1 # Copyright (C) 2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # "lei add-external --mirror" support (also "public-inbox-clone");
5 package PublicInbox::LeiMirror;
8 use parent qw(PublicInbox::IPC);
9 use PublicInbox::Config;
10 use PublicInbox::AutoReap;
11 use IO::Uncompress::Gunzip qw(gunzip $GunzipError);
12 use IO::Compress::Gzip qw(gzip $GzipError);
13 use PublicInbox::Spawn qw(popen_rd spawn);
15 use Fcntl qw(SEEK_SET O_CREAT O_EXCL O_WRONLY);
18 sub _wq_done_wait { # dwaitpid callback (via wq_eof)
20 my ($mrr, $lei) = @$arg;
21 my $f = "$mrr->{dst}/mirror.done";
23 $lei->child_error($?);
24 } elsif (!unlink($f)) {
25 warn("unlink($f): $!\n") unless $!{ENOENT};
27 if ($lei->{cmd} ne 'public-inbox-clone') {
28 $lei->lazy_cb('add-external', '_finish_'
29 )->($lei, $mrr->{dst});
31 $lei->qerr("# mirrored $mrr->{src} => $mrr->{dst}");
36 # for old installations without manifest.js.gz
39 my $uri = URI->new($self->{src});
40 my $lei = $self->{lei};
41 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
42 my $cmd = $curl->for_uri($lei, $uri, '--compressed');
43 my $opt = { 0 => $lei->{0}, 2 => $lei->{2} };
44 my $fh = popen_rd($cmd, undef, $opt);
45 my $html = do { local $/; <$fh> } // die "read(curl $uri): $!";
46 close($fh) or return $lei->child_error($?, "@$cmd failed");
48 # we grep with URL below, we don't want Subject/From headers
49 # making us clone random URLs
50 my @html = split(/<hr>/, $html);
51 my @urls = ($html[-1] =~ m!\bgit clone --mirror ([a-z\+]+://\S+)!g);
52 my $url = $uri->as_string;
53 chop($url) eq '/' or die "BUG: $uri not canonicalized";
55 # since this is for old instances w/o manifest.js.gz, try v1 first
56 return clone_v1($self) if grep(m!\A\Q$url\E/*\z!, @urls);
57 if (my @v2_urls = grep(m!\A\Q$url\E/[0-9]+\z!, @urls)) {
59 my ($n) = (m!/([0-9]+)\z!);
62 return clone_v2($self, \%v2_epochs);
65 # filter out common URLs served by WWW (e.g /$MSGID/T/)
66 if (@urls && $url =~ s!/+[^/]+\@[^/]+/.*\z!! &&
67 grep(m!\A\Q$url\E/*\z!, @urls)) {
69 E: confused by scraping <$uri>, did you mean <$url>?
73 E: confused by scraping <$uri>, got ambiguous results:
76 die "E: scraping <$uri> revealed nothing\n";
82 $opt->{$_} = $lei->{$_} for (0..2);
83 # we support "-c $key=$val" for arbitrary git config options
84 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
85 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
86 push @cmd, qw(clone --mirror);
87 push @cmd, '-q' if $lei->{opt}->{quiet};
88 push @cmd, '-v' if $lei->{opt}->{verbose};
89 # XXX any other options to support?
90 # --reference is tricky with multiple epochs...
95 my ($ft, $dst, $open_mode) = @_;
96 my $fn = $ft->filename;
98 my $mode = @st ? ($st[2] & 07777) : ($open_mode & ~umask);
99 chmod($mode, $ft) or croak "E: chmod $fn: $!";
100 rename($fn, $dst) or croak "E: rename($fn => $ft): $!";
101 $ft->unlink_on_destroy(0);
104 sub _get_txt { # non-fatal
105 my ($self, $endpoint, $file, $mode) = @_;
106 my $uri = URI->new($self->{src});
107 my $lei = $self->{lei};
108 my $path = $uri->path;
109 chop($path) eq '/' or die "BUG: $uri not canonicalized";
110 $uri->path("$path/$endpoint");
111 my $ft = File::Temp->new(TEMPLATE => "$file-XXXX", DIR => $self->{dst});
112 my $opt = { 0 => $lei->{0}, 1 => $lei->{1}, 2 => $lei->{2} };
113 my $cmd = $self->{curl}->for_uri($lei, $uri,
114 qw(--compressed -R -o), $ft->filename);
115 my $cerr = run_reap($lei, $cmd, $opt);
116 return "$uri missing" if ($cerr >> 8) == 22;
117 return "# @$cmd failed (non-fatal)" if $cerr;
118 ft_rename($ft, "$self->{dst}/$file", $mode);
122 # tries the relatively new /$INBOX/_/text/config/raw endpoint
125 my $dst = $self->{dst};
126 if (!-d $dst || !mkdir($dst)) {
128 File::Path::mkpath($dst);
129 -d $dst or die "mkpath($dst): $!\n";
131 my $err = _get_txt($self,
132 qw(_/text/config/raw inbox.config.example), 0444);
133 return warn($err, "\n") if $err;
134 my $f = "$self->{dst}/inbox.config.example";
135 my $cfg = PublicInbox::Config->git_config_dump($f, $self->{lei}->{2});
