1 # Copyright (C) 2014-2015 all contributors <meta@public-inbox.org>
2 # License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
4 # Main web interface for mailing list archives
6 # We focus on the lowest common denominators here:
7 # - targeted at text-only console browsers (w3m, links, etc..)
8 # - Only basic HTML, CSS only for line-wrapping <pre> text content for GUIs
9 # - No JavaScript, graphics or icons allowed.
10 # - Must not rely on static content
11 # - UTF-8 is only for user-content, 7-bit US-ASCII for us
12 package PublicInbox::WWW;
16 use PublicInbox::Config qw(try_cat);
17 use URI::Escape qw(uri_escape_utf8 uri_unescape);
18 use constant SSOMA_URL => 'http://ssoma.public-inbox.org/';
19 use constant PI_URL => 'http://public-inbox.org/';
20 require PublicInbox::Git;
21 our $LISTNAME_RE = qr!\A/([\w\.\-]+)!;
22 our $MID_RE = qr!([^/]+)!;
23 our $END_RE = qr!(f/|T/|t/|t\.mbox(?:\.gz)?|t\.atom|raw|)!;
27 my ($cgi, $method) = @_;
28 $pi_config ||= PublicInbox::Config->new;
29 my $ctx = { cgi => $cgi, pi_config => $pi_config };
30 if ($method !~ /\AGET|HEAD\z/) {
31 return r(405, 'Method Not Allowed');
33 my $path_info = $cgi->path_info;
35 # top-level indices and feeds
36 if ($path_info eq '/') {
38 } elsif ($path_info =~ m!$LISTNAME_RE\z!o) {
39 invalid_list($ctx, $1) || r301($ctx, $1);
40 } elsif ($path_info =~ m!$LISTNAME_RE(?:/|/index\.html)?\z!o) {
41 invalid_list($ctx, $1) || get_index($ctx);
42 } elsif ($path_info =~ m!$LISTNAME_RE/(?:atom\.xml|new\.atom)\z!o) {
43 invalid_list($ctx, $1) || get_atom($ctx);
45 } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/$END_RE\z!o) {
46 msg_page($ctx, $1, $2, $3);
48 # in case people leave off the trailing slash:
49 } elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/(f|T|t)\z!o) {
50 my ($listname, $mid, $suffix) = ($1, $2, $3);
51 $suffix .= $suffix =~ /\A[tT]\z/ ? '/#u' : '/';
52 r301($ctx, $listname, $mid, $suffix);
54 # convenience redirects order matters
55 } elsif ($path_info =~ m!$LISTNAME_RE/([^/]{2,})\z!o) {
59 legacy_redirects($ctx, $path_info);
63 # for CoW-friendliness, MOOOOO!
65 require PublicInbox::Feed;
66 require PublicInbox::View;
67 require PublicInbox::Thread;
73 require PublicInbox::Search;
74 require PublicInbox::SearchView;
75 require PublicInbox::Mbox;
76 require IO::Compress::Gzip;
80 # private functions below
84 if ($ctx && $ctx->{mid}) {
85 require PublicInbox::ExtMsg;
87 return PublicInbox::ExtMsg::ext_msg($ctx);
92 # simple response for errors
93 sub r { [ $_[0], ['Content-Type' => 'text/plain'], [ join(' ', @_, "\n") ] ] }
95 # returns undef if valid, array ref response if invalid
97 my ($ctx, $listname) = @_;
98 my $git_dir = $pi_config->get($listname, "mainrepo");
99 if (defined $git_dir) {
100 $ctx->{git_dir} = $git_dir;
101 $ctx->{git} = PublicInbox::Git->new($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);
114 $ctx->{mid} = $mid = uri_unescape($mid);
115 if ($mid =~ /\A[a-f0-9]{40}\z/) {
116 if ($mid = mid2blob($ctx)) {
117 require Email::Simple;
118 use PublicInbox::MID qw/mid_clean/;
119 $mid = Email::Simple->new($mid);
120 $ctx->{mid} = mid_clean($mid->header('Message-ID'));
126 # /$LISTNAME/new.atom -> Atom feed, includes replies
129 require PublicInbox::Feed;
130 PublicInbox::Feed::generate($ctx);
133 # /$LISTNAME/?r=$GIT_COMMIT -> HTML only
136 require PublicInbox::Feed;
137 my $srch = searcher($ctx);
139 if (defined $ctx->{cgi}->param('q')) {
140 require PublicInbox::SearchView;
141 PublicInbox::SearchView::sres_top_html($ctx);
