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;
7 use parent qw(PublicInbox::IPC);
8 use IO::Uncompress::Gunzip qw(gunzip $GunzipError);
9 use IO::Compress::Gzip qw(gzip $GzipError);
10 use PublicInbox::Spawn qw(popen_rd spawn run_die);
14 use Fcntl qw(SEEK_SET O_CREAT O_EXCL O_WRONLY);
17 use PublicInbox::Config;
18 use PublicInbox::Inbox;
19 use PublicInbox::LeiCurl;
20 use PublicInbox::OnDestroy;
21 use Digest::SHA qw(sha256_hex sha1_hex);
22 use POSIX qw(strftime);
24 our $LIVE; # pid => callback
25 our $FGRP_TODO; # objstore -> [ fgrp mirror objects ]
26 our $TODO; # reference => [ non-fgrp mirror objects ]
29 $LIVE && (!$_[0]->{lei}->{child_error} ||
30 $_[0]->{lei}->{opt}->{'keep-going'});
33 sub _wq_done_wait { # dwaitpid callback (via wq_eof)
35 my ($mrr, $lei) = @$arg;
37 $lei->child_error($?);
38 } elsif (!$lei->{child_error}) {
39 if (!$mrr->{dry_run} && $lei->{cmd} ne 'public-inbox-clone') {
40 require PublicInbox::LeiAddExternal;
41 PublicInbox::LeiAddExternal::_finish_add_external(
44 $lei->qerr("# mirrored $mrr->{src} => $mrr->{dst}");
49 # for old installations without manifest.js.gz
51 my ($self, $fallback_manifest) = @_;
52 my $uri = URI->new($self->{src});
53 my $lei = $self->{lei};
54 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
55 my $cmd = $curl->for_uri($lei, $uri, '--compressed');
56 my $opt = { 0 => $lei->{0}, 2 => $lei->{2} };
57 my $fh = popen_rd($cmd, undef, $opt);
58 my $html = do { local $/; <$fh> } // die "read(curl $uri): $!";
59 close($fh) or return $lei->child_error($?, "@$cmd failed");
61 # we grep with URL below, we don't want Subject/From headers
62 # making us clone random URLs. This assumes remote instances
63 # prior to public-inbox 1.7.0
64 # 5b96edcb1e0d8252 (www: move mirror instructions to /text/, 2021-08-28)
65 my @html = split(/<hr>/, $html);
66 my @urls = ($html[-1] =~ m!\bgit clone --mirror ([a-z\+]+://\S+)!g);
67 if (!@urls && $fallback_manifest) {
69 W: failed to extract URLs from $uri, trying manifest.js.gz...
71 return start_clone_url($self);
73 my $url = $uri->as_string;
74 chop($url) eq '/' or die "BUG: $uri not canonicalized";
76 # since this is for old instances w/o manifest.js.gz, try v1 first
77 return clone_v1($self) if grep(m!\A\Q$url\E/*\z!, @urls);
78 if (my @v2_urls = grep(m!\A\Q$url\E/[0-9]+\z!, @urls)) {
80 my ($n) = (m!/([0-9]+)\z!);
81 $n => [ URI->new($_), '' ]
83 clone_v2_prep($self, \%v2_epochs);
84 delete local $lei->{opt}->{epoch};
89 # filter out common URLs served by WWW (e.g /$MSGID/T/)
90 if (@urls && $url =~ s!/+[^/]+\@[^/]+/.*\z!! &&
91 grep(m!\A\Q$url\E/*\z!, @urls)) {
93 E: confused by scraping <$uri>, did you mean <$url>?
97 E: confused by scraping <$uri>, got ambiguous results:
100 die "E: scraping <$uri> revealed nothing\n";
104 my ($lei, $opt) = @_;
106 $opt->{$_} = $lei->{$_} for (0..2);
107 # we support "-c $key=$val" for arbitrary git config options
108 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
109 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
110 push @cmd, qw(clone --mirror);
111 push @cmd, '-q' if $lei->{opt}->{quiet} ||
112 ($lei->{opt}->{jobs} // 1) > 1;
113 push @cmd, '-v' if $lei->{opt}->{verbose};
114 # XXX any other options to support?
115 # --reference is tricky with multiple epochs, but handled
116 # automatically if using manifest.js.gz
120 sub ft_rename ($$$;$) {
121 my ($ft, $dst, $open_mode, $fh) = @_;
122 my @st = stat($fh // $dst);
123 my $mode = @st ? ($st[2] & 07777) : ($open_mode & ~umask);
124 chmod($mode, $ft) or croak "E: chmod($ft): $!";
126 File::Copy::mv($ft->filename, $dst) or croak "E: mv($ft => $dst): $!";
127 $ft->unlink_on_destroy(0);
131 my ($self, $jobs) = @_;
132 $jobs //= $self->{-jobs} //= $self->{lei}->{opt}->{jobs} // 1;
133 $jobs = 1 if $jobs < 1;
134 while (keys(%$LIVE) >= $jobs) {
135 my $pid = waitpid(-1, 0) // die "waitpid(-1): $!";
136 if (my $x = delete $LIVE->{$pid}) {
140 warn "reaped unknown PID=$pid ($?)\n";
145 sub _get_txt_start { # non-fatal
146 my ($self, $endpoint, $fini) = @_;
147 my $uri = URI->new($self->{cur_src} // $self->{src});
148 my $lei = $self->{lei};
149 my $path = $uri->path;
150 chop($path) eq '/' or die "BUG: $uri not canonicalized";
151 $uri->path("$path/$endpoint");
152 my $f = (split(m!/!, $endpoint))[-1];
153 my $ft = File::Temp->new(TEMPLATE => "$f-XXXX", TMPDIR => 1);
154 my $opt = { 0 => $lei->{0}, 1 => $lei->{1}, 2 => $lei->{2} };
155 my $cmd = $self->{curl}->for_uri($lei, $uri, qw(--compressed -R -o),
158 $lei->qerr("# @$cmd");
159 return if $self->{dry_run};
160 $self->{"-get_txt.$endpoint"} = [ $ft, $cmd, $uri ];
161 $LIVE->{spawn($cmd, undef, $opt)} =
162 [ \&_get_txt_done, $self, $endpoint, $fini ];
165 sub _get_txt_done { # returns true on error (non-fatal), undef on success
166 my ($self, $endpoint) = @_;
