1 # Copyright (C) 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 # calls _finish_add_external
29 $lei->lazy_cb('add-external', '_finish_'
30 )->($lei, $mrr->{dst});
32 $lei->qerr("# mirrored $mrr->{src} => $mrr->{dst}");
37 # for old installations without manifest.js.gz
40 my $uri = URI->new($self->{src});
41 my $lei = $self->{lei};
42 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
43 my $cmd = $curl->for_uri($lei, $uri, '--compressed');
44 my $opt = { 0 => $lei->{0}, 2 => $lei->{2} };
45 my $fh = popen_rd($cmd, undef, $opt);
46 my $html = do { local $/; <$fh> } // die "read(curl $uri): $!";
47 close($fh) or return $lei->child_error($?, "@$cmd failed");
49 # we grep with URL below, we don't want Subject/From headers
50 # making us clone random URLs
51 my @html = split(/<hr>/, $html);
52 my @urls = ($html[-1] =~ m!\bgit clone --mirror ([a-z\+]+://\S+)!g);
53 my $url = $uri->as_string;
54 chop($url) eq '/' or die "BUG: $uri not canonicalized";
56 # since this is for old instances w/o manifest.js.gz, try v1 first
57 return clone_v1($self) if grep(m!\A\Q$url\E/*\z!, @urls);
58 if (my @v2_urls = grep(m!\A\Q$url\E/[0-9]+\z!, @urls)) {
60 my ($n) = (m!/([0-9]+)\z!);
61 $n => [ URI->new($_), '' ]
63 return clone_v2($self, \%v2_epochs);
66 # filter out common URLs served by WWW (e.g /$MSGID/T/)
67 if (@urls && $url =~ s!/+[^/]+\@[^/]+/.*\z!! &&
68 grep(m!\A\Q$url\E/*\z!, @urls)) {
70 E: confused by scraping <$uri>, did you mean <$url>?
74 E: confused by scraping <$uri>, got ambiguous results:
77 die "E: scraping <$uri> revealed nothing\n";
83 $opt->{$_} = $lei->{$_} for (0..2);
84 # we support "-c $key=$val" for arbitrary git config options
85 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
86 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
87 push @cmd, qw(clone --mirror);
88 push @cmd, '-q' if $lei->{opt}->{quiet};
89 push @cmd, '-v' if $lei->{opt}->{verbose};
90 # XXX any other options to support?
91 # --reference is tricky with multiple epochs...
96 my ($ft, $dst, $open_mode) = @_;
97 my $fn = $ft->filename;
99 my $mode = @st ? ($st[2] & 07777) : ($open_mode & ~umask);
100 chmod($mode, $ft) or croak "E: chmod $fn: $!";
101 rename($fn, $dst) or croak "E: rename($fn => $ft): $!";
102 $ft->unlink_on_destroy(0);
105 sub _get_txt { # non-fatal
106 my ($self, $endpoint, $file, $mode) = @_;
107 my $uri = URI->new($self->{cur_src} // $self->{src});
108 my $lei = $self->{lei};
109 my $path = $uri->path;
110 chop($path) eq '/' or die "BUG: $uri not canonicalized";
111 $uri->path("$path/$endpoint");
112 my $dst = $self->{cur_dst} // $self->{dst};
113 my $ft = File::Temp->new(TEMPLATE => "$file-XXXX", DIR => $dst);
114 my $opt = { 0 => $lei->{0}, 1 => $lei->{1}, 2 => $lei->{2} };
115 my $cmd = $self->{curl}->for_uri($lei, $uri,
116 qw(--compressed -R -o), $ft->filename);
117 my $cerr = run_reap($lei, $cmd, $opt);
118 return "$uri missing" if ($cerr >> 8) == 22;
119 return "# @$cmd failed (non-fatal)" if $cerr;
120 ft_rename($ft, "$dst/$file", $mode);
124 # tries the relatively new /$INBOX/_/text/config/raw endpoint
127 my $dst = $self->{cur_dst} // $self->{dst};
128 if (!-d $dst || !mkdir($dst)) {
130 File::Path::mkpath($dst);
131 -d $dst or die "mkpath($dst): $!\n";
133 my $err = _get_txt($self,
134 qw(_/text/config/raw inbox.config.example), 0444);
135 return warn($err, "\n") if $err;
136 my $f = "$dst/inbox.config.example";
137 my $cfg = PublicInbox::Config->git_config_dump($f, $self->{lei}->{2});
