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+)\.html\z!o) {
60 invalid_list_mid(\%ctx, $1, $2) || get_thread(\%ctx);
62 } elsif ($path_info =~ m!$LISTNAME_RE/t/(\S+)/mbox(\.gz)?\z!ox ||
63 $path_info =~ m!$LISTNAME_RE/t/(\S+)\.mbox(\.gz)?\z!o) {
65 invalid_list_mid(\%ctx, $1, $2) ||
66 get_thread_mbox(\%ctx, $sfx);
68 } elsif ($path_info =~ m!$LISTNAME_RE/f/\S+\.txt\z!o) {
69 invalid_list_mid(\%ctx, $1, $2) || redirect_mid_txt(\%ctx);
71 # convenience redirects, order matters
72 } elsif ($path_info =~ m!$LISTNAME_RE/(m|f|t|s)/(\S+)\z!o) {
74 invalid_list_mid(\%ctx, $1, $3) || redirect_mid(\%ctx, $2);
81 # for CoW-friendliness, MOOOOO!
83 require PublicInbox::Feed;
84 require PublicInbox::View;
85 require PublicInbox::Thread;
86 require PublicInbox::GitCatFile;
92 require PublicInbox::Search;
93 require PublicInbox::Mbox;
94 require IO::Compress::Gzip;
98 # private functions below
100 sub r404 { r(404, 'Not Found') }
102 # simple response for errors
103 sub r { [ $_[0], ['Content-Type' => 'text/plain'], [ join(' ', @_, "\n") ] ] }
105 # returns undef if valid, array ref response if invalid
107 my ($ctx, $listname) = @_;
108 my $git_dir = $pi_config->get($listname, "mainrepo");
109 if (defined $git_dir) {
110 $ctx->{git_dir} = $git_dir;
111 $ctx->{listname} = $listname;
117 # returns undef if valid, array ref response if invalid
118 sub invalid_list_mid {
119 my ($ctx, $listname, $mid) = @_;
120 my $ret = invalid_list($ctx, $listname, $mid);
121 $ctx->{mid} = uri_unescape($mid) unless $ret;
125 # /$LISTNAME/atom.xml -> Atom feed, includes replies
128 require PublicInbox::Feed;
129 PublicInbox::Feed::generate($ctx);
132 # /$LISTNAME/?r=$GIT_COMMIT -> HTML only
135 require PublicInbox::Feed;
136 my $srch = searcher($ctx);
138 PublicInbox::Feed::generate_html_index($ctx);
141 # just returns a string ref for the blob in the current ctx
144 require PublicInbox::MID;
145 my $path = PublicInbox::MID::mid2path($ctx->{mid});
146 my @cmd = ('git', "--git-dir=$ctx->{git_dir}",
147 qw(cat-file blob), "HEAD:$path");
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
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)
171 my $x = mid2blob($ctx) or return r404();
173 require PublicInbox::View;
174 my $pfx = msg_pfx($ctx);
175 my $foot = footer($ctx);
177 my $mime = Email::MIME->new($x);
179 [ 200, [ 'Content-Type' => 'text/html; charset=UTF-8' ],
180 [ PublicInbox::View::msg_html($ctx, $mime, $pfx, $foot) ] ];
183 # /$LISTNAME/f/$MESSAGE_ID.html -> HTML content (fullquotes)
186 my $x = mid2blob($ctx) or return r404();
188 require PublicInbox::View;
189 my $foot = footer($ctx);
191 my $mime = Email::MIME->new($x);
193 [ 200, [ 'Content-Type' => 'text/html; charset=UTF-8' ],
194 [ PublicInbox::View::msg_html($ctx, $mime, undef, $foot)] ];
197 # /$LISTNAME/t/$MESSAGE_ID.html
200 my $srch = searcher($ctx) or return need_search($ctx);
201 require PublicInbox::View;
202 my $foot = footer($ctx);
203 PublicInbox::View::thread_html($ctx, $foot, $srch);
