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 !$raw;
34 if (!_default_text($ctx, $key, \$txt)) {
36 $txt = "404 Not Found ($key)\n";
38 my $env = $ctx->{env};
40 my $h = delete $ctx->{-res_hdr};
41 $txt = gzf_maybe($h, $env)->zflush($txt) if $code == 200;
42 push @$h, 'Content-Type', 'text/plain',
43 'Content-Length', length($txt);
44 return [ $code, $h, [ $txt ] ]
47 # enforce trailing slash for "wget -r" compatibility
48 if (!$have_tslash && $code == 200) {
49 my $url = $ctx->{ibx}->base_url($env);
50 $url .= "_/text/$key/";
52 return [ 302, [ 'Content-Type', 'text/plain',
54 [ "Redirecting to $url\n" ] ];
57 # Follow git commit message conventions,
58 # first line is the Subject/title
59 my ($title) = ($txt =~ /\A([^\n]*)/s);
60 $ctx->{-title_html} = ascii_html($title);
61 my $nslash = ($key =~ tr!/!/!);
62 $ctx->{-upfx} = '../../../' . ('../' x $nslash);
63 my $l = PublicInbox::Linkify->new;
66 $hl->do_hl_text(\$txt);
68 $txt = ascii_html($txt);
70 $txt = '<pre>' . $l->linkify_2($txt) . '</pre>';
71 $txt =~ s!^search$!<a\nid=search>search</a>!sm;
72 $txt =~ s!\bPOP3\b!<a\nid=pop3>POP3</a>!;
73 $txt =~ s!\b(Newsgroups?)\b!<a\nid=nntp>$1</a>!;
74 $txt =~ s!\bIMAP\b!<a\nid=imap>IMAP</a>!;
75 PublicInbox::WwwStream::html_oneshot($ctx, $code, $txt);
78 sub _srch_prefix ($$) {
82 my $help = $ibx->isrch->help;
84 for ($i = 0; $i < @$help; $i += 2) {
85 my $pfx = $help->[$i];
87 $pad = $n if $n > $pad;
89 $htxt .= $help->[$i + 1];
93 my $padding = ' ' x ($pad + 4);
94 $htxt =~ s/^/$padding/gms;
95 $htxt =~ s/^$padding(\S+)\0/" $1".(' ' x ($pad - length($1)))/egms;
96 $htxt =~ s/\f\n/\n/gs;
101 sub _colors_help ($$) {
102 my ($ctx, $txt) = @_;
103 my $ibx = $ctx->{ibx};
104 my $env = $ctx->{env};
105 my $base_url = $ibx->base_url($env);
106 $$txt .= "color customization for $base_url\n";
109 public-inbox provides a stable set of CSS classes for users to
110 customize colors for highlighting diffs and code.
112 Users of browsers such as dillo, Firefox, or some browser
113 extensions may start by downloading the following sample CSS file
114 to control the colors they see:
116 ${base_url}userContent.css
122 $$txt .= PublicInbox::UserContent::sample($ibx, $env) . "```\n";
125 # git-config section names are quoted in the config file, so escape them
128 $name =~ s/\\/\\\\/g;
133 sub _coderepo_config ($$) {
134 my ($ctx, $txt) = @_;
135 my $cr = $ctx->{ibx}->{coderepo} // return;
136 # note: this doesn't preserve cgitrc layout, since we parse cgitrc
137 # and drop the original structure
138 $$txt .= "\tcoderepo = $_\n" for @$cr;
141 ; `coderepo' entries allows blob reconstruction via patch emails if
142 ; the inbox is indexed with Xapian. `@@ <from-range> <to-range> @@'
143 ; line number ranges in `[PATCH]' emails link to /$INBOX_NAME/$OID/s/,
144 ; an HTTP endpoint which reconstructs git blobs via git-apply(1).