138 my $ibx = $self->{ibx} = {};
139 for my $sec (grep(/\Apublicinbox\./, @{$cfg->{-section_order}})) {
140 for (qw(address newsgroup nntpmirror)) {
141 $ibx->{$_} = $cfg->{"$sec.$_"};
146 sub set_description ($) {
148 my $dst = $self->{cur_dst} // $self->{dst};
149 my $f = "$dst/description";
150 open my $fh, '+>>', $f or die "open($f): $!";
151 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
152 chomp(my $d = do { local $/; <$fh> } // die "read($f): $!");
153 if ($d eq '($INBOX_DIR/description missing)' ||
154 $d =~ /^Unnamed repository/ || $d !~ /\S/) {
155 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
156 truncate($fh, 0) or die "truncate($f): $!";
157 my $src = $self->{cur_src} // $self->{src};
158 print $fh "mirror of $src\n" or die "print($f): $!";
159 close $fh or die "close($f): $!";
163 sub index_cloned_inbox {
164 my ($self, $iv) = @_;
165 my $lei = $self->{lei};
166 my $err = _get_txt($self, qw(description description), 0666);
167 warn($err, "\n") if $err; # non fatal
168 eval { set_description($self) };
171 # n.b. public-inbox-clone works w/o (SQLite || Xapian)
172 # lei is useless without Xapian + SQLite
173 if ($lei->{cmd} ne 'public-inbox-clone') {
174 my $ibx = delete($self->{ibx}) // {
175 address => [ 'lei@example.com' ],
178 $ibx->{inboxdir} = $self->{cur_dst} // $self->{dst};
179 PublicInbox::Inbox->new($ibx);
180 PublicInbox::InboxWritable->new($ibx);
182 for my $sw ($lei->index_opt) {
183 my ($k) = ($sw =~ /\A([\w-]+)/);
184 $opt->{$k} = $lei->{opt}->{$k};
186 # force synchronous dwaitpid for v2:
187 local $PublicInbox::DS::in_loop = 0;
188 my $cfg = PublicInbox::Config->new(undef, $lei->{2});
189 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
190 local %ENV = (%ENV, %$env) if $env;
191 PublicInbox::Admin::progress_prepare($opt, $lei->{2});
192 PublicInbox::Admin::index_inbox($ibx, undef, $opt);
194 return if defined $self->{cur_dst};
195 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
199 my ($lei, $cmd, $opt) = @_;
200 $lei->qerr("# @$cmd");
201 my $ar = PublicInbox::AutoReap->new(spawn($cmd, undef, $opt));
204 $? = 0; # don't let it influence normal exit
210 my $lei = $self->{lei};
211 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
212 my $uri = URI->new($self->{cur_src} // $self->{src});
213 defined($lei->{opt}->{epoch}) and
214 die "$uri is a v1 inbox, --epoch is not supported\n";
215 my $pfx = $curl->torsocks($lei, $uri) or return;
216 my $dst = $self->{cur_dst} // $self->{dst};
217 my $cmd = [ @$pfx, clone_cmd($lei, my $opt = {}),
218 $uri->as_string, $dst ];
219 my $cerr = run_reap($lei, $cmd, $opt);
220 return $lei->child_error($cerr, "@$cmd failed") if $cerr;
222 write_makefile($dst, 1);
223 index_cloned_inbox($self, 1);
226 sub parse_epochs ($$) {
227 my ($opt_epochs, $v2_epochs) = @_; # $epochs "LOW..HIGH"
228 $opt_epochs // return; # undef => all epochs
229 my ($lo, $dotdot, $hi, @extra) = split(/(\.\.)/, $opt_epochs);
230 undef($lo) if ($lo // '') eq '';
231 my $re = qr/\A~?[0-9]+\z/;
232 if (@extra || (($lo // '0') !~ $re) ||
233 (($hi // '0') !~ $re) ||
234 !(grep(defined, $lo, $hi))) {
236 --epoch=$opt_epochs not in the form of `LOW..HIGH', `LOW..', nor `..HIGH'
239 my @n = sort { $a <=> $b } keys %$v2_epochs;
240 for (grep(defined, $lo, $hi)) {
243 "`$_' exceeds maximum available epoch ($n[-1])\n";
245 "`$_' is lower than minimum available epoch ($n[0])\n";
246 } elsif (/\A~([0-9]+)/) {
248 $n[$off] // die "`$_' is out of range\n";
250 } else { die "`$_' not understood\n" }