208 ref($cgi) eq 'CGI' ? $cgi->self_url : $cgi->uri->as_string;
211 sub redirect_list_index {
213 do_redirect(self_url($cgi) . "/");
217 my ($ctx, $pfx) = @_;
218 my $url = self_url($ctx->{cgi});
220 if (lc($pfx) eq 't') {
221 $anchor = '#u'; # <u id='#u'> is used to highlight in View.pm
223 do_redirect($url . ".html$anchor");
226 # only hit when somebody tries to guess URLs manually:
227 sub redirect_mid_txt {
228 my ($ctx, $pfx) = @_;
229 my $listname = $ctx->{listname};
230 my $url = self_url($ctx->{cgi});
231 $url =~ s!/$listname/f/(\S+\.txt)\z!/$listname/m/$1!;
238 [ Location => $url, 'Content-Type' => 'text/plain' ],
239 [ "Redirecting to $url\n" ]
244 my ($ctx, $key) = @_;
245 my $val = $ctx->{$key};
246 (defined $val && length $val) or die "BUG: bad ctx, $key unusable\n";
253 if (open(my $fh, '<', $path)) {
263 return '' unless $ctx;
264 my $git_dir = ctx_get($ctx, 'git_dir');
266 # favor user-supplied footer
267 my $footer = try_cat("$git_dir/public-inbox/footer.html");
268 if (defined $footer) {
270 $ctx->{footer} = $footer;
274 # auto-generate a footer
275 my $listname = ctx_get($ctx, 'listname');
276 my $desc = try_cat("$git_dir/description");
277 $desc = '$GIT_DIR/description missing' unless defined $desc;
280 my $urls = try_cat("$git_dir/cloneurl");
281 my @urls = split(/\r?\n/, $urls || '');
282 my $nurls = scalar @urls;
284 $urls = '($GIT_DIR/cloneurl missing)';
285 } elsif ($nurls == 1) {
286 $urls = "git URL for <a\nhref=\"" . SSOMA_URL .
287 '">ssoma</a>: ' . $urls[0];
289 $urls = "git URLs for <a\nhref=\"" . SSOMA_URL .
290 "\">ssoma</a>:\n" . join("\n", map { "\t$_" } @urls);
293 my $addr = $pi_config->get($listname, 'address');
294 if (ref($addr) eq 'ARRAY') {
295 $addr = $addr->[0]; # first address is primary
298 $addr = "<a\nhref=\"mailto:$addr\">$addr</a>";
300 $ctx->{footer} = join("\n",
302 "A <a\nhref=\"" . PI_URL . '">public-inbox</a>, ' .
303 'anybody may post in plain-text (not HTML):',
309 # search support is optional, returns undef if Xapian is not installed
310 # or not configured for the given GIT_DIR
314 require PublicInbox::Search;
315 $ctx->{srch} = PublicInbox::Search->new($ctx->{git_dir});
322 <html><head><title>Search not available for this
323 public-inbox</title><body><pre>Search is not available for this public-inbox
324 <a href="../">Return to index</a></pre></body></html>
326 [ 501, [ 'Content-Type' => 'text/html; charset=UTF-8' ], [ $msg ] ];
331 my $href = PublicInbox::Hval::ascii_html(uri_escape_utf8($ctx->{mid}));
335 # /$LISTNAME/t/$MESSAGE_ID/mbox -> thread as mbox
336 # /$LISTNAME/t/$MESSAGE_ID/mbox.gz -> thread as gzipped mbox
337 # note: I'm not a big fan of other compression formats since they're
338 # significantly more expensive on CPU than gzip and less-widely available,
339 # especially on older systems. Stick to zlib since that's what git uses.
340 sub get_thread_mbox {
341 my ($ctx, $sfx) = @_;
342 my $srch = searcher($ctx) or return need_search($ctx);
343 require PublicInbox::Mbox;
344 PublicInbox::Mbox::thread_mbox($ctx, $srch, $sfx);