167 my ($fh, $cmd, $uri) = @{delete $self->{"-get_txt.$endpoint"}};
169 $? = 0; # don't influence normal lei exit
170 return warn("$uri missing\n") if ($cerr >> 8) == 22;
171 return warn("# @$cmd failed (non-fatal)\n") if $cerr;
172 seek($fh, SEEK_SET, 0) or die "seek: $!";
173 $self->{"mtime.$endpoint"} = (stat($fh))[9];
175 $self->{"txt.$endpoint"} = <$fh>;
179 sub _write_inbox_config {
181 my $buf = delete($self->{'txt._/text/config/raw'}) // return;
182 my $dst = $self->{cur_dst} // $self->{dst};
183 my $f = "$dst/inbox.config.example";
184 open my $fh, '>', $f or die "open($f): $!";
185 print $fh $buf or die "print: $!";
186 chmod(0444 & ~umask, $fh) or die "chmod($f): $!";
187 my $mtime = delete $self->{'mtime._/text/config/raw'};
188 $fh->flush or die "flush($f): $!";
189 if (defined $mtime) {
190 utime($mtime, $mtime, $fh) or die "utime($f): $!";
192 my $cfg = PublicInbox::Config->git_config_dump($f, $self->{lei}->{2});
193 my $ibx = $self->{ibx} = {};
194 for my $sec (grep(/\Apublicinbox\./, @{$cfg->{-section_order}})) {
195 for (qw(address newsgroup nntpmirror)) {
196 $ibx->{$_} = $cfg->{"$sec.$_"};
201 sub set_description ($) {
203 my $dst = $self->{cur_dst} // $self->{dst};
204 chomp(my $orig = PublicInbox::Git::try_cat("$dst/description"));
206 while (defined($d) && ($d =~ m!^\(\$INBOX_DIR/description missing\)! ||
207 $d =~ /^Unnamed repository/ || $d !~ /\S/)) {
208 $d = delete($self->{'txt.description'});
210 $d //= 'mirror of '.($self->{cur_src} // $self->{src});
211 atomic_write($dst, 'description', $d."\n") if $d ne $orig;
214 sub index_cloned_inbox {
215 my ($self, $iv) = @_;
216 my $lei = $self->{lei};
218 # n.b. public-inbox-clone works w/o (SQLite || Xapian)
219 # lei is useless without Xapian + SQLite
220 if ($lei->{cmd} ne 'public-inbox-clone') {
221 require PublicInbox::InboxWritable;
222 require PublicInbox::Admin;
223 my $ibx = delete($self->{ibx}) // {
224 address => [ 'lei@example.com' ],
227 $ibx->{inboxdir} = $self->{cur_dst} // $self->{dst};
228 PublicInbox::Inbox->new($ibx);
229 PublicInbox::InboxWritable->new($ibx);
231 for my $sw ($lei->index_opt) {
232 my ($k) = ($sw =~ /\A([\w-]+)/);
233 $opt->{$k} = $lei->{opt}->{$k};
235 # force synchronous dwaitpid for v2:
236 local $PublicInbox::DS::in_loop = 0;
237 my $cfg = PublicInbox::Config->new(undef, $lei->{2});
238 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
239 local %ENV = (%ENV, %$env) if $env;
240 PublicInbox::Admin::progress_prepare($opt, $lei->{2});
241 PublicInbox::Admin::index_inbox($ibx, undef, $opt);
243 return if defined $self->{cur_dst}; # one of many repos to clone
247 my ($lei, $cmd, $opt) = @_;
248 $lei->qerr("# @$cmd");
249 waitpid(spawn($cmd, undef, $opt), 0) // die "waitpid: $!";
251 $? = 0; # don't let it influence normal exit
256 my ($self, $cmd, $opt, $fini) = @_;
258 $self->{lei}->qerr("# @$cmd");
259 return if $self->{dry_run};
260 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_cmd, $self, $cmd, $fini ]
263 sub fetch_args ($$) {
264 my ($lei, $opt) = @_;
265 my @cmd; # (git --git-dir=...) to be added by caller
266 $opt->{$_} = $lei->{$_} for (0..2);
267 # we support "-c $key=$val" for arbitrary git config options
268 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
269 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
271 push @cmd, '-q' if $lei->{opt}->{quiet} ||
272 ($lei->{opt}->{jobs} // 1) > 1;
273 push @cmd, '-v' if $lei->{opt}->{verbose};
274 push(@cmd, '-p') if $lei->{opt}->{prune};
278 sub upr { # feed `git update-ref --stdin -z' verbosely
279 my ($lei, $w, $op, @rest) = @_; # ($ref, $oid) = @rest
280 $lei->qerr("# $op @rest") if $lei->{opt}->{verbose};
281 print $w "$op ", join("\0", @rest, '') or die "print(w): $!";
286 return if !keep_going($fgrp);
287 my $srcfh = delete $fgrp->{srcfh} or return;
288 my $dstfh = delete $fgrp->{dstfh} or return;
289 seek($srcfh, SEEK_SET, 0) or die "seek(src): $!";
290 seek($dstfh, SEEK_SET, 0) or die "seek(dst): $!";
291 my %src = map { chomp; split(/\0/) } (<$srcfh>);
293 my %dst = map { chomp; split(/\0/) } (<$dstfh>);
295 pipe(my ($r, $w)) or die "pipe: $!";
296 my $cmd = [ 'git', "--git-dir=$fgrp->{cur_dst}",
297 qw(update-ref --stdin -z) ];
298 my $lei = $fgrp->{lei};
299 my $pack = PublicInbox::OnDestroy->new($$, \&pack_dst, $fgrp);
300 start_cmd($fgrp, $cmd, { 0 => $r, 2 => $lei->{2} }, $pack);
301 close $r or die "close(r): $!";
302 return if $fgrp->{dry_run};
303 for my $ref (keys %dst) {
304 my $new = delete $src{$ref};
305 my $old = $dst{$ref};
308 upr($lei, $w, 'update', $ref, $new, $old);
310 upr($lei, $w, 'delete', $ref, $old);
313 while (my ($ref, $oid) = each %src) {
314 upr($lei, $w, 'create', $ref, $oid);
316 close($w) or warn "E: close(update-ref --stdin): $! (need git 1.8.5+)\n";
319 sub pack_dst { # packs lightweight satellite repos