252 defined($lo) && defined($hi) && $lo > $hi and die
253 "low value (`$lo') exceeds high (`$hi')\n";
254 $lo //= $n[0] if $dotdot;
255 $hi //= $n[-1] if $dotdot;
259 if (defined $v2_epochs->{$_}) {
263 "# epoch $_ is not available (non-fatal, $lo..$hi)\n";
269 sub init_placeholder ($$) {
270 my ($src, $edst) = @_;
271 PublicInbox::Import::init_bare($edst);
272 my $f = "$edst/config";
273 open my $fh, '>>', $f or die "open($f): $!";
274 print $fh <<EOM or die "print($f): $!";
277 fetch = +refs/*:refs/*
280 ; This git epoch was created read-only and "public-inbox-fetch"
281 ; will not fetch updates for it unless write permission is added.
283 close $fh or die "close:($f): $!";
286 sub clone_v2 ($$;$) {
287 my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
288 my $lei = $self->{lei};
289 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
290 my $first_uri = (map { $_->[0] } values %$v2_epochs)[0];
291 my $pfx = $curl->torsocks($lei, $first_uri) or return;
292 my $dst = $self->{cur_dst} // $self->{dst};
293 my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
294 my (@src_edst, @read_only, @skip);
295 for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
296 my ($uri, $key) = @{$v2_epochs->{$nr}};
297 my $src = $uri->as_string;
299 $src =~ m!/([0-9]+)(?:\.git)?\z! or die <<"";
300 failed to extract epoch number from $src
302 $1 + 0 == $nr or die "BUG: <$uri> miskeyed $1 != $nr";
303 $edst .= "/git/$nr.git";
304 if (!$want || $want->{$nr}) {
305 push @src_edst, $src, $edst;
306 } else { # create a placeholder so users only need to chmod +w
307 init_placeholder($src, $edst);
308 push @read_only, $edst;
312 # filter out the epochs we skipped
313 $self->{-culled_manifest} = 1 if delete(@$m{@skip});
314 my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
316 my $on_destroy = $lk->lock_for_scope($$);
317 my @cmd = clone_cmd($lei, my $opt = {});
318 while (my ($src, $edst) = splice(@src_edst, 0, 2)) {
319 my $cmd = [ @$pfx, @cmd, $src, $edst ];
320 my $cerr = run_reap($lei, $cmd, $opt);
321 return $lei->child_error($cerr, "@$cmd failed") if $cerr;
323 require PublicInbox::MultiGit;
324 my $mg = PublicInbox::MultiGit->new($dst, 'all.git', 'git');
325 $mg->fill_alternates;
326 for my $i ($mg->git_epochs) { $mg->epoch_cfg_set($i) }
327 for my $edst (@read_only) {
328 my @st = stat($edst) or die "stat($edst): $!";
329 chmod($st[2] & 0555, $edst) or die "chmod(a-w, $edst): $!";
331 write_makefile($dst, 2);
332 undef $on_destroy; # unlock
333 index_cloned_inbox($self, 2);
336 sub decode_manifest ($$$) {
337 my ($fh, $fn, $uri) = @_;
339 my $gz = do { local $/; <$fh> } // die "slurp($fn): $!";
340 gunzip(\$gz => \$js, MultiStream => 1) or
341 die "gunzip($uri): $GunzipError\n";
342 my $m = eval { PublicInbox::Config->json->decode($js) };
343 die "$uri: error decoding `$js': $@\n" if $@;
344 ref($m) eq 'HASH' or die "$uri unknown type: ".ref($m);
348 sub multi_inbox ($$$) {
349 my ($self, $path, $m) = @_;
351 # assuming everything not v2 is v1, for now
352 my @v1 = sort grep(!m!.+/git/[0-9]+\.git\z!, keys %$m);
353 my @v2_epochs = sort grep(m!.+/git/[0-9]+\.git\z!, keys %$m);
357 m!\A/(.+)/git/[0-9]+\.git\z! or die "BUG: $_";
358 push @{$v2->{$1}}, $_;
360 my $n = scalar(keys %$v2) + scalar(@v1);
361 my $ret; # { v1 => [ ... ], v2 => { $inbox_name => [ epochs ] }}
362 $ret->{v1} = \@v1 if @v1;
363 $ret->{v2} = $v2 if keys %$v2;
366 # PSGI mount prefixes and manifest.js.gz prefixes don't always align...