136 my $ibx = $self->{ibx} = {};
137 for my $sec (grep(/\Apublicinbox\./, @{$cfg->{-section_order}})) {
138 for (qw(address newsgroup nntpmirror)) {
139 $ibx->{$_} = $cfg->{"$sec.$_"};
144 sub set_description ($) {
146 my $f = "$self->{dst}/description";
147 open my $fh, '+>>', $f or die "open($f): $!";
148 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
149 chomp(my $d = do { local $/; <$fh> } // die "read($f): $!");
150 if ($d eq '($INBOX_DIR/description missing)' ||
151 $d =~ /^Unnamed repository/ || $d !~ /\S/) {
152 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
153 truncate($fh, 0) or die "truncate($f): $!";
154 print $fh "mirror of $self->{src}\n" or die "print($f): $!";
155 close $fh or die "close($f): $!";
159 sub index_cloned_inbox {
160 my ($self, $iv) = @_;
161 my $lei = $self->{lei};
162 my $err = _get_txt($self, qw(description description), 0666);
163 warn($err, "\n") if $err; # non fatal
164 eval { set_description($self) };
167 # n.b. public-inbox-clone works w/o (SQLite || Xapian)
168 # lei is useless without Xapian + SQLite
169 if ($lei->{cmd} ne 'public-inbox-clone') {
170 my $ibx = delete($self->{ibx}) // {
171 address => [ 'lei@example.com' ],
174 $ibx->{inboxdir} = $self->{dst};
175 PublicInbox::Inbox->new($ibx);
176 PublicInbox::InboxWritable->new($ibx);
178 for my $sw ($lei->index_opt) {
179 my ($k) = ($sw =~ /\A([\w-]+)/);
180 $opt->{$k} = $lei->{opt}->{$k};
182 # force synchronous dwaitpid for v2:
183 local $PublicInbox::DS::in_loop = 0;
184 my $cfg = PublicInbox::Config->new(undef, $lei->{2});
185 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
186 local %ENV = (%ENV, %$env) if $env;
187 PublicInbox::Admin::progress_prepare($opt, $lei->{2});
188 PublicInbox::Admin::index_inbox($ibx, undef, $opt);
190 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
194 my ($lei, $cmd, $opt) = @_;
195 $lei->qerr("# @$cmd");
196 my $ar = PublicInbox::AutoReap->new(spawn($cmd, undef, $opt));
199 $? = 0; # don't let it influence normal exit
205 my $lei = $self->{lei};
206 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
207 my $uri = URI->new($self->{src});
208 defined($lei->{opt}->{epoch}) and
209 die "$uri is a v1 inbox, --epoch is not supported\n";
210 my $pfx = $curl->torsocks($lei, $uri) or return;
211 my $cmd = [ @$pfx, clone_cmd($lei, my $opt = {}),
212 $uri->as_string, $self->{dst} ];
213 my $cerr = run_reap($lei, $cmd, $opt);
214 return $lei->child_error($cerr, "@$cmd failed") if $cerr;
216 write_makefile($self->{dst}, 1);
217 index_cloned_inbox($self, 1);
220 sub parse_epochs ($$) {
221 my ($opt_epochs, $v2_epochs) = @_; # $epcohs "LOW..HIGH"
222 $opt_epochs // return; # undef => all epochs
223 my ($lo, $dotdot, $hi, @extra) = split(/(\.\.)/, $opt_epochs);
224 undef($lo) if ($lo // '') eq '';
225 my $re = qr/\A~?[0-9]+\z/;
226 if (@extra || (($lo // '0') !~ $re) ||
227 (($hi // '0') !~ $re) ||
228 !(grep(defined, $lo, $hi))) {
230 --epoch=$opt_epochs not in the form of `LOW..HIGH', `LOW..', nor `..HIGH'
233 my @n = sort { $a <=> $b } keys %$v2_epochs;
234 for (grep(defined, $lo, $hi)) {
237 "`$_' exceeds maximum available epoch ($n[-1])\n";
239 "`$_' is lower than minimum available epoch ($n[0])\n";
240 } elsif (/\A~([0-9]+)/) {
242 $n[$off] // die "`$_' is out of range\n";
244 } else { die "`$_' not understood\n" }
246 defined($lo) && defined($hi) && $lo > $hi and die
247 "low value (`$lo') exceeds high (`$hi')\n";
248 $lo //= $n[0] if $dotdot;
249 $hi //= $n[-1] if $dotdot;
253 if (defined $v2_epochs->{$_}) {
257 "# epoch $_ is not available (non-fatal, $lo..$hi)\n";
263 sub init_placeholder ($$) {
264 my ($src, $edst) = @_;
265 PublicInbox::Import::init_bare($edst);
266 my $f = "$edst/config";
267 open my $fh, '>>', $f or die "open($f): $!";
268 print $fh <<EOM or die "print($f): $!";
271 fetch = +refs/*:refs/*
274 ; This git epoch was created read-only and "public-inbox-fetch"
275 ; will not fetch updates for it unless write permission is added.
