1 # Copyright (C) all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # used for displaying help texts and other non-mail content
5 package PublicInbox::WwwText;
8 use PublicInbox::Linkify;
9 use PublicInbox::WwwStream;
10 use PublicInbox::Hval qw(ascii_html prurl);
11 use HTTP::Date qw(time2str);
12 use URI::Escape qw(uri_escape_utf8);
13 use PublicInbox::GzipFilter qw(gzf_maybe);
14 our $QP_URL = 'https://xapian.org/docs/queryparser.html';
15 our $WIKI_URL = 'https://en.wikipedia.org/wiki';
17 require PublicInbox::HlMod;
18 PublicInbox::HlMod->new
21 # /$INBOX/_/text/$KEY/ # KEY may contain slashes
22 # For now, "help" is the only supported $KEY
27 $key //= 'help'; # this 302s to _/text/help/
29 # get the raw text the same way we get mboxrds
30 my $raw = ($key =~ s!/raw\z!!);
31 my $have_tslash = ($key =~ s!/\z!!) if !$raw;
34 my $hdr = [ 'Content-Type', 'text/plain', 'Content-Length', undef ];
35 if (!_default_text($ctx, $key, $hdr, \$txt)) {
37 $txt = "404 Not Found ($key)\n";
39 my $env = $ctx->{env};
41 $txt = gzf_maybe($hdr, $env)->zflush($txt) if $code == 200;
42 $hdr->[3] = length($txt);
43 return [ $code, $hdr, [ $txt ] ]
46 # enforce trailing slash for "wget -r" compatibility
47 if (!$have_tslash && $code == 200) {
48 my $url = $ctx->{ibx}->base_url($env);
49 $url .= "_/text/$key/";
51 return [ 302, [ 'Content-Type', 'text/plain',
53 [ "Redirecting to $url\n" ] ];
56 # Follow git commit message conventions,
57 # first line is the Subject/title
58 my ($title) = ($txt =~ /\A([^\n]*)/s);
59 $ctx->{-title_html} = ascii_html($title);
60 my $nslash = ($key =~ tr!/!/!);
61 $ctx->{-upfx} = '../../../' . ('../' x $nslash);
62 my $l = PublicInbox::Linkify->new;
65 $hl->do_hl_text(\$txt);
67 $txt = ascii_html($txt);
69 $txt = '<pre>' . $l->linkify_2($txt) . '</pre>';
70 $txt =~ s!\bPOP3\b!<a\nid=pop3>POP3</a>!;
71 $txt =~ s!\b(Newsgroups?)\b!<a\nid=nntp>$1</a>!;
72 $txt =~ s!\bIMAP\b!<a\nid=imap>IMAP</a>!;
73 PublicInbox::WwwStream::html_oneshot($ctx, $code, \$txt);
76 sub _srch_prefix ($$) {
80 my $help = $ibx->isrch->help;
82 for ($i = 0; $i < @$help; $i += 2) {
83 my $pfx = $help->[$i];
85 $pad = $n if $n > $pad;
87 $htxt .= $help->[$i + 1];
91 my $padding = ' ' x ($pad + 4);
92 $htxt =~ s/^/$padding/gms;
93 $htxt =~ s/^$padding(\S+)\0/" $1".(' ' x ($pad - length($1)))/egms;
94 $htxt =~ s/\f\n/\n/gs;
99 sub _colors_help ($$) {
100 my ($ctx, $txt) = @_;
101 my $ibx = $ctx->{ibx};
102 my $env = $ctx->{env};
103 my $base_url = $ibx->base_url($env);
104 $$txt .= "color customization for $base_url\n";
107 public-inbox provides a stable set of CSS classes for users to
108 customize colors for highlighting diffs and code.
110 Users of browsers such as dillo, Firefox, or some browser
111 extensions may start by downloading the following sample CSS file
112 to control the colors they see:
114 ${base_url}userContent.css
120 $$txt .= PublicInbox::UserContent::sample($ibx, $env) . "```\n";
123 # git-config section names are quoted in the config file, so escape them
126 $name =~ s/\\/\\\\/g;
131 sub _coderepo_config ($$) {
132 my ($ctx, $txt) = @_;
133 my $cr = $ctx->{ibx}->{coderepo} // return;
134 # note: this doesn't preserve cgitrc layout, since we parse cgitrc
135 # and drop the original structure
136 $$txt .= "\tcoderepo = $_\n" for @$cr;
139 ; `coderepo' entries allows blob reconstruction via patch emails if
140 ; the inbox is indexed with Xapian. `@@ <from-range> <to-range> @@'
141 ; line number ranges in `[PATCH]' emails link to /$INBOX_NAME/$OID/s/,
142 ; an HTTP endpoint which reconstructs git blobs via git-apply(1).