368 until (grep(m!\A\Q$$path\E/git/[0-9]+\.git\z!,
369 @v2_epochs) == @v2_epochs) {
370 $$path =~ s!\A(/[^/]+)/!/! or last;
374 while (!defined($m->{$$path}) && $$path =~ s!\A(/[^/]+)/!/!) {
378 ($path_pfx, $n, $ret);
381 # FIXME: this gets confused by single inbox instance w/ global manifest.js.gz
384 my $uri = URI->new($self->{src});
385 my $lei = $self->{lei};
386 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
387 my $path = $uri->path;
388 chop($path) eq '/' or die "BUG: $uri not canonicalized";
389 $uri->path($path . '/manifest.js.gz');
390 my $pdir = $lei->rel2abs($self->{dst});
391 $pdir =~ s!/[^/]+/?\z!!;
392 my $ft = File::Temp->new(TEMPLATE => 'm-XXXX',
393 UNLINK => 1, DIR => $pdir, SUFFIX => '.tmp');
394 my $fn = $ft->filename;
395 my ($bn) = ($fn =~ m!/([^/]+)\z!);
396 my $cmd = $curl->for_uri($lei, $uri, '-R', '-o', $bn);
397 my $opt = { -C => $pdir };
398 $opt->{$_} = $lei->{$_} for (0..2);
399 my $cerr = run_reap($lei, $cmd, $opt);
401 return try_scrape($self) if ($cerr >> 8) == 22; # 404 missing
402 return $lei->child_error($cerr, "@$cmd failed");
404 my $m = eval { decode_manifest($ft, $fn, $uri) };
407 return try_scrape($self);
409 my ($path_pfx, $n, $multi) = multi_inbox($self, \$path, $m);
410 if (my $v2 = delete $multi->{v2}) {
411 for my $name (sort keys %$v2) {
412 my $epochs = delete $v2->{$name};
413 my %v2_epochs = map {
414 $uri->path($n > 1 ? $path_pfx.$path.$_
416 my ($e) = ("$uri" =~ m!/([0-9]+)\.git\z!);
417 $e // die "no [0-9]+\.git in `$uri'";
418 $e => [ $uri->clone, $_ ];
420 ("$uri" =~ m!\A(.+/)git/[0-9]+\.git\z!) or
421 die "BUG: `$uri' !~ m!/git/[0-9]+.git!";
422 local $self->{cur_src} = $1;
423 local $self->{cur_dst} = $self->{dst};
424 if ($n > 1 && $uri->path =~ m!\A\Q$path_pfx$path\E/(.+)/
425 git/[0-9]+\.git\z!x) {
426 $self->{cur_dst} .= "/$1";
428 index($self->{cur_dst}, "\n") >= 0 and die <<EOM;
429 E: `$self->{cur_dst}' must not contain newline
431 clone_v2($self, \%v2_epochs, $m);
434 if (my $v1 = delete $multi->{v1}) {
435 my $p = $path_pfx.$path;
436 chop($p) if substr($p, -1, 1) eq '/';
438 for my $name (@$v1) {
439 local $self->{cur_src} = "$uri";
440 local $self->{cur_dst} = $self->{dst};
442 $self->{cur_dst} .= $name;
443 $self->{cur_src} .= $name;
445 index($self->{cur_dst}, "\n") >= 0 and die <<EOM;
446 E: `$self->{cur_dst}' must not contain newline
448 $self->{cur_src} .= '/';
452 if (delete $self->{-culled_manifest}) { # set by clone_v2
453 # write the smaller manifest if epochs were skipped so
454 # users won't have to delete manifest if they +w an
455 # epoch they no longer want to skip
456 my $json = PublicInbox::Config->json->encode($m);
457 my $mtime = (stat($fn))[9];
458 gzip(\$json => $fn) or die "gzip: $GzipError";
459 utime($mtime, $mtime, $fn) or die "utime(..., $fn): $!";
