 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)/\z!o) {
44 invalid_list_mid(\%ctx, $1, $2) || get_mid_html(\%ctx);
45 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)/raw\z!o) {
46 invalid_list_mid(\%ctx, $1, $2) || get_mid_txt(\%ctx);
47 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)\.txt\z!o) {
48 invalid_list_mid(\%ctx, $1, $2) || get_mid_txt(\%ctx);
49 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)\.html\z!o) {
50 invalid_list_mid(\%ctx, $1, $2) || get_mid_html(\%ctx);
53 } elsif ($path_info =~ m!$LISTNAME_RE/f/(\S+)/\z!o) {
54 invalid_list_mid(\%ctx, $1, $2) || get_full_html(\%ctx);
55 } elsif ($path_info =~ m!$LISTNAME_RE/f/(\S+)\.html\z!o) {
56 invalid_list_mid(\%ctx, $1, $2) || get_full_html(\%ctx);
59 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)/\z!o) {
60 invalid_list_mid(\%ctx, $1, $2) || get_thread(\%ctx);
62 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)\.html\z!o) {
63 invalid_list_mid(\%ctx, $1, $2) || get_thread(\%ctx);
65 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)/mbox(\.gz)?\z!ox ||
66 $path_info =~ m!$LISTNAME_RE/t/(\S+)\.mbox(\.gz)?\z!o) {
68 invalid_list_mid(\%ctx, $1, $2) ||
69 get_thread_mbox(\%ctx, $sfx);
71 } elsif ($path_info =~ m!$LISTNAME_RE/f/\S+\.txt\z!o) {
72 invalid_list_mid(\%ctx, $1, $2) || redirect_mid_txt(\%ctx);
74 # convenience redirects, order matters
75 } elsif ($path_info =~ m!$LISTNAME_RE/(m|f|t|s)/(\S+)\z!o) {
77 invalid_list_mid(\%ctx, $1, $3) || redirect_mid(\%ctx, $2);
84 # for CoW-friendliness, MOOOOO!
86 require PublicInbox::Feed;
87 require PublicInbox::View;
88 require PublicInbox::Thread;
89 require PublicInbox::GitCatFile;
95 require PublicInbox::Search;
96 require PublicInbox::Mbox;
97 require IO::Compress::Gzip;
101 # private functions below
103 sub r404 { r(404, 'Not Found') }
105 # simple response for errors
106 sub r { [ $_[0], ['Content-Type' => 'text/plain'], [ join(' ', @_, "\n") ] ] }
108 # returns undef if valid, array ref response if invalid
110 my ($ctx, $listname) = @_;
111 my $git_dir = $pi_config->get($listname, "mainrepo");
112 if (defined $git_dir) {
113 $ctx->{git_dir} = $git_dir;
114 $ctx->{listname} = $listname;
120 # returns undef if valid, array ref response if invalid
121 sub invalid_list_mid {
122 my ($ctx, $listname, $mid) = @_;
123 my $ret = invalid_list($ctx, $listname, $mid);
124 $ctx->{mid} = uri_unescape($mid) unless $ret;
128 # /$LISTNAME/atom.xml -> Atom feed, includes replies
131 require PublicInbox::Feed;
132 PublicInbox::Feed::generate($ctx);
135 # /$LISTNAME/?r=$GIT_COMMIT -> HTML only
138 require PublicInbox::Feed;
139 my $srch = searcher($ctx);
141 PublicInbox::Feed::generate_html_index($ctx);
144 # just returns a string ref for the blob in the current ctx
147 require PublicInbox::MID;
148 my $path = PublicInbox::MID::mid2path($ctx->{mid});
149 my @cmd = ('git', "--git-dir=$ctx->{git_dir}",
150 qw(cat-file blob), "HEAD:$path");
151 my $pid = open my $fh, '-|';
152 defined $pid or die "fork failed: $!\n";
154 open STDERR, '>', '/dev/null'; # ignore errors
155 exec @cmd or die "exec failed: $!\n";
157 my $blob = eval { local $/; <$fh> };
159 $? == 0 ? \$blob : undef;
163 # /$LISTNAME/m/$MESSAGE_ID.txt -> raw mbox
166 my $x = mid2blob($ctx) or return r404();
167 require PublicInbox::Mbox;
168 PublicInbox::Mbox::emit1($x);
171 # /$LISTNAME/m/$MESSAGE_ID.html -> HTML content (short quotes)
174 my $x = mid2blob($ctx) or return r404();
176 require PublicInbox::View;
177 my $pfx = msg_pfx($ctx);
178 my $foot = footer($ctx);
180 my $mime = Email::MIME->new($x);
182 [ 200, [ 'Content-Type' => 'text/html; charset=UTF-8' ],