143 PublicInbox::Feed::generate_html_index($ctx);
147 # just returns a string ref for the blob in the current ctx
150 require PublicInbox::MID;
151 my $path = PublicInbox::MID::mid2path($ctx->{mid});
152 $ctx->{git}->cat_file("HEAD:$path");
155 # /$LISTNAME/$MESSAGE_ID/raw -> raw mbox
158 my $x = mid2blob($ctx) or return r404($ctx);
159 require PublicInbox::Mbox;
160 PublicInbox::Mbox::emit1($ctx, $x);
163 # /$LISTNAME/$MESSAGE_ID/ -> HTML content (short quotes)
166 my $x = mid2blob($ctx) or return r404($ctx);
168 require PublicInbox::View;
169 my $foot = footer($ctx);
171 my $mime = Email::MIME->new($x);
173 [ 200, [ 'Content-Type' => 'text/html; charset=UTF-8' ],
174 [ PublicInbox::View::msg_html($ctx, $mime, 'f/', $foot) ] ];
177 # /$LISTNAME/$MESSAGE_ID/f/ -> HTML content (fullquotes)
180 my $x = mid2blob($ctx) or return r404($ctx);
182 require PublicInbox::View;
183 my $foot = footer($ctx);
185 my $mime = Email::MIME->new($x);
187 [ 200, [ 'Content-Type' => 'text/html; charset=UTF-8' ],
188 [ PublicInbox::View::msg_html($ctx, $mime, undef, $foot)] ];
191 # /$LISTNAME/$MESSAGE_ID/t/
193 my ($ctx, $flat) = @_;
194 my $srch = searcher($ctx) or return need_search($ctx);
195 require PublicInbox::View;
196 my $foot = footer($ctx);
197 $ctx->{flat} = $flat;
198 PublicInbox::View::thread_html($ctx, $foot, $srch);
204 ref($cgi) eq 'CGI' ? $cgi->self_url : $cgi->uri->as_string;
208 my ($ctx, $key) = @_;
209 my $val = $ctx->{$key};
210 (defined $val && $val ne '') or die "BUG: bad ctx, $key unusable\n";
216 return '' unless $ctx;
217 my $git_dir = ctx_get($ctx, 'git_dir');
219 # favor user-supplied footer
220 my $footer = try_cat("$git_dir/public-inbox/footer.html");
221 if (defined $footer) {
223 $ctx->{footer} = $footer;
227 # auto-generate a footer
228 my $listname = ctx_get($ctx, 'listname');
229 my $desc = try_cat("$git_dir/description");
230 $desc = '$GIT_DIR/description missing' unless defined $desc;
233 my $urls = try_cat("$git_dir/cloneurl");
234 my @urls = split(/\r?\n/, $urls || '');
235 my $nurls = scalar @urls;
237 $urls = '($GIT_DIR/cloneurl missing)';
238 } elsif ($nurls == 1) {
239 $urls = "git URL for <a\nhref=\"" . SSOMA_URL .
240 '">ssoma</a>: ' . $urls[0];
242 $urls = "git URLs for <a\nhref=\"" . SSOMA_URL .
243 "\">ssoma</a>:\n" . join("\n", map { "\t$_" } @urls);
246 my $addr = $pi_config->get($listname, 'address');
247 if (ref($addr) eq 'ARRAY') {
248 $addr = $addr->[0]; # first address is primary
251 $addr = "<a\nhref=\"mailto:$addr\">$addr</a>";
253 $ctx->{footer} = join("\n",
255 "A <a\nhref=\"" . PI_URL . '">public-inbox</a>, ' .
256 'anybody may post in plain-text (not HTML):',
262 # search support is optional, returns undef if Xapian is not installed
263 # or not configured for the given GIT_DIR
267 require PublicInbox::Search;
268 $ctx->{srch} = PublicInbox::Search->new($ctx->{git_dir});
275 <html><head><title>Search not available for this
276 public-inbox</title><body><pre>Search is not available for this public-inbox
277 <a href="../">Return to index</a></pre></body></html>
279 [ 501, [ 'Content-Type' => 'text/html; charset=UTF-8' ], [ $msg ] ];
282 # /$LISTNAME/$MESSAGE_ID/t.mbox -> thread as mbox
283 # /$LISTNAME/$MESSAGE_ID/t.mbox.gz -> thread as gzipped mbox
284 # note: I'm not a big fan of other compression formats since they're
285 # significantly more expensive on CPU than gzip and less-widely available,
286 # especially on older systems. Stick to zlib since that's what git uses.