461 ft_rename($ft, "$self->{dst}/manifest.js.gz", 0666);
462 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
465 sub start_clone_url {
467 return try_manifest($self) if $self->{src} =~ m!\Ahttps?://!;
468 die "TODO: non-HTTP/HTTPS clone of $self->{src} not supported, yet";
471 sub do_mirror { # via wq_io_do
473 my $lei = $self->{lei};
474 umask($lei->{client_umask}) if defined $lei->{client_umask};
476 my $iv = $lei->{opt}->{'inbox-version'};
478 return clone_v1($self) if $iv == 1;
479 return try_scrape($self) if $iv == 2;
480 die "bad --inbox-version=$iv\n";
482 return start_clone_url($self) if $self->{src} =~ m!://!;
483 die "TODO: cloning local directories not supported, yet";
485 $lei->fail($@) if $@;
489 my ($cls, $lei, $src, $dst) = @_;
490 my $self = bless { src => $src, dst => $dst }, $cls;
491 if ($src =~ m!https?://!) {
493 require PublicInbox::LeiCurl;
495 require PublicInbox::Lock;
496 require PublicInbox::Inbox;
497 require PublicInbox::Admin;
498 require PublicInbox::InboxWritable;
500 my ($op_c, $ops) = $lei->workers_start($self, 1);
502 $self->wq_io_do('do_mirror', []);
504 $lei->wait_wq_events($op_c, $ops);
507 sub ipc_atfork_child {
509 $self->{lei}->_lei_atfork_child;
510 $self->SUPER::ipc_atfork_child;
514 my ($dir, $ibx_ver) = @_;
515 my $f = "$dir/Makefile";
516 if (sysopen my $fh, $f, O_CREAT|O_EXCL|O_WRONLY) {
517 print $fh <<EOM or die "print($f) $!";
518 # This is a v$ibx_ver public-inbox, see the public-inbox-v$ibx_ver-format(5)
519 # manpage for more information on the format. This Makefile is
520 # intended as a familiar wrapper for users unfamiliar with
521 # public-inbox-* commands.
523 # See the respective manpages for public-inbox-fetch(1),
524 # public-inbox-index(1), etc for more information on
525 # some of the commands used by this Makefile.
527 # This Makefile will not be modified nor read by public-inbox,
528 # so you may edit it freely with your own convenience targets
529 # and notes. public-inbox-fetch will recreate it if removed.
531 print $fh <<'EOM' or die "print($f): $!";
532 # the default target:
534 @echo Common targets:
535 @echo ' make fetch - fetch from remote git repostorie(s)'
536 @echo ' make update - fetch and update index '
538 @echo Rarely needed targets:
539 @echo ' make reindex - may be needed for new features/bugfixes'
540 @echo ' make compact - rewrite Xapian storage to save space'
545 @if ! public-inbox-fetch --exit-code; \
548 test $$c -eq 127 && exit 0; \
550 elif test -f msgmap.sqlite3 || test -f public-inbox/msgmap.sqlite3; \
552 public-inbox-index; \
554 echo 'public-inbox index not initialized'; \
555 echo 'see public-inbox-index(1) man page'; \
558 public-inbox-index --reindex
562 .PHONY : help fetch update reindex compact
564 close $fh or die "close($f): $!";
566 die "open($f): $!" unless $!{EEXIST};