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, $method) = @_;
27 if ($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(\%ctx, $cgi);
37 } elsif ($path_info =~ m!$LISTNAME_RE(?:/|/index\.html)?\z!o) {
38 invalid_list(\%ctx, $1) || get_index(\%ctx, $cgi);
39 } elsif ($path_info =~ m!$LISTNAME_RE/atom\.xml\z!o) {
40 invalid_list(\%ctx, $1) || get_atom(\%ctx, $cgi);
42 # single-message pages
43 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)\.txt\z!o) {
44 invalid_list_mid(\%ctx, $1, $2) || get_mid_txt(\%ctx, $cgi);
45 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)\.html\z!o) {
46 invalid_list_mid(\%ctx, $1, $2) || get_mid_html(\%ctx, $cgi);
49 } elsif ($path_info =~ m!$LISTNAME_RE/f/(\S+)\.html\z!o) {
50 invalid_list_mid(\%ctx, $1, $2) || get_full_html(\%ctx, $cgi);
53 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)\.html\z!o) {
54 invalid_list_mid(\%ctx, $1, $2) || get_thread(\%ctx, $cgi);
56 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)\.mbox(\.gz)?\z!o) {
58 invalid_list_mid(\%ctx, $1, $2) ||
59 get_thread_mbox(\%ctx, $cgi, $sfx);
61 } elsif ($path_info =~ m!$LISTNAME_RE/f/\S+\.txt\z!o) {
62 invalid_list_mid(\%ctx, $1, $2) ||
63 redirect_mid_txt(\%ctx, $cgi);
65 # convenience redirects, order matters
66 } elsif ($path_info =~ m!$LISTNAME_RE/(m|f|t|s)/(\S+)\z!o) {
68 invalid_list_mid(\%ctx, $1, $3) ||
69 redirect_mid(\%ctx, $cgi, $2);
76 # for CoW-friendliness, MOOOOO!
78 require PublicInbox::Feed;
79 require PublicInbox::View;
80 require PublicInbox::Thread;
81 require PublicInbox::GitCatFile;
87 require PublicInbox::Search;
88 require PublicInbox::Mbox;
89 require IO::Compress::Gzip;
93 # private functions below
95 sub r404 { r(404, 'Not Found') }
97 # simple response for errors
98 sub r { [ $_[0], ['Content-Type' => 'text/plain'], [ join(' ', @_, "\n") ] ] }
100 # returns undef if valid, array ref response if invalid
102 my ($ctx, $listname) = @_;
103 my $git_dir = $pi_config->get($listname, "mainrepo");
104 if (defined $git_dir) {
105 $ctx->{git_dir} = $git_dir;
106 $ctx->{listname} = $listname;
112 # returns undef if valid, array ref response if invalid
113 sub invalid_list_mid {
114 my ($ctx, $listname, $mid) = @_;
115 my $ret = invalid_list($ctx, $listname, $mid);
116 $ctx->{mid} = uri_unescape($mid) unless $ret;
120 # /$LISTNAME/atom.xml -> Atom feed, includes replies
122 my ($ctx, $cgi) = @_;
123 $ctx->{pi_config} = $pi_config;
125 require PublicInbox::Feed;
126 PublicInbox::Feed::generate($ctx);
129 # /$LISTNAME/?r=$GIT_COMMIT -> HTML only
131 my ($ctx, $cgi) = @_;
132 require PublicInbox::Feed;
133 my $srch = searcher($ctx);
134 $ctx->{pi_config} = $pi_config;
137 PublicInbox::Feed::generate_html_index($ctx);
140 # just returns a string ref for the blob in the current ctx
143 require PublicInbox::MID;
144 my $path = PublicInbox::MID::mid2path($ctx->{mid});
145 my @cmd = ('git', "--git-dir=$ctx->{git_dir}",
146 qw(cat-file blob), "HEAD:$path");
147 my $cmd = join(' ', @cmd);
148 my $pid = open my $fh, '-|';
149 defined $pid or die "fork failed: $!\n";
151 open STDERR, '>', '/dev/null'; # ignore errors
152 exec @cmd or die "exec failed: $!\n";
154 my $blob = eval { local $/; <$fh> };
156 $? == 0 ? \$blob : undef;
160 # /$LISTNAME/m/$MESSAGE_ID.txt -> raw mbox
162 my ($ctx, $cgi) = @_;
163 my $x = mid2blob($ctx) or return r404();
164 require PublicInbox::Mbox;
165 PublicInbox::Mbox::emit1($x);
168 # /$LISTNAME/m/$MESSAGE_ID.html -> HTML content (short quotes)
170 my ($ctx, $cgi) = @_;
171 my $x = mid2blob($ctx);
172 return r404() unless $x;
174 require PublicInbox::View;
175 my $pfx = msg_pfx($ctx);
176 my $foot = footer($ctx);
178 my $mime = Email::MIME->new($x);
179 my $srch = searcher($ctx);
180 [ 200, [ 'Content-Type' => 'text/html; charset=UTF-8' ],
