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);
23 our $LIVE; # pid => callback
25 sub _wq_done_wait { # dwaitpid callback (via wq_eof)
27 my ($mrr, $lei) = @$arg;
28 my $f = "$mrr->{dst}/mirror.done";
30 $lei->child_error($?);
31 } elsif (!$mrr->{dry_run} && !unlink($f)) {
32 warn("unlink($f): $!\n") unless $!{ENOENT};
34 if (!$mrr->{dry_run} && $lei->{cmd} ne 'public-inbox-clone') {
35 require PublicInbox::LeiAddExternal;
36 PublicInbox::LeiAddExternal::_finish_add_external(
39 $lei->qerr("# mirrored $mrr->{src} => $