321 pack_refs($fgrp, $fgrp->{cur_dst});
325 my ($self, $git_dir) = @_;
326 my $cmd = [ 'git', "--git-dir=$git_dir", qw(pack-refs --all --prune) ];
327 start_cmd($self, $cmd, { 2 => $self->{lei}->{2} });
333 my $first = $fgrpv->[0] // die 'BUG: no fgrpv->[0]';
334 return if !keep_going($first);
335 pack_refs($first, $first->{-osdir}); # objstore refs always packed
336 for my $fgrp (@$fgrpv) {
337 my $rn = $fgrp->{-remote};
338 my %opt = ( 2 => $fgrp->{lei}->{2} );
340 my $update_ref = PublicInbox::OnDestroy->new($$,
341 \&fgrp_update, $fgrp);
343 my $src = [ 'git', "--git-dir=$fgrp->{-osdir}", 'for-each-ref',
344 "--format=refs/%(refname:lstrip=3)%00%(objectname)",
345 "refs/remotes/$rn/" ];
346 open(my $sfh, '+>', undef) or die "open(src): $!";
347 $fgrp->{srcfh} = $sfh;
348 start_cmd($fgrp, $src, { %opt, 1 => $sfh }, $update_ref);
349 my $dst = [ 'git', "--git-dir=$fgrp->{cur_dst}", 'for-each-ref',
350 '--format=%(refname)%00%(objectname)' ];
351 open(my $dfh, '+>', undef) or die "open(dst): $!";
352 $fgrp->{dstfh} = $dfh;
353 start_cmd($fgrp, $dst, { %opt, 1 => $dfh }, $update_ref);
359 my $todo = $FGRP_TODO;
360 $FGRP_TODO = \'BUG on further use';
361 keys(%$todo) or return;
363 # Rely on the fgrptmp remote groups in the config file rather
364 # than listing all remotes since the remote name list may exceed
365 # system argv limits:
368 my @git = (@{$self->{-torsocks}}, 'git');
369 my $j = $self->{lei}->{opt}->{jobs};
372 local $self->{lei}->{opt}->{jobs} = 1;
373 (fetch_args($self->{lei}, $opt),
374 qw(--no-tags --multiple));
376 push(@fetch, "-j$j") if $j;
377 while (my ($osdir, $fgrpv) = each %$todo) {
378 my $f = "$osdir/config";
380 # clobber group from previous run atomically
381 my $cmd = ['git', "--git-dir=$osdir", qw(config -f),
382 $f, '--unset-all', "remotes.$grp"];
383 $self->{lei}->qerr("# @$cmd");
384 if (!$self->{dry_run}) {
385 my $pid = spawn($cmd, undef, { 2 => $self->{lei}->{2} });
386 waitpid($pid, 0) // die "waitpid: $!";
387 die "E: @$cmd: \$?=$?" if ($? && ($? >> 8) != 5);
389 # update the config atomically via O_APPEND while
390 # respecting git-config locking
391 sysopen(my $lk, "$f.lock", O_CREAT|O_EXCL|O_WRONLY)
392 or die "open($f.lock): $!";
393 open my $fh, '>>', $f or die "open(>>$f): $!";
395 my $buf = join('', "[remotes]\n",
396 map { "\t$grp = $_->{-remote}\n" } @$fgrpv);
397 print $fh $buf or die "print($f): $!";
398 close $fh or die "close($f): $!";
399 unlink("$f.lock") or die "unlink($f.lock): $!";
402 $cmd = [ @git, "--git-dir=$osdir", @fetch, $grp ];
403 my $end = PublicInbox::OnDestroy->new($$, \&fgrpv_done, $fgrpv);
404 start_cmd($self, $cmd, $opt, $end);
408 # keep this idempotent for future use by public-inbox-fetch
410 my ($self, $uri) = @_;
411 $self->{-ent} // return;
412 my $os = $self->{-objstore} // return;
413 my $fg = $self->{-ent}->{forkgroup} // return;
414 my $dir = "$os/$fg.git";
415 if (!-d $dir && !$self->{dry_run}) {
416 PublicInbox::Import::init_bare($dir);
417 my @cmd = ('git', "--git-dir=$dir", 'config');
418 my $opt = { 2 => $self->{lei}->{2} };
419 for ('repack.useDeltaIslands=true',
420 'pack.island=refs/remotes/([^/]+)/') {
421 run_die([@cmd, split(/=/, $_, 2)], undef, $opt);
424 my $key = $self->{-key} // die 'BUG: no -key';
425 my $rn = substr(sha256_hex($key), 0, 16);
426 if (!-d $self->{cur_dst} && !$self->{dry_run}) {
427 my $alt = File::Spec->rel2abs("$dir/objects");
428 PublicInbox::Import::init_bare($self->{cur_dst});
429 my $o = "$self->{cur_dst}/objects";
430 my $f = "$o/info/alternates";
431 my $l = File::Spec->abs2rel($alt, File::Spec->rel2abs($o));
432 open my $fh, '+>>', $f or die "open($f): $!";
433 seek($fh, SEEK_SET, 0) or die "seek($f): $!";
434 chomp(my @cur = <$fh>);
435 if (!grep(/\A\Q$l\E\z/, @cur)) {
436 say $fh $l or die "say($f): $!";
438 close $fh or die "close($f): $!";
439 $f = "$self->{cur_dst}/config";
440 open $fh, '+>>', $f or die "open:($f): $!";
441 print $fh <<EOM or die "print($f): $!";
442 ; rely on the "$rn" remote in the
443 ; $fg fork group for fetches
444 ; only uncomment the following iff you detach from fork groups
447 ; fetch = +refs/*:refs/*
450 close $fh or die "close($f): $!";
453 %$self, -osdir => $dir, -remote => $rn, -uri => $uri
458 my ($self, $cb, @arg) = @_;
459 return if !keep_going($self);
460 my $fh = delete $self->{-show_ref} // die 'BUG: no show-ref output';
461 seek($fh, SEEK_SET, 0) or die "seek(show_ref): $!";
462 $self->{-ent} // die 'BUG: no -ent';
463 my $A = $self->{-ent}->{fingerprint} // die 'BUG: no fingerprint';
464 my $B = sha1_hex(do { local $/; <$fh> } // die("read(show_ref): $!"));
465 return $cb->($self, @arg) if $A ne $B;
466 $self->{lei}->qerr("# $self->{-key} up-to-date");
470 my ($self, $cb, @arg) = @_;
471 # $cb is either resume_fetch or fgrp_enqueue
472 $self->{-ent} // return $cb->($self, @arg);