277 close $fh or die "close:($f): $!";
280 sub clone_v2 ($$;$) {
281 my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
282 my $lei = $self->{lei};
283 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
284 my $pfx = $curl->torsocks($lei, (values %$v2_epochs)[0]) or return;
285 my $dst = $self->{dst};
286 my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
287 my (@src_edst, @read_only, @skip_nr);
288 for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
289 my $uri = $v2_epochs->{$nr};
290 my $src = $uri->as_string;
292 $src =~ m!/([0-9]+)(?:\.git)?\z! or die <<"";
293 failed to extract epoch number from $src
295 $1 + 0 == $nr or die "BUG: <$uri> miskeyed $1 != $nr";
296 $edst .= "/git/$nr.git";
297 if (!$want || $want->{$nr}) {
298 push @src_edst, $src, $edst;
299 } else { # create a placeholder so users only need to chmod +w
300 init_placeholder($src, $edst);
301 push @read_only, $edst;
305 if (@skip_nr) { # filter out the epochs we skipped
306 my $re = join('|', @skip_nr);
307 my @del = grep(m!/git/$re\.git\z!, keys %$m);
309 $self->{-culled_manifest} = 1;
311 my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
313 my $on_destroy = $lk->lock_for_scope($$);
314 my @cmd = clone_cmd($lei, my $opt = {});
315 while (my ($src, $edst) = splice(@src_edst, 0, 2)) {
316 my $cmd = [ @$pfx, @cmd, $src, $edst ];
317 my $cerr = run_reap($lei, $cmd, $opt);
318 return $lei->child_error($cerr, "@$cmd failed") if $cerr;
320 require PublicInbox::MultiGit;
321 my $mg = PublicInbox::MultiGit->new($dst, 'all.git', 'git');
322 $mg->fill_alternates;
323 for my $i ($mg->git_epochs) { $mg->epoch_cfg_set($i) }
324 for my $edst (@read_only) {
325 my @st = stat($edst) or die "stat($edst): $!";
326 chmod($st[2] & 0555, $edst) or die "chmod(a-w, $edst): $!";
328 write_makefile($self->{dst}, 2);
329 undef $on_destroy; # unlock
330 index_cloned_inbox($self, 2);
333 # PSGI mount prefixes and manifest.js.gz prefixes don't always align...
334 sub deduce_epochs ($$) {
336 my ($v1_ent, @v2_epochs);
340 $v1_ent = $m->{$path};
341 @v2_epochs = grep(m!\A\Q$path\E/git/[0-9]+\.git\z!, keys %$m);
342 } while (!defined($v1_ent) && !@v2_epochs &&
343 $path =~ s!\A(/[^/]+)/!/! and $path_pfx .= $1);
344 ($path_pfx, $v1_ent ? $path : undef, @v2_epochs);
347 sub decode_manifest ($$$) {
348 my ($fh, $fn, $uri) = @_;
350 my $gz = do { local $/; <$fh> } // die "slurp($fn): $!";
351 gunzip(\$gz => \$js, MultiStream => 1) or
352 die "gunzip($uri): $GunzipError\n";
353 my $m = eval { PublicInbox::Config->json->decode($js) };
354 die "$uri: error decoding `$js': $@\n" if $@;
355 ref($m) eq 'HASH' or die "$uri unknown type: ".ref($m);
361 my $uri = URI->new($self->{src});
362 my $lei = $self->{lei};
363 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
364 my $path = $uri->path;
365 chop($path) eq '/' or die "BUG: $uri not canonicalized";
366 $uri->path($path . '/manifest.js.gz');
367 my $pdir = $lei->rel2abs($self->{dst});
368 $pdir =~ s!/[^/]+/?\z!!;
369 my $ft = File::Temp->new(TEMPLATE => 'm-XXXX',
370 UNLINK => 1, DIR => $pdir, SUFFIX => '.tmp');
371 my $fn = $ft->filename;
372 my ($bn) = ($fn =~ m!/([^/]+)\z!);
373 my $cmd = $curl->for_uri($lei, $uri, '-R', '-o', $bn);
374 my $opt = { -C => $pdir };
375 $opt->{$_} = $lei->{$_} for (0..2);
376 my $cerr = run_reap($lei, $cmd, $opt);
378 return try_scrape($self) if ($cerr >> 8) == 22; # 404 missing
379 return $lei->child_error($cerr, "@$cmd failed");
381 my $m = eval { decode_manifest($ft, $fn, $uri) };
384 return try_scrape($self);
386 my ($path_pfx, $v1_path, @v2_epochs) = deduce_epochs($m, $path);