287 sub get_thread_mbox {
288 my ($ctx, $sfx) = @_;
289 my $srch = searcher($ctx) or return need_search($ctx);
290 require PublicInbox::Mbox;
291 PublicInbox::Mbox::thread_mbox($ctx, $srch, $sfx);
295 # /$LISTNAME/$MESSAGE_ID/t.atom -> thread as Atom feed
296 sub get_thread_atom {
298 searcher($ctx) or return need_search($ctx);
299 $ctx->{self_url} = self_url($ctx->{cgi});
300 require PublicInbox::Feed;
301 PublicInbox::Feed::generate_thread_atom($ctx);
304 sub legacy_redirects {
305 my ($ctx, $path_info) = @_;
307 # single-message pages
308 if ($path_info =~ m!$LISTNAME_RE/m/(\S+)/\z!o) {
310 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)/raw\z!o) {
311 r301($ctx, $1, $2, 'raw');
313 } elsif ($path_info =~ m!$LISTNAME_RE/f/(\S+)/\z!o) {
314 r301($ctx, $1, $2, 'f/');
317 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)/\z!o) {
318 r301($ctx, $1, $2, 't/#u');
320 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)/mbox(\.gz)?\z!o) {
321 r301($ctx, $1, $2, "t.mbox$3");
323 # even older legacy redirects
324 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)\.html\z!o) {
327 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)\.html\z!o) {
328 r301($ctx, $1, $2, 't/#u');
330 } elsif ($path_info =~ m!$LISTNAME_RE/f/(\S+)\.html\z!o) {
331 r301($ctx, $1, $2, 'f/');
333 } elsif ($path_info =~ m!$LISTNAME_RE/(?:m|f)/(\S+)\.txt\z!o) {
334 r301($ctx, $1, $2, 'raw');
336 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)(\.mbox(?:\.gz)?)\z!o) {
337 r301($ctx, $1, $2, "t$3");
339 # legacy convenience redirects, order still matters
340 } elsif ($path_info =~ m!$LISTNAME_RE/m/(\S+)\z!o) {
342 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)\z!o) {
343 r301($ctx, $1, $2, 't/#u');
344 } elsif ($path_info =~ m!$LISTNAME_RE/f/(\S+)\z!o) {
345 r301($ctx, $1, $2, 'f/');
347 # some Message-IDs have slashes in them and the HTTP server
348 # may try to be clever and unescape them :<
349 } elsif ($path_info =~ m!$LISTNAME_RE/(\S+/\S+)/$END_RE\z!o) {
350 msg_page($ctx, $1, $2, $3);
352 # in case people leave off the trailing slash:
353 } elsif ($path_info =~ m!$LISTNAME_RE/(\S+/\S+)/(f|T|t)\z!o) {
354 r301($ctx, $1, $2, $3 eq 't' ? 't/#u' : $3);
361 my ($ctx, $listname, $mid, $suffix) = @_;
362 my $cgi = $ctx->{cgi};
365 if (ref($cgi) eq 'CGI') {
366 $url = $cgi->url(-base) . '/';
367 $qs = $cgi->query_string;
368 } else { # Plack::Request
369 $url = $cgi->base->as_string;
370 $qs = $cgi->env->{QUERY_STRING};
373 $url .= $listname . '/';
374 $url .= (uri_escape_utf8($mid) . '/') if (defined $mid);
375 $url .= $suffix if (defined $suffix);
376 $url .= "?$qs" if $qs ne '';
379 [ Location => $url, 'Content-Type' => 'text/plain' ],
380 [ "Redirecting to $url\n" ] ]
384 my ($ctx, $list, $mid, $e) = @_;
385 unless (invalid_list_mid($ctx, $list, $mid)) {
386 '' eq $e and return get_mid_html($ctx);
387 't/' eq $e and return get_thread($ctx);
388 't.atom' eq $e and return get_thread_atom($ctx);
389 't.mbox' eq $e and return get_thread_mbox($ctx);
390 't.mbox.gz' eq $e and return get_thread_mbox($ctx, '.gz');
391 'T/' eq $e and return get_thread($ctx, 1);
392 'raw' eq $e and return get_mid_txt($ctx);
393 'f/' eq $e and return get_full_html($ctx);