473 my $new = $self->{-ent}->{fingerprint} // return $cb->($self, @arg);
474 my $key = $self->{-key} // die 'BUG: no -key';
475 if (my $cur_ent = $self->{-local_manifest}->{$key}) {
476 # runs go_fetch->DESTROY run if eq
477 return if $cur_ent->{fingerprint} eq $new;
479 my $dst = $self->{cur_dst} // $self->{dst};
480 my $cmd = ['git', "--git-dir=$dst", 'show-ref'];
481 my $opt = { 2 => $self->{lei}->{2} };
482 open($opt->{1}, '+>', undef) or die "open(tmp): $!";
483 $self->{-show_ref} = $opt->{1};
484 my $done = PublicInbox::OnDestroy->new($$, \&fp_done, $self, $cb, @arg);
485 start_cmd($self, $cmd, $opt, $done);
489 my ($self, $uri, $fini) = @_;
490 return if !keep_going($self);
491 my $dst = $self->{cur_dst} // $self->{dst};
492 my @git = ('git', "--git-dir=$dst");
493 my $opt = { 2 => $self->{lei}->{2} };
494 my $rn = 'origin'; # configurable?
495 for ("url=$uri", "fetch=+refs/*:refs/*", 'mirror=true') {
496 my @kv = split(/=/, $_, 2);
497 $kv[0] = "remote.$rn.$kv[0]";
498 next if $self->{dry_run};
499 run_die([@git, 'config', @kv], undef, $opt);
501 my $cmd = [ @{$self->{-torsocks}}, @git,
502 fetch_args($self->{lei}, $opt), $rn ];
503 push @$cmd, '-P' if $self->{lei}->{prune}; # --prune-tags implied
504 start_cmd($self, $cmd, $opt, $fini);
508 my ($fgrp, $end) = @_; # $end calls fgrp_fetch_all
509 return if !keep_going($fgrp);
510 my $opt = { 2 => $fgrp->{lei}->{2} };
511 # --no-tags is required to avoid conflicts
512 my $u = $fgrp->{-uri} // die 'BUG: no {-uri}';
513 my $rn = $fgrp->{-remote} // die 'BUG: no {-remote}';
514 my @cmd = ('git', "--git-dir=$fgrp->{-osdir}", 'config');
515 for ("url=$u", "fetch=+refs/*:refs/remotes/$rn/*", 'tagopt=--no-tags') {
516 my @kv = split(/=/, $_, 2);
517 $kv[0] = "remote.$rn.$kv[0]";
518 $fgrp->{dry_run} ? $fgrp->{lei}->qerr("# @cmd @kv") :
519 run_die([@cmd, @kv], undef, $opt);
521 push @{$FGRP_TODO->{$fgrp->{-osdir}}}, $fgrp;
525 my ($self, $end) = @_;
526 my $lei = $self->{lei};
527 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
528 my $uri = URI->new($self->{cur_src} // $self->{src});
529 my $path = $uri->path;
530 $path =~ s!/*\z!! and $uri->path($path);
531 defined($lei->{opt}->{epoch}) and
532 die "$uri is a v1 inbox, --epoch is not supported\n";
533 $self->{-torsocks} //= $curl->torsocks($lei, $uri) or return;
534 my $dst = $self->{cur_dst} // $self->{dst};
535 my $fini = PublicInbox::OnDestroy->new($$, \&v1_done, $self);
536 my $resume = -d $dst;
537 if (my $fgrp = forkgroup_prep($self, $uri)) {
538 $fgrp->{-fini} = $fini;
539 $resume ? cmp_fp_do($fgrp, \&fgrp_enqueue, $end)
540 : fgrp_enqueue($fgrp, $end);
542 cmp_fp_do($self, \&resume_fetch, $uri, $fini);
543 } else { # normal clone
544 my $cmd = [ @{$self->{-torsocks}},
545 clone_cmd($lei, my $opt = {}), "$uri", $dst ];
546 if (defined($self->{-ent})) {
547 if (defined(my $ref = $self->{-ent}->{reference})) {
548 -e "$self->{dst}$ref" and
549 push @$cmd, '--reference',
553 start_cmd($self, $cmd, $opt, $fini);
555 if (!$self->{-is_epoch} && $lei->{opt}->{'inbox-config'} =~
556 /\A(?:always|v1)\z/s) {
557 _get_txt_start($self, '_/text/config/raw', $fini);
560 my $d = $self->{-ent} ? $self->{-ent}->{description} : undef;
561 $self->{'txt.description'} = $d if defined $d;
562 (!defined($d) && !$end) and
563 _get_txt_start($self, 'description', $fini);
565 $end or do_reap($self, 1); # for non-manifest clone
568 sub parse_epochs ($$) {
569 my ($opt_epochs, $v2_epochs) = @_; # $epochs "LOW..HIGH"
570 $opt_epochs // return; # undef => all epochs
571 my ($lo, $dotdot, $hi, @extra) = split(/(\.\.)/, $opt_epochs);
572 undef($lo) if ($lo // '') eq '';
573 my $re = qr/\A~?[0-9]+\z/;
574 if (@extra || (($lo // '0') !~ $re) ||
575 (($hi // '0') !~ $re) ||
576 !(grep(defined, $lo, $hi))) {
578 --epoch=$opt_epochs not in the form of `LOW..HIGH', `LOW..', nor `..HIGH'
581 my @n = sort { $a <=> $b } keys %$v2_epochs;
582 for (grep(defined, $lo, $hi)) {
585 "`$_' exceeds maximum available epoch ($n[-1])\n";
587 "`$_' is lower than minimum available epoch ($n[0])\n";
588 } elsif (/\A~([0-9]+)/) {
590 $n[$off] // die "`$_' is out of range\n";
592 } else { die "`$_' not understood\n" }
594 defined($lo) && defined($hi) && $lo > $hi and die
595 "low value (`$lo') exceeds high (`$hi')\n";
596 $lo //= $n[0] if $dotdot;
597 $hi //= $n[-1] if $dotdot;
601 if (defined $v2_epochs->{$_}) {
605 "# epoch $_ is not available (non-fatal, $lo..$hi)\n";
611 sub init_placeholder ($$$) {
612 my ($src, $edst, $ent) = @_;
613 PublicInbox::Import::init_bare($edst);
614 my $f = "$edst/config";
615 open my $fh, '>>', $f or die "open($f): $!";
616 print $fh <<EOM or die "print($f): $!";
619 fetch = +refs/*:refs/*
622 ; This git epoch was created read-only and "public-inbox-fetch"
623 ; will not fetch updates for it unless write permission is added.