181 [ PublicInbox::View->msg_html($mime, $pfx, $foot, $srch) ] ];
184 # /$LISTNAME/f/$MESSAGE_ID.html -> HTML content (fullquotes)
186 my ($ctx, $cgi) = @_;
187 my $x = mid2blob($ctx);
188 return r404() unless $x;
189 require PublicInbox::View;
190 my $foot = footer($ctx);
192 my $mime = Email::MIME->new($x);
193 my $srch = searcher($ctx);
194 [ 200, [ 'Content-Type' => 'text/html; charset=UTF-8' ],
195 [ PublicInbox::View->msg_html($mime, undef, $foot, $srch)] ];
198 # /$LISTNAME/t/$MESSAGE_ID.html
200 my ($ctx, $cgi) = @_;
201 my $srch = searcher($ctx) or return need_search($ctx);
202 require PublicInbox::View;
203 my $foot = footer($ctx);
204 PublicInbox::View::thread_html($ctx, $foot, $srch);
209 ref($cgi) eq 'CGI' ? $cgi->self_url : $cgi->uri->as_string;
212 sub redirect_list_index {
213 my ($ctx, $cgi) = @_;
214 do_redirect(self_url($cgi) . "/");
218 my ($ctx, $cgi, $pfx) = @_;
219 my $url = self_url($cgi);
221 if (lc($pfx) eq 't') {
222 $anchor = '#u'; # <u id='#u'> is used to highlight in View.pm
224 do_redirect($url . ".html$anchor");
227 # only hit when somebody tries to guess URLs manually:
228 sub redirect_mid_txt {
229 my ($ctx, $cgi, $pfx) = @_;
230 my $listname = $ctx->{listname};
231 my $url = self_url($cgi);
232 $url =~ s!/$listname/f/(\S+\.txt)\z!/$listname/m/$1!;
239 [ Location => $url, 'Content-Type' => 'text/plain' ],
240 [ "Redirecting to $url\n" ]
245 my ($ctx, $key) = @_;
246 my $val = $ctx->{$key};
247 (defined $val && length $val) or die "BUG: bad ctx, $key unusable\n";
254 if (open(my $fh, '<', $path)) {
264 return '' unless $ctx;
265 my $git_dir = ctx_get($ctx, 'git_dir');
267 # favor user-supplied footer
268 my $footer = try_cat("$git_dir/public-inbox/footer.html");
269 if (defined $footer) {
271 $ctx->{footer} = $footer;
275 # auto-generate a footer
276 my $listname = ctx_get($ctx, 'listname');
277 my $desc = try_cat("$git_dir/description");
278 $desc = '$GIT_DIR/description missing' unless defined $desc;
281 my $urls = try_cat("$git_dir/cloneurl");
282 my @urls = split(/\r?\n/, $urls || '');
283 my $nurls = scalar @urls;
285 $urls = '($GIT_DIR/cloneurl missing)';
286 } elsif ($nurls == 1) {
287 $urls = "git URL for <a\nhref=\"" . SSOMA_URL .
288 '">ssoma</a>: ' . $urls[0];
290 $urls = "git URLs for <a\nhref=\"" . SSOMA_URL .
291 "\">ssoma</a>:\n" . join("\n", map { "\t$_" } @urls);
294 my $addr = $pi_config->get($listname, 'address');
295 if (ref($addr) eq 'ARRAY') {
296 $addr = $addr->[0]; # first address is primary
299 $addr = "<a\nhref=\"mailto:$addr\">$addr</a>";
301 $ctx->{footer} = join("\n",
303 "A <a\nhref=\"" . PI_URL . '">public-inbox</a>, ' .
304 'anybody may post in plain-text (not HTML):',
310 # search support is optional, returns undef if Xapian is not installed
311 # or not configured for the given GIT_DIR
315 require PublicInbox::Search;
316 $ctx->{srch} = PublicInbox::Search->new($ctx->{git_dir});
323 <html><head><title>Search not available for this
324 public-inbox</title><body><pre>Search is not available for this public-inbox
325 <a href="../">Return to index</a></pre></body></html>
327 [ 501, [ 'Content-Type' => 'text/html; charset=UTF-8' ], [ $msg ] ];
332 my $href = PublicInbox::Hval::ascii_html(uri_escape_utf8($ctx->{mid}));
336 # /$LISTNAME/t/$MESSAGE_ID.mbox -> thread as mbox
337 # /$LISTNAME/t/$MESSAGE_ID.mbox.gz -> thread as gzipped mbox
338 # note: I'm not a big fan of other compression formats since they're
339 # significantly more expensive on CPU than gzip and less-widely available,
340 # especially on older systems. Stick to zlib since that's what git uses.
341 sub get_thread_mbox {
342 my ($ctx, $cgi, $sfx) = @_;
343 my $srch = searcher($ctx) or return need_search($ctx);
344 require PublicInbox::Mbox;
345 PublicInbox::Mbox::thread_mbox($ctx, $srch, $sfx);