388 # It may be possible to have v1 + v2 in parallel someday:
389 warn(<<EOM) if defined $v1_path;
390 # `$v1_path' appears to be a v1 inbox while v2 epochs exist:
392 # ignoring $v1_path (use --inbox-version=1 to force v1 instead)
394 my %v2_epochs = map {
395 $uri->path($path_pfx.$_);
396 my ($n) = ("$uri" =~ m!/([0-9]+)\.git\z!);
399 clone_v2($self, \%v2_epochs, $m);
400 } elsif (defined $v1_path) {
403 die "E: confused by <$uri>, possible matches:\n\t",
404 join(', ', sort keys %$m), "\n";
406 if (delete $self->{-culled_manifest}) { # set by clone_v2
407 # write the smaller manifest if epochs were skipped so
408 # users won't have to delete manifest if they +w an
409 # epoch they no longer want to skip
410 my $json = PublicInbox::Config->json->encode($m);
411 gzip(\$json => $fn) or die "gzip: $GzipError";
413 ft_rename($ft, "$self->{dst}/manifest.js.gz", 0666);
416 sub start_clone_url {
418 return try_manifest($self) if $self->{src} =~ m!\Ahttps?://!;
419 die "TODO: non-HTTP/HTTPS clone of $self->{src} not supported, yet";
422 sub do_mirror { # via wq_io_do
424 my $lei = $self->{lei};
425 umask($lei->{client_umask}) if defined $lei->{client_umask};
427 my $iv = $lei->{opt}->{'inbox-version'};
429 return clone_v1($self) if $iv == 1;
430 return try_scrape($self) if $iv == 2;
431 die "bad --inbox-version=$iv\n";
433 return start_clone_url($self) if $self->{src} =~ m!://!;
434 die "TODO: cloning local directories not supported, yet";
436 $lei->fail($@) if $@;
440 my ($cls, $lei, $src, $dst) = @_;
441 my $self = bless { src => $src, dst => $dst }, $cls;
442 if ($src =~ m!https?://!) {
444 require PublicInbox::LeiCurl;
446 require PublicInbox::Lock;
447 require PublicInbox::Inbox;
448 require PublicInbox::Admin;
449 require PublicInbox::InboxWritable;
451 my ($op_c, $ops) = $lei->workers_start($self, 1);
453 $self->wq_io_do('do_mirror', []);
455 $lei->wait_wq_events($op_c, $ops);
458 sub ipc_atfork_child {
460 $self->{lei}->_lei_atfork_child;
461 $self->SUPER::ipc_atfork_child;
465 my ($dir, $ibx_ver) = @_;
466 my $f = "$dir/Makefile";
467 if (sysopen my $fh, $f, O_CREAT|O_EXCL|O_WRONLY) {
468 print $fh <<EOM or die "print($f) $!";
469 # This is a v$ibx_ver public-inbox, see the public-inbox-v$ibx_ver-format(5)
470 # manpage for more information on the format. This Makefile is
471 # intended as a familiar wrapper for users unfamiliar with
472 # public-inbox-* commands.
474 # See the respective manpages for public-inbox-fetch(1),
475 # public-inbox-index(1), etc for more information on
476 # some of the commands used by this Makefile.
478 # This Makefile will not be modified nor read by public-inbox,
479 # so you may edit it freely with your own convenience targets
480 # and notes. public-inbox-fetch will recreate it if removed.
482 print $fh <<'EOM' or die "print($f): $!";
483 # the default target:
485 @echo Common targets:
486 @echo ' make fetch - fetch from remote git repostorie(s)'
487 @echo ' make update - fetch and update index '
489 @echo Rarely needed targets:
490 @echo ' make reindex - may be needed for new features/bugfixes'
491 @echo ' make compact - rewrite Xapian storage to save space'
496 @if ! public-inbox-fetch --exit-code; \
499 test $$c -eq 127 && exit 0; \
501 elif test -f msgmap.sqlite3 || test -f public-inbox/msgmap.sqlite3; \
503 public-inbox-index; \
505 echo 'public-inbox index not initialized'; \
506 echo 'see public-inbox-index(1) man page'; \
509 public-inbox-index --reindex
513 .PHONY : help fetch update reindex compact
515 close $fh or die "close($f): $!";
517 die "open($f): $!" unless $!{EEXIST};