624 ; Hint: chmod +w $edst
626 if (defined($ent->{owner})) {
627 print $fh <<EOM or die "print($f): $!";
629 owner = $ent->{owner}
632 close $fh or die "close($f): $!";
633 my %map = (head => 'HEAD', description => undef);
634 while (my ($key, $fn) = each %map) {
635 my $val = $ent->{$key} // next;
638 open $fh, '>', $fn or die "open($fn): $!";
639 print $fh $val, "\n" or die "print($fn): $!";
640 close $fh or die "close($fn): $!";
644 sub reap_cmd { # async, called via SIGCHLD
645 my ($self, $cmd) = @_;
647 $? = 0; # don't let it influence normal exit
648 $self->{lei}->child_error($cerr, "@$cmd failed (\$?=$cerr)") if $cerr;
653 return if !keep_going($self);
654 my $fh = delete $self->{-show_ref_up} // die 'BUG: no show-ref output';
655 seek($fh, SEEK_SET, 0) or die "seek(show_ref): $!";
656 $self->{-ent} // die 'BUG: no -ent';
657 my $A = $self->{-ent}->{fingerprint} // die 'BUG: no fingerprint';
658 my $B = sha1_hex(do { local $/; <$fh> } // die("read(show_ref): $!"));
660 $self->{-ent}->{fingerprint} = $B;
661 push @{$self->{chg}->{fp_mismatch}}, $self->{-key};
664 sub atomic_write ($$$) {
665 my ($dn, $bn, $raw) = @_;
666 my $ft = File::Temp->new(DIR => $dn, TEMPLATE => "$bn-XXXX");
667 print $ft $raw or die "print($ft): $!";
668 $ft->flush or die "flush($ft): $!";
669 ft_rename($ft, "$dn/$bn", 0666);
672 # modifies the to-be-written manifest entry, and sets values from it, too
675 my $key = $self->{-key} // die 'BUG: no -key';
676 my $new = $self->{-ent}->{fingerprint};
677 my $cur = $self->{-local_manifest}->{$key}->{fingerprint} // "\0";
678 my $dst = $self->{cur_dst} // $self->{dst};
679 if (defined($new) && $new ne $cur) {
680 my $cmd = ['git', "--git-dir=$dst", 'show-ref'];
681 my $opt = { 2 => $self->{lei}->{2} };
682 open($opt->{1}, '+>', undef) or die "open(tmp): $!";
683 $self->{-show_ref_up} = $opt->{1};
684 my $done = PublicInbox::OnDestroy->new($$, \&up_fp_done, $self);
685 start_cmd($self, $cmd, $opt, $done);
687 $new = $self->{-ent}->{head};
688 $cur = $self->{-local_manifest}->{$key}->{head} // "\0";
689 if (defined($new) && $new ne $cur) {
690 # n.b. grokmirror writes raw contents to $dst/HEAD w/o locking
691 my $cmd = [ 'git', "--git-dir=$dst" ];
692 if ($new =~ s/\Aref: //) {
693 push @$cmd, qw(symbolic-ref HEAD), $new;
694 } elsif ($new =~ /\A[a-f0-9]{40,}\z/) {
695 push @$cmd, qw(update-ref --no-deref HEAD), $new;
698 warn "W: $key: {head} => `$new' not understood\n";
700 start_cmd($self, $cmd, { 2 => $self->{lei}->{2} }) if $cmd;
702 if (my $symlinks = $self->{-ent}->{symlinks}) {
703 my $top = File::Spec->rel2abs($self->{dst});
704 for my $p (@$symlinks) {
707 my (undef, $dn, $bn) = File::Spec->splitpath($ln);
708 File::Path::mkpath($dn);
709 my $tgt = "$top/$key";
710 $tgt = File::Spec->abs2rel($tgt, $dn);
713 next if readlink($ln) eq $tgt;
714 unlink($ln) or die "unlink($ln): $!";
716 push @{$self->{chg}->{badlink}}, $p;
719 symlink($tgt, $ln) or die "symlink($tgt, $ln): $!";
722 if (defined(my $t = $self->{-ent}->{modified})) {
723 my ($dn, $bn) = ("$dst/info/web", 'last-modified');
724 my $orig = PublicInbox::Git::try_cat("$dn/$bn");
725 $t = strftime('%F %T', gmtime($t))." +0000\n";
726 File::Path::mkpath($dn);
727 atomic_write($dn, $bn, $t) if $orig ne $t;
730 $new = $self->{-ent}->{owner} // return;
731 $cur = $self->{-local_manifest}->{$key}->{owner} // "\0";
732 return if $cur eq $new;
733 my $cmd = [ qw(git config -f), "$dst/config", 'gitweb.owner', $new ];
734 start_cmd($self, $cmd, { 2 => $self->{lei}->{2} });
737 sub v1_done { # called via OnDestroy
739 return if $self->{dry_run} || !keep_going($self);
740 _write_inbox_config($self);
741 my $dst = $self->{cur_dst} // $self->{dst};
742 update_ent($self) if $self->{-ent};
743 my $o = "$dst/objects";
744 if (open(my $fh, '<', my $fn = "$o/info/alternates")) {;
745 my $base = File::Spec->rel2abs($o);
750 $_ = File::Spec->abs2rel($_, $base);
751 $ft //= File::Temp->new(TEMPLATE => '.XXXX',
755 print $ft @l or die "print($ft): $!";
756 $ft->flush or die "flush($ft): $!";
757 ft_rename($ft, $fn, 0666, $fh);
760 eval { set_description($self) };
762 return if ($self->{-is_epoch} ||
763 $self->{lei}->{opt}->{'inbox-config'} ne 'always');
764 write_makefile($dst, 1);
765 index_cloned_inbox($self, 1);
768 sub v2_done { # called via OnDestroy
770 return if $self->{dry_run} || !keep_going($self);
771 my $dst = $self->{cur_dst} // $self->{dst};
772 require PublicInbox::Lock;
773 my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