146 my $pi_cfg = $ctx->{www}->{pi_cfg};
147 for my $cr_name (@$cr) {
148 my $urls = $pi_cfg->get_all("coderepo.$cr_name.cgiturl");
149 my $path = "/path/to/$cr_name";
150 $cr_name = dq_escape($cr_name);
152 $$txt .= qq([coderepo "$cr_name"]\n);
153 if ($urls && scalar(@$urls)) {
155 $$txt .= join(" ||\n\t;\t", map {;
157 if ($path !~ m![a-z0-9_/\.\-]!i) {
158 $dst = '"'.dq_escape($dst).'"';
160 qq(git clone $_ $dst);
164 $$txt .= "\tdir = $path\n";
165 $$txt .= "\tcgiturl = https://example.com/";
166 $$txt .= uri_escape_utf8($cr_name, '^A-Za-z0-9\-\._~/')."\n";
170 # n.b. this is a perfect candidate for memoization
171 sub inbox_config ($$) {
172 my ($ctx, $txt) = @_;
173 my $ibx = $ctx->{ibx};
174 push @{$ctx->{-res_hdr}},
175 'Content-Disposition', 'inline; filename=inbox.config';
176 my $t = eval { $ibx->mm->created_at };
177 push(@{$ctx->{-res_hdr}}, 'Last-Modified', time2str($t)) if $t;
178 my $name = dq_escape($ibx->{name});
179 my $inboxdir = '/path/to/top-level-inbox';
180 my $base_url = $ibx->base_url($ctx->{env});
182 ; Example public-inbox config snippet for a mirror of
184 ; See public-inbox-config(5) manpage for more details:
185 ; https://public-inbox.org/public-inbox-config.html
186 [publicinbox "$name"]
188 ; note: public-inbox before v1.2.0 used `mainrepo' instead of
189 ; `inboxdir', both remain supported after 1.2
191 url = https://example.com/$name/
192 url = http://example.onion/$name/
194 for my $k (qw(address listid infourl watchheader)) {
195 defined(my $v = $ibx->{$k}) or next;
196 $$txt .= "\t$k = $_\n" for @$v;
198 if (my $altid = $ibx->{altid}) {
199 my $altid_map = $ibx->altid_map;
201 ; altid DBs may be used to provide numeric article ID lookup from
202 ; old, pre-existing sources. You can recreate them via curl(1),
203 ; gzip(1), and sqlite3(1) as documented:
205 for (sort keys %$altid_map) {
206 $$txt .= "\t;\tcurl -d '' $base_url$_.sql.gz | \\\n" .
207 "\t;\tgzip -dc | \\\n" .
208 "\t;\tsqlite3 $inboxdir/$_.sqlite3\n";
209 $$txt .= "\taltid = serial:$_:file=$_.sqlite3\n";
213 for my $k (qw(filter newsgroup obfuscate replyto)) {
214 defined(my $v = $ibx->{$k}) or next;
215 $$txt .= "\t$k = $v\n";
217 $$txt .= "\timapmirror = $_\n" for (@{$ibx->imap_url($ctx)});
218 $$txt .= "\tnntpmirror = $_\n" for (@{$ibx->nntp_url($ctx)});
219 _coderepo_config($ctx, $txt);
223 # n.b. this is a perfect candidate for memoization
224 sub extindex_config ($$) {
225 my ($ctx, $txt) = @_;
226 my $ibx = $ctx->{ibx};
227 push @{$ctx->{-res_hdr}},
228 'Content-Disposition', 'inline; filename=extindex.config';
229 my $name = dq_escape($ibx->{name});
230 my $base_url = $ibx->base_url($ctx->{env});
232 ; Example public-inbox config snippet for the external index (extindex) at:
234 ; See public-inbox-config(5)manpage for more details:
235 ; https://public-inbox.org/public-inbox-config.html
237 topdir = /path/to/extindex-topdir
238 url = https://example.com/$name/
239 url = http://example.onion/$name/
241 for my $k (qw(infourl)) {
242 defined(my $v = $ibx->{$k}) or next;
243 $$txt .= "\t$k = $v\n";
245 _coderepo_config($ctx, $txt);
249 sub coderepos_raw ($$) {
250 my ($ctx, $top_url) = @_;
251 my $cr = $ctx->{ibx}->{coderepo} // return ();
252 my $cfg = $ctx->{www}->{pi_cfg};
253 my $buf = 'Code repositories for project(s) associated with this '.