144 my $pi_cfg = $ctx->{www}->{pi_cfg};
145 for my $cr_name (@$cr) {
146 my $urls = $pi_cfg->get_all("coderepo.$cr_name.cgiturl");
147 my $path = "/path/to/$cr_name";
148 $cr_name = dq_escape($cr_name);
150 $$txt .= qq([coderepo "$cr_name"]\n);
151 if ($urls && scalar(@$urls)) {
153 $$txt .= join(" ||\n\t;\t", map {;
155 if ($path !~ m![a-z0-9_/\.\-]!i) {
156 $dst = '"'.dq_escape($dst).'"';
158 qq(git clone $_ $dst);
162 $$txt .= "\tdir = $path\n";
163 $$txt .= "\tcgiturl = https://example.com/";
164 $$txt .= uri_escape_utf8($cr_name, '^A-Za-z0-9\-\._~/')."\n";
168 # n.b. this is a perfect candidate for memoization
169 sub inbox_config ($$$) {
170 my ($ctx, $hdr, $txt) = @_;
171 my $ibx = $ctx->{ibx};
172 push @$hdr, 'Content-Disposition', 'inline; filename=inbox.config';
173 my $t = eval { $ibx->mm->created_at };
174 push(@$hdr, 'Last-Modified', time2str($t)) if $t;
175 my $name = dq_escape($ibx->{name});
176 my $inboxdir = '/path/to/top-level-inbox';
177 my $base_url = $ibx->base_url($ctx->{env});
179 ; Example public-inbox config snippet for a mirror of
181 ; See public-inbox-config(5) manpage for more details:
182 ; https://public-inbox.org/public-inbox-config.html
183 [publicinbox "$name"]
185 ; note: public-inbox before v1.2.0 used `mainrepo' instead of
186 ; `inboxdir', both remain supported after 1.2
188 url = https://example.com/$name/
189 url = http://example.onion/$name/
191 for my $k (qw(address listid infourl watchheader)) {
192 defined(my $v = $ibx->{$k}) or next;
193 $$txt .= "\t$k = $_\n" for @$v;
195 if (my $altid = $ibx->{altid}) {
196 my $altid_map = $ibx->altid_map;
198 ; altid DBs may be used to provide numeric article ID lookup from
199 ; old, pre-existing sources. You can recreate them via curl(1),
200 ; gzip(1), and sqlite3(1) as documented:
202 for (sort keys %$altid_map) {
203 $$txt .= "\t;\tcurl -d '' $base_url$_.sql.gz | \\\n" .
204 "\t;\tgzip -dc | \\\n" .
205 "\t;\tsqlite3 $inboxdir/$_.sqlite3\n";
206 $$txt .= "\taltid = serial:$_:file=$_.sqlite3\n";
210 for my $k (qw(filter newsgroup obfuscate replyto)) {
211 defined(my $v = $ibx->{$k}) or next;
212 $$txt .= "\t$k = $v\n";
214 $$txt .= "\timapmirror = $_\n" for (@{$ibx->imap_url($ctx)});
215 $$txt .= "\tnntpmirror = $_\n" for (@{$ibx->nntp_url($ctx)});
216 _coderepo_config($ctx, $txt);
220 # n.b. this is a perfect candidate for memoization
221 sub extindex_config ($$$) {
222 my ($ctx, $hdr, $txt) = @_;
223 my $ibx = $ctx->{ibx};
224 push @$hdr, 'Content-Disposition', 'inline; filename=extindex.config';
225 my $name = dq_escape($ibx->{name});
226 my $base_url = $ibx->base_url($ctx->{env});
228 ; Example public-inbox config snippet for the external index (extindex) at:
230 ; See public-inbox-config(5)manpage for more details:
231 ; https://public-inbox.org/public-inbox-config.html
233 topdir = /path/to/extindex-topdir
234 url = https://example.com/$name/
235 url = http://example.onion/$name/
237 for my $k (qw(infourl)) {
238 defined(my $v = $ibx->{$k}) or next;
239 $$txt .= "\t$k = $v\n";
241 _coderepo_config($ctx, $txt);
245 sub coderepos_raw ($$) {
246 my ($ctx, $top_url) = @_;
247 my $cr = $ctx->{ibx}->{coderepo} // return ();
248 my $cfg = $ctx->{www}->{pi_cfg};
249 my @ret = ('Code repositories for project(s) associated with this '.