774 my $lck = $lk->lock_for_scope($$);
775 _write_inbox_config($self);
776 require PublicInbox::MultiGit;
777 my $mg = PublicInbox::MultiGit->new($dst, 'all.git', 'git');
778 $mg->fill_alternates;
779 for my $i ($mg->git_epochs) { $mg->epoch_cfg_set($i) }
780 for my $edst (@{delete($self->{-read_only}) // []}) {
781 my @st = stat($edst) or die "stat($edst): $!";
782 chmod($st[2] & 0555, $edst) or die "chmod(a-w, $edst): $!";
784 write_makefile($dst, 2);
786 eval { set_description($self) };
788 index_cloned_inbox($self, 2);
791 sub clone_v2_prep ($$;$) {
792 my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
793 my $lei = $self->{lei};
794 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
795 my $first_uri = (map { $_->[0] } values %$v2_epochs)[0];
796 $self->{-torsocks} //= $curl->torsocks($lei, $first_uri) or return;
797 my $dst = $self->{cur_dst} // $self->{dst};
798 my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
799 my $task = $m ? bless { %$self }, __PACKAGE__ : $self;
801 my $fini = PublicInbox::OnDestroy->new($$, \&v2_done, $task);
802 for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
803 my ($uri, $key) = @{$v2_epochs->{$nr}};
804 my $src = $uri->as_string;
806 $src =~ m!/([0-9]+)(?:\.git)?\z! or die <<"";
807 failed to extract epoch number from $src
809 $1 + 0 == $nr or die "BUG: <$uri> miskeyed $1 != $nr";
810 $edst .= "/git/$nr.git";
814 die("BUG: `$key' not in manifest.js.gz");
815 if (defined(my $d = $ent->{description})) {
816 $d =~ s/ \[epoch [0-9]+\]\z//s;
820 if (!$want || $want->{$nr}) {
821 my $etask = bless { %$task, -key => $key }, __PACKAGE__;
822 $etask->{-ent} = $ent; # may have {reference}
823 $etask->{cur_src} = $src;
824 $etask->{cur_dst} = $edst;
825 $etask->{-is_epoch} = $fini;
826 my $ref = $ent->{reference} // '';
827 push @{$TODO->{$ref}}, $etask;
828 $self->{any_want}->{$key} = 1;
829 } else { # create a placeholder so users only need to chmod +w
830 init_placeholder($src, $edst, $ent);
831 push @{$task->{-read_only}}, $edst;
835 # filter out the epochs we skipped
836 $self->{chg}->{manifest} = 1 if $m && delete(@$m{@skip});
838 (!$self->{dry_run} && !-d $dst) and File::Path::mkpath($dst);
840 $lei->{opt}->{'inbox-config'} =~ /\A(?:always|v2)\z/s and
841 _get_txt_start($task, '_/text/config/raw', $fini);
843 defined($desc) ? ($task->{'txt.description'} = $desc) :
844 _get_txt_start($task, 'description', $fini);
847 sub decode_manifest ($$$) {
848 my ($fh, $fn, $uri) = @_;
850 my $gz = do { local $/; <$fh> } // die "slurp($fn): $!";
851 gunzip(\$gz => \$js, MultiStream => 1) or
852 die "gunzip($uri): $GunzipError\n";
853 my $m = eval { PublicInbox::Config->json->decode($js) };
854 die "$uri: error decoding `$js': $@\n" if $@;
855 ref($m) eq 'HASH' or die "$uri unknown type: ".ref($m);
859 sub load_current_manifest ($) {
861 my $fn = $self->{-manifest} // return;
862 if (open(my $fh, '<', $fn)) {
863 decode_manifest($fh, $fn, $fn);
864 } elsif ($!{ENOENT}) { # non-fatal, we can just do it slowly
865 warn "open($fn): $!\n" if !$self->{-initial_clone};
868 die "open($fn): $!\n";
872 sub multi_inbox ($$$) {
873 my ($self, $path, $m) = @_;
874 my $incl = $self->{lei}->{opt}->{include};
875 my $excl = $self->{lei}->{opt}->{exclude};
877 # assuming everything not v2 is v1, for now
878 my @v1 = sort grep(!m!.+/git/[0-9]+\.git\z!, keys %$m);
879 my @v2_epochs = sort grep(m!.+/git/[0-9]+\.git\z!, keys %$m);
883 m!\A(/.+)/git/[0-9]+\.git\z! or die "BUG: $_";
884 push @{$v2->{$1}}, $_;
886 my $n = scalar(keys %$v2) + scalar(@v1);
887 my @orig = defined($incl // $excl) ? (keys %$v2, @v1) : ();
889 my $re = '(?:'.join('\\z|', map {
890 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
892 my @gone = delete @$v2{grep(!/$re/, keys %$v2)};
893 delete @$m{map { @$_ } @gone} and $self->{chg}->{manifest} = 1;
894 delete @$m{grep(!/$re/, @v1)} and $self->{chg}->{manifest} = 1;
895 @v1 = grep(/$re/, @v1);
898 my $re = '(?:'.join('\\z|', map {
899 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
901 my @gone = delete @$v2{grep(/$re/, keys %$v2)};
902 delete @$m{map { @$_ } @gone} and $self->{chg}->{manifest} = 1;
903 delete @$m{grep(/$re/, @v1)} and $self->{chg}->{manifest} = 1;
904 @v1 = grep(!/$re/, @v1);
906 my $ret; # { v1 => [ ... ], v2 => { "/$inbox_name" => [ epochs ] }}
907 $ret->{v1} = \@v1 if @v1;
908 $ret->{v2} = $v2 if keys %$v2;
909 $ret //= @orig ? "Nothing to clone, available repositories:\n\t".