183 [ PublicInbox::View::msg_html($ctx, $mime, $pfx, $foot) ] ];
186 # /$LISTNAME/f/$MESSAGE_ID.html -> HTML content (fullquotes)
189 my $x = mid2blob($ctx) or return r404();
191 require PublicInbox::View;
192 my $foot = footer($ctx);
194 my $mime = Email::MIME->new($x);
196 [ 200, [ 'Content-Type' => 'text/html; charset=UTF-8' ],
197 [ PublicInbox::View::msg_html($ctx, $mime, undef, $foot)] ];
200 # /$LISTNAME/t/$MESSAGE_ID.html
203 my $srch = searcher($ctx) or return need_search($ctx);
204 require PublicInbox::View;
205 my $foot = footer($ctx);
206 PublicInbox::View::thread_html($ctx, $foot, $srch);
211 ref($cgi) eq 'CGI' ? $cgi->self_url : $cgi->uri->as_string;
214 sub redirect_list_index {
216 do_redirect(self_url($cgi) . "/");
220 my ($ctx, $pfx) = @_;
221 my $url = self_url($ctx->{cgi});
223 if (lc($pfx) eq 't') {
224 $anchor = '#u'; # <u id='#u'> is used to highlight in View.pm
226 do_redirect($url . "/$anchor");
229 # only hit when somebody tries to guess URLs manually:
230 sub redirect_mid_txt {
231 my ($ctx, $pfx) = @_;
232 my $listname = $ctx->{listname};
233 my $url = self_url($ctx->{cgi});
234 $url =~ s!/$listname/f/(\S+\.txt)\z!/$listname/m/$1!;
241 [ Location => $url, 'Content-Type' => 'text/plain' ],
242 [ "Redirecting to $url\n" ]
247 my ($ctx, $key) = @_;
248 my $val = $ctx->{$key};
249 (defined $val && length $val) or die "BUG: bad ctx, $key unusable\n";
256 if (open(my $fh, '<', $path)) {
266 return '' unless $ctx;
267 my $git_dir = ctx_get($ctx, 'git_dir');
269 # favor user-supplied footer
270 my $footer = try_cat("$git_dir/public-inbox/footer.html");
271 if (defined $footer) {
273 $ctx->{footer} = $footer;
277 # auto-generate a footer
278 my $listname = ctx_get($ctx, 'listname');
279 my $desc = try_cat("$git_dir/description");
280 $desc = '$GIT_DIR/description missing' unless defined $desc;
283 my $urls = try_cat("$git_dir/cloneurl");
284 my @urls = split(/\r?\n/, $urls || '');
285 my $nurls = scalar @urls;
287 $urls = '($GIT_DIR/cloneurl missing)';
288 } elsif ($nurls == 1) {
289 $urls = "git URL for <a\nhref=\"" . SSOMA_URL .
290 '">ssoma</a>: ' . $urls[0];
292 $urls = "git URLs for <a\nhref=\"" . SSOMA_URL .
293 "\">ssoma</a>:\n" . join("\n", map { "\t$_" } @urls);
296 my $addr = $pi_config->get($listname, 'address');
297 if (ref($addr) eq 'ARRAY') {
298 $addr = $addr->[0]; # first address is primary
301 $addr = "<a\nhref=\"mailto:$addr\">$addr</a>";
303 $ctx->{footer} = join("\n",
305 "A <a\nhref=\"" . PI_URL . '">public-inbox</a>, ' .
306 'anybody may post in plain-text (not HTML):',
312 # search support is optional, returns undef if Xapian is not installed
313 # or not configured for the given GIT_DIR
317 require PublicInbox::Search;
318 $ctx->{srch} = PublicInbox::Search->new($ctx->{git_dir});
325 <html><head><title>Search not available for this
326 public-inbox</title><body><pre>Search is not available for this public-inbox
327 <a href="../">Return to index</a></pre></body></html>
329 [ 501, [ 'Content-Type' => 'text/html; charset=UTF-8' ], [ $msg ] ];
334 my $href = PublicInbox::Hval::ascii_html(uri_escape_utf8($ctx->{mid}));
338 # /$LISTNAME/t/$MESSAGE_ID/mbox -> thread as mbox
339 # /$LISTNAME/t/$MESSAGE_ID/mbox.gz -> thread as gzipped mbox
340 # note: I'm not a big fan of other compression formats since they're
341 # significantly more expensive on CPU than gzip and less-widely available,
342 # especially on older systems. Stick to zlib since that's what git uses.
343 sub get_thread_mbox {
344 my ($ctx, $sfx) = @_;
345 my $srch = searcher($ctx) or return need_search($ctx);
346 require PublicInbox::Mbox;
347 PublicInbox::Mbox::thread_mbox($ctx, $srch, $sfx);