250 $ctx->{ibx}->thing_type . "\n");
251 for my $cr_name (@$cr) {
252 my $urls = $cfg->get_all("coderepo.$cr_name.cgiturl");
255 my $u = m!\A(?:[a-z\+]+:)?//!i ? $_ :
257 $ret[0] .= "\n\t" . prurl($ctx->{env}, $u);
260 $ret[0] .= qq[\n\t$cr_name.git (no URL configured)];
263 @ret; # may be empty, this sub is called as an arg for join()
266 sub _add_non_http_urls ($$) {
267 my ($ctx, $txt) = @_;
268 $ctx->{ibx}->can('nntp_url') or return; # TODO extindex can have IMAP
269 my $urls = $ctx->{ibx}->imap_url($ctx);
271 $urls = join("\n ", @$urls);
272 $urls =~ s!://([^/@]+)/!://;AUTH=ANONYMOUS\@$1/!sg;
275 IMAP subfolder(s) are available under:
277 # each subfolder (starting with `0') holds 50K messages at most
280 $urls = $ctx->{ibx}->nntp_url($ctx);
282 $$txt .= @$urls == 1 ? "\nNewsgroup" : "\nNewsgroups are";
283 $$txt .= ' available over NNTP:';
284 $$txt .= "\n " . join("\n ", @$urls) . "\n";
286 $urls = $ctx->{ibx}->pop3_url($ctx);
288 $urls = join("\n ", @$urls);
291 POP3 access is available:
294 The POP3 password is: anonymous
295 The POP3 username is: \$(uuidgen)\@$ctx->{ibx}->{newsgroup}
296 where \$(uuidgen) in the output of the `uuidgen' command on your system.
297 The UUID in the username functions as a private cookie (don't share it).
298 Idle accounts will expire periodically.
303 sub _add_onion_note ($) {
305 $$txt =~ m!\b[^:]+://\w+\.onion/!i and $$txt .= <<EOM
307 note: .onion URLs require Tor: https://www.torproject.org/
312 sub _mirror_help ($$) {
313 my ($ctx, $txt) = @_;
314 my $ibx = $ctx->{ibx};
315 my $base_url = $ibx->base_url($ctx->{env});
316 chop $base_url; # no trailing slash for "git clone"
317 my $dir = (split(m!/!, $base_url))[-1];
318 my %seen = ($base_url => 1);
319 my $top_url = $base_url;
320 $top_url =~ s!/[^/]+\z!/!;
321 $$txt .= "public-inbox mirroring instructions\n\n";
322 if ($ibx->can('cloneurl')) { # PublicInbox::Inbox
324 "This public inbox may be cloned and mirrored by anyone:\n";
326 my $max = $ibx->max_git_epoch;
327 # TODO: some of these URLs may be too long and we may need to
328 # do something like code_footer() above, but these are local
330 if (defined($max)) { # v2
331 for my $i (0..$max) {
332 # old epochs my be deleted:
333 -d "$ibx->{inboxdir}/git/$i.git" or next;
334 my $url = "$base_url/$i";
336 push @urls, "$url $dir/git/$i.git";
338 my $nr = scalar(@urls);
340 chomp($$txt .= <<EOM);
342 # this inbox consists of $nr epochs: (no need to clone all of them)
344 $urls[0] .= " # oldest";
345 $urls[-1] .= " # newest";
348 push @urls, $base_url;
350 # FIXME: epoch splits can be different in other repositories,
351 # use the "cloneurl" file as-is for now:
352 for my $u (@{$ibx->cloneurl}) {
357 $$txt .= join('', map { " git clone --mirror $_\n" } @urls);
358 my $addrs = $ibx->{address} // 'inbox@example.com';
359 my $ng = $ibx->{newsgroup} // '';
360 substr($ng, 0, 0, ' --ng ') if $ng;
361 $addrs = join(' ', @$addrs) if ref($addrs) eq 'ARRAY';
362 my $v = defined $max ? '-V2' : '-V1';
365 # If you have public-inbox 1.1+ installed, you may
366 # initialize and index your mirror using the following commands:
367 public-inbox-init $v$ng \\
368 $ibx->{name} ./$dir $base_url \\
370 public-inbox-index ./$dir
372 } else { # PublicInbox::ExtSearch
374 This is an external index which is an amalgamation of several public inboxes.
375 Each public inbox needs to be mirrored individually.
377 my $v = $ctx->{www}->{pi_cfg}->{lc('publicInbox.wwwListing')};
378 if (($v // '') =~ /\A(?:all|match=domain)\z/) {
380 A list of them is available at $top_url
384 my $cfg_link = "$base_url/_/text/config/raw";
387 Example config snippet for mirrors: $cfg_link
389 _add_non_http_urls($ctx, $txt);
390 _add_onion_note($txt);
392 my $code_url = prurl($ctx->{env}, $PublicInbox::WwwStream::CODE_URL);
393 $$txt .= join("\n\n",
394 coderepos_raw($ctx, $top_url), # may be empty
395 "AGPL code for this site:\n git clone $code_url");
399 sub _default_text ($$$$) {
400 my ($ctx, $key, $hdr, $txt) = @_;
401 if ($key eq 'mirror') {
402 return _mirror_help($ctx, $txt);
403 } elsif ($key eq 'color') {
404 return _colors_help($ctx, $txt);
405 } elsif ($key eq 'config') {
406 return $ctx->{ibx}->can('cloneurl') ?
407 inbox_config($ctx, $hdr, $txt) :
408 extindex_config($ctx, $hdr, $txt);
410 return if $key ne 'help'; # TODO more keys?
412 my $ibx = $ctx->{ibx};
413 my $base_url = $ibx->base_url($ctx->{env});
415 public-inbox help for $base_url
420 public-inbox uses Message-ID identifiers in URLs.
421 One may look up messages by substituting Message-IDs
422 (without the leading '<' or trailing '>') into the URL.
423 Forward slash ('/') characters in the Message-IDs
424 need to be escaped as "%2F" (without quotes).
426 Thus, it is possible to retrieve any message by its
427 Message-ID by going to:
429 $base_url<Message-ID>/
430 (without the '<' or '>')
432 Message-IDs are described at:
438 # n.b. we use the Xapian DB for any regeneratable,
439 # order-of-arrival-independent data.
445 This public-inbox has search functionality provided by Xapian.
447 It supports typical AND, OR, NOT, '+', '-' queries present
448 in other search engines.
450 We also support search prefixes to limit the scope of the
451 search to certain fields.
453 Prefixes supported in this installation include:
456 _srch_prefix($ibx, $txt);
459 Most prefixes are probabilistic, meaning they support stemming
460 and wildcards ('*'). Ranges (such as 'd:') and boolean prefixes
461 do not support stemming or wildcards.
462 The upstream Xapian query parser documentation fully explains
474 Message threading is enabled for this public-inbox,
475 additional endpoints for message threads are available:
477 * $base_url<Message-ID>/T/#u
479 Loads the thread belonging to the given <Message-ID>
480 in flat chronological order. The "#u" anchor
481 focuses the browser on the given <Message-ID>.
483 * $base_url<Message-ID>/t/#u
485 Loads the thread belonging to the given <Message-ID>
486 in threaded order with nesting. For deep threads,
487 this requires a wide display or horizontal scrolling.
489 Both of these HTML endpoints are suitable for offline reading
490 using the thread overview at the bottom of each page.
492 The gzipped mbox for a thread is available for downloading and
493 importing into your favorite mail client:
495 * $base_url<Message-ID>/t.mbox.gz
497 We use the mboxrd variant of the mbox format described at:
501 Users of feed readers may follow a particular thread using:
503 * $base_url<Message-ID>/t.atom
505 Which loads the thread in Atom Syndication Standard
506 described at Wikipedia and RFC4287:
508 $WIKI_URL/Atom_(standard)
509 https://tools.ietf.org/html/rfc4287
511 Atom Threading Extensions (RFC4685) are supported:
513 https://tools.ietf.org/html/rfc4685
518 _add_non_http_urls($ctx, \(my $note = ''));
519 $note and $note =~ s/^/ /gms and $$txt .= <<EOF;
528 This help text is maintained by public-inbox developers
529 reachable via plain-text email at: meta\@public-inbox.org
530 Their inbox is archived at: https://public-inbox.org/meta/
532 # TODO: support admin contact info in ~/.public-inbox/config