910 join("\n\t", sort @orig)
911 : "Nothing available to clone\n";
914 # PSGI mount prefixes and manifest.js.gz prefixes don't always align...
916 until (grep(m!\A\Q$$path\E/git/[0-9]+\.git\z!,
917 @v2_epochs) == @v2_epochs) {
918 $$path =~ s!\A(/[^/]+)/!/! or last;
922 while (!defined($m->{$$path}) && $$path =~ s!\A(/[^/]+)/!/!) {
926 ($path_pfx, $n, $ret);
932 $TODO = \'BUG on further use';
933 my $end = PublicInbox::OnDestroy->new($$, \&fgrp_fetch_all, $self);
935 my $nodep = delete $todo->{''};
937 # do not download unwanted deps
938 my $any_want = delete $self->{any_want};
939 my @unwanted = grep { !$any_want->{$_} } keys %$todo;
940 my @nodep = delete(@$todo{@unwanted});
941 push(@$nodep, @$_) for @nodep;
943 # handle no-dependency repos, first
946 return if !keep_going($self);
949 # resolve references, deepest, first:
950 while (scalar keys %$todo) {
951 for my $x (keys %$todo) {
953 # resolve multi-level references
954 while ($m && defined($nxt = $m->{$x}->{reference})) {
955 exists($todo->{$nxt}) or last;
956 die <<EOM if ++$nr > 1000;
957 E: dependency loop detected (`$x' => `$nxt')
961 my $y = delete $todo->{$x} // next; # already done
964 return if !keep_going($self);
966 last; # restart %$todo iteration
970 # $end->DESTROY will call fgrp_fetch_all once all references
971 # in $LIVE are gone, and do_reap will eventually drain $LIVE
976 sub dump_manifest ($$) {
978 # write the smaller manifest if epochs were skipped so
979 # users won't have to delete manifest if they +w an
980 # epoch they no longer want to skip
981 my $json = PublicInbox::Config->json->encode($m);
982 my $mtime = (stat($ft))[9];
983 seek($ft, SEEK_SET, 0) or die "seek($ft): $!";
984 truncate($ft, 0) or die "truncate($ft): $!";
985 gzip(\$json => $ft) or die "gzip($ft): $GzipError";
986 $ft->flush or die "flush($ft): $!";
987 utime($mtime, $mtime, "$ft") or die "utime(..., $ft): $!";
990 # FIXME: this gets confused by single inbox instance w/ global manifest.js.gz
993 my $uri = URI->new($self->{src});
994 my $lei = $self->{lei};
995 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
996 $self->{-torsocks} //= $curl->torsocks($lei, $uri) or return;
997 my $path = $uri->path;
998 chop($path) eq '/' or die "BUG: $uri not canonicalized";
999 $uri->path($path . '/manifest.js.gz');
1000 my $manifest = $self->{-manifest} // "$self->{dst}/manifest.js.gz";
1001 my %opt = (UNLINK => 1, SUFFIX => '.tmp', TMPDIR => 1);
1002 if (!$self->{dry_run} && $manifest =~ m!\A(.+?)/[^/]+\z! and -d $1) {
1003 $opt{DIR} = $1; # allows fast rename(2) w/o EXDEV
1004 delete $opt{TMPDIR};
1006 my $ft = File::Temp->new(TEMPLATE => '.manifest-XXXX', %opt);
1007 my $cmd = $curl->for_uri($lei, $uri, qw(-R -o), $ft->filename);
1008 push(@$cmd, '-z', $manifest) if -f $manifest;
1009 my $mf_url = "$uri";
1010 %opt = map { $_ => $lei->{$_} } (0..2);
1011 my $cerr = run_reap($lei, $cmd, \%opt);
1013 return try_scrape($self) if ($cerr >> 8) == 22; # 404 missing
1014 return $lei->child_error($cerr, "@$cmd failed");
1017 # bail out if curl -z/--timecond hit 304 Not Modified, $ft will be empty
1018 return $lei->qerr("# $manifest unchanged") if -f $manifest && !-s $ft;
1020 my $m = eval { decode_manifest($ft, $ft, $uri) };
1023 return try_scrape($self);
1025 local $self->{chg} = {};
1026 local $self->{-local_manifest} = load_current_manifest($self);
1027 my ($path_pfx, $n, $multi) = multi_inbox($self, \$path, $m);
1028 return $lei->child_error(1, $multi) if !ref($multi);
1029 my $v2 = delete $multi->{v2};
1031 for my $name (sort keys %$v2) {
1032 my $epochs = delete $v2->{$name};
1033 my %v2_epochs = map {
1034 $uri->path($n > 1 ? $path_pfx.$path.$_
1036 my ($e) = ("$uri" =~ m!/([0-9]+)\.git\z!);
1037 $e // die "no [0-9]+\.git in `$uri'";
1038 $e => [ $uri->clone, $_ ];
1040 ("$uri" =~ m!\A(.+/)git/[0-9]+\.git\z!) or
1041 die "BUG: `$uri' !~ m!/git/[0-9]+.git!";
1042 local $self->{cur_src} = $1;
1043 local $self->{cur_dst} = $self->{dst};
1044 if ($n > 1 && $uri->path =~ m!\A\Q$path_pfx$path\E/(.+)/
1045 git/[0-9]+\.git\z!x) {
1046 $self->{cur_dst} .= "/$1";
1048 index($self->{cur_dst}, "\n") >= 0 and die <<EOM;
1049 E: `$self->{cur_dst}' must not contain newline
1051 clone_v2_prep($self, \%v2_epochs, $m);
1052 return if !keep_going($self);
1055 if (my $v1 = delete $multi->{v1}) {
1056 my $p = $path_pfx.$path;
1057 chop($p) if substr($p, -1, 1) eq '/';
1059 for my $name (@$v1) {
1060 my $task = bless { %$self }, __PACKAGE__;
1061 $task->{-ent} = $m->{$name} //
1062 die("BUG: no `$name' in manifest");
1063 $task->{cur_src} = "$uri";
1064 $task->{cur_dst} = $task->{dst};