254 $ctx->{ibx}->thing_type . "\n";
255 for my $git (@{$ctx->{www}->{pi_cfg}->repo_objs($ctx->{ibx})}) {
256 for ($git->pub_urls($ctx->{env})) {
257 my $u = m!\A(?:[a-z\+]+:)?//!i ? $_ : $top_url.$_;
258 $buf .= "\n\t" . prurl($ctx->{env}, $u);
264 sub _add_non_http_urls ($$) {
265 my ($ctx, $txt) = @_;
266 $ctx->{ibx}->can('nntp_url') or return; # TODO extindex can have IMAP
267 my $urls = $ctx->{ibx}->imap_url($ctx);
269 $urls = join("\n ", @$urls);
270 $urls =~ s!://([^/@]+)/!://;AUTH=ANONYMOUS\@$1/!sg;
273 IMAP subfolder(s) are available under:
275 # each subfolder (starting with `0') holds 50K messages at most
278 $urls = $ctx->{ibx}->nntp_url($ctx);
280 $$txt .= @$urls == 1 ? "\nNewsgroup" : "\nNewsgroups are";
281 $$txt .= ' available over NNTP:';
282 $$txt .= "\n " . join("\n ", @$urls) . "\n";
284 $urls = $ctx->{ibx}->pop3_url($ctx);
286 $urls = join("\n ", @$urls);
289 POP3 access is available:
292 The POP3 password is: anonymous
293 The POP3 username is: \$(uuidgen)\@$ctx->{ibx}->{newsgroup}
294 where \$(uuidgen) in the output of the `uuidgen' command on your system.
295 The UUID in the username functions as a private cookie (don't share it).
296 Idle accounts will expire periodically.
301 sub _add_onion_note ($) {
303 $$txt =~ m!\b[^:]+://\w+\.onion/!i and $$txt .= <<EOM
305 note: .onion URLs require Tor: https://www.torproject.org/
310 sub _mirror_help ($$) {
311 my ($ctx, $txt) = @_;
312 my $ibx = $ctx->{ibx};
313 my $base_url = $ibx->base_url($ctx->{env});
314 chop $base_url; # no trailing slash for "git clone"
315 my $dir = (split(m!/!, $base_url))[-1];
316 my %seen = ($base_url => 1);
317 my $top_url = $base_url;
318 $top_url =~ s!/[^/]+\z!/!;
319 $$txt .= "public-inbox mirroring instructions\n\n";
320 if ($ibx->can('cloneurl')) { # PublicInbox::Inbox
322 "This public inbox may be cloned and mirrored by anyone:\n";
324 my $max = $ibx->max_git_epoch;
325 # TODO: some of these URLs may be too long and we may need to
326 # do something like code_footer() above, but these are local
328 if (defined($max)) { # v2
329 for my $i (0..$max) {
330 # old epochs my be deleted:
331 -d "$ibx->{inboxdir}/git/$i.git" or next;
332 my $url = "$base_url/$i";
334 push @urls, "$url $dir/git/$i.git";
336 my $nr = scalar(@urls);
338 chomp($$txt .= <<EOM);
340 # this inbox consists of $nr epochs: (no need to clone all of them)
342 $urls[0] .= " # oldest";
343 $urls[-1] .= " # newest";
346 push @urls, $base_url;
348 # FIXME: epoch splits can be different in other repositories,
349 # use the "cloneurl" file as-is for now:
350 for my $u (@{$ibx->cloneurl}) {
355 $$txt .= join('', map { " git clone --mirror $_\n" } @urls);
356 my $addrs = $ibx->{address} // 'inbox@example.com';
357 my $ng = $ibx->{newsgroup} // '';
358 substr($ng, 0, 0, ' --ng ') if $ng;
359 $addrs = join(' ', @$addrs) if ref($addrs) eq 'ARRAY';
360 my $v = defined $max ? '-V2' : '-V1';
363 # If you have public-inbox 1.1+ installed, you may
364 # initialize and index your mirror using the following commands:
365 public-inbox-init $v$ng \\
366 $ibx->{name} ./$dir $base_url \\
368 public-inbox-index ./$dir
370 } else { # PublicInbox::ExtSearch
372 This is an external index which is an amalgamation of several public inboxes.
373 Each public inbox needs to be mirrored individually.
375 my $v = $ctx->{www}->{pi_cfg}->{lc('publicInbox.wwwListing')};
376 if (($v // '') =~ /\A(?:all|match=domain)\z/) {
378 A list of them is available at $top_url
382 my $cfg_link = "$base_url/_/text/config/raw";
385 Example config snippet for mirrors: $cfg_link
387 _add_non_http_urls($ctx, $txt);
388 _add_onion_note($txt);
390 my $code_url = prurl($ctx->{env}, $PublicInbox::WwwStream::CODE_URL);
391 $$txt .= join("\n\n",
392 coderepos_raw($ctx, $top_url), # may be empty
393 "AGPL code for this site:\n git clone $code_url");
397 sub _default_text ($$$) {
398 my ($ctx, $key, $txt) = @_;
399 if ($key eq 'mirror') {
400 return _mirror_help($ctx, $txt);
401 } elsif ($key eq 'color') {
402 return _colors_help($ctx, $txt);
403 } elsif ($key eq 'config') {
404 return $ctx->{ibx}->can('cloneurl') ?
405 inbox_config($ctx, $txt) :
406 extindex_config($ctx, $txt);
408 return if $key ne 'help'; # TODO more keys?
410 my $ibx = $ctx->{ibx};
411 my $base_url = $ibx->base_url($ctx->{env});
413 public-inbox help for $base_url
418 public-inbox uses Message-ID identifiers in URLs.
419 One may look up messages by substituting Message-IDs
420 (without the leading '<' or trailing '>') into the URL.
421 Forward slash ('/') characters in the Message-IDs
422 need to be escaped as "%2F" (without quotes).
424 Thus, it is possible to retrieve any message by its
425 Message-ID by going to:
427 $base_url<Message-ID>/
428 (without the '<' or '>')
430 Message-IDs are described at:
436 # n.b. we use the Xapian DB for any regeneratable,
437 # order-of-arrival-independent data.
443 This public-inbox has search functionality provided by Xapian.
445 It supports typical AND, OR, NOT, '+', '-' queries present
446 in other search engines.
448 We also support search prefixes to limit the scope of the
449 search to certain fields.
451 Prefixes supported in this installation include:
454 _srch_prefix($ibx, $txt);
457 Most prefixes are probabilistic, meaning they support stemming
458 and wildcards ('*'). Ranges (such as 'd:') and boolean prefixes
459 do not support stemming or wildcards.
460 The upstream Xapian query parser documentation fully explains
472 Message threading is enabled for this public-inbox,
473 additional endpoints for message threads are available:
475 * $base_url<Message-ID>/T/#u
477 Loads the thread belonging to the given <Message-ID>
478 in flat chronological order. The "#u" anchor
479 focuses the browser on the given <Message-ID>.
481 * $base_url<Message-ID>/t/#u
483 Loads the thread belonging to the given <Message-ID>
484 in threaded order with nesting. For deep threads,
485 this requires a wide display or horizontal scrolling.
487 Both of these HTML endpoints are suitable for offline reading
488 using the thread overview at the bottom of each page.
490 The gzipped mbox for a thread is available for downloading and
491 importing into your favorite mail client:
493 * $base_url<Message-ID>/t.mbox.gz
495 We use the mboxrd variant of the mbox format described at:
499 Users of feed readers may follow a particular thread using:
501 * $base_url<Message-ID>/t.atom
503 Which loads the thread in Atom Syndication Standard
504 described at Wikipedia and RFC4287:
506 $WIKI_URL/Atom_(standard)
507 https://tools.ietf.org/html/rfc4287
509 Atom Threading Extensions (RFC4685) are supported:
511 https://tools.ietf.org/html/rfc4685
516 _add_non_http_urls($ctx, \(my $note = ''));
517 $note and $note =~ s/^/ /gms and $$txt .= <<EOF;
526 This help text is maintained by public-inbox developers
527 reachable via plain-text email at: meta\@public-inbox.org
528 Their inbox is archived at: https://public-inbox.org/meta/
530 # TODO: support admin contact info in ~/.public-inbox/config