1065 $task->{-key} = $name;
1067 $task->{cur_dst} .= $name;
1068 $task->{cur_src} .= $name;
1070 index($task->{cur_dst}, "\n") >= 0 and die <<EOM;
1071 E: `$task->{cur_dst}' must not contain newline
1073 $task->{cur_src} .= '/';
1074 my $dep = $task->{-ent}->{reference} // '';
1075 push @{$TODO->{$dep}}, $task; # for clone_all
1076 $self->{any_want}->{$name} = 1;
1079 delete local $lei->{opt}->{epoch} if defined($v2);
1080 clone_all($self, $m);
1081 return if $self->{dry_run} || !keep_going($self);
1083 # set by clone_v2_prep/-I/--exclude
1084 my $mis = delete $self->{chg}->{fp_mismatch};
1086 my $t = (stat($ft))[9];
1087 $t = strftime('%F %k:%M:%S %z', localtime($t));
1089 W: Fingerprints for the following repositories do not match
1091 W: These repositories may have updated since $t:
1093 warn "\t", $_, "\n" for @$mis;
1094 warn <<EOM if !$self->{lei}->{opt}->{prune};
1095 W: The above fingerprints may never match without --prune
1098 dump_manifest($m => $ft) if delete($self->{chg}->{manifest}) || $mis;
1099 my $bad = delete $self->{chg}->{badlink};
1100 warn(<<EOM, map { ("\t", $_, "\n") } @$bad) if $bad;
1101 W: The following exist and have not been converted to symlinks
1103 ft_rename($ft, $manifest, 0666);
1106 sub start_clone_url {
1108 return try_manifest($self) if $self->{src} =~ m!\Ahttps?://!;
1109 die "TODO: non-HTTP/HTTPS clone of $self->{src} not supported, yet";
1112 sub do_mirror { # via wq_io_do or public-inbox-clone
1114 my $lei = $self->{lei};
1115 $self->{dry_run} = 1 if $lei->{opt}->{'dry-run'};
1116 umask($lei->{client_umask}) if defined $lei->{client_umask};
1117 $self->{-initial_clone} = 1 if !-d $self->{dst};
1119 my $ic = $lei->{opt}->{'inbox-config'} //= 'always';
1120 $ic =~ /\A(?:v1|v2|always|never)\z/s or die <<"";
1121 --inbox-config must be one of `always', `v2', `v1', or `never'
1123 # we support --objstore= and --manifest= with '' (empty string)
1124 for my $default (qw(objstore manifest.js.gz)) {
1125 my ($k) = (split(/\./, $default))[0];
1126 my $v = $lei->{opt}->{$k} // next;
1127 $v = $default if $v eq '';
1128 $v = "$self->{dst}/$v" if $v !~ m!\A\.{0,2}/!;
1129 $self->{"-$k"} = $v;
1133 local $FGRP_TODO = {};
1134 my $iv = $lei->{opt}->{'inbox-version'} //
1135 return start_clone_url($self);
1136 return clone_v1($self) if $iv == 1;
1137 die "bad --inbox-version=$iv\n" if $iv != 2;
1138 die <<EOM if $self->{src} !~ m!://!;
1139 cloning local v2 inboxes not supported
1141 try_scrape($self, 1);
1143 $lei->fail($@) if $@;
1147 my ($cls, $lei, $src, $dst) = @_;
1148 my $self = bless { src => $src, dst => $dst }, $cls;
1149 $lei->request_umask;
1150 my ($op_c, $ops) = $lei->workers_start($self, 1);
1151 $lei->{wq1} = $self;
1152 $self->wq_io_do('do_mirror', []);
1154 $lei->wait_wq_events($op_c, $ops);
1157 sub ipc_atfork_child {
1159 $self->{lei}->_lei_atfork_child;
1160 $self->SUPER::ipc_atfork_child;
1163 sub write_makefile {
1164 my ($dir, $ibx_ver) = @_;
1165 my $f = "$dir/Makefile";
1166 if (sysopen my $fh, $f, O_CREAT|O_EXCL|O_WRONLY) {
1167 print $fh <<EOM or die "print($f) $!";
1168 # This is a v$ibx_ver public-inbox, see the public-inbox-v$ibx_ver-format(5)
1169 # manpage for more information on the format. This Makefile is
1170 # intended as a familiar wrapper for users unfamiliar with
1171 # public-inbox-* commands.
1173 # See the respective manpages for public-inbox-fetch(1),
1174 # public-inbox-index(1), etc for more information on
1175 # some of the commands used by this Makefile.
1177 # This Makefile will not be modified nor read by public-inbox,
1178 # so you may edit it freely with your own convenience targets
1179 # and notes. public-inbox-fetch will recreate it if removed.
1181 print $fh <<'EOM' or die "print($f): $!";
1182 # the default target:
1184 @echo Common targets:
1185 @echo ' make fetch - fetch from remote git repostorie(s)'
1186 @echo ' make update - fetch and update index '
1188 @echo Rarely needed targets:
1189 @echo ' make reindex - may be needed for new features/bugfixes'
1190 @echo ' make compact - rewrite Xapian storage to save space'
1191 @echo ' make index - initial index after clone
1196 @if ! public-inbox-fetch --exit-code; \
1199 test $$c -eq 127 && exit 0; \
1201 elif test -f msgmap.sqlite3 || test -f public-inbox/msgmap.sqlite3; \
1203 public-inbox-index; \
1205 echo 'public-inbox index not initialized'; \
1206 echo 'see public-inbox-index(1) man page'; \
1211 public-inbox-index --reindex
1213 public-inbox-compact
1215 .PHONY : help fetch update index reindex compact
1217 close $fh or die "close($f): $!";
1219 die "open($f): $!" unless $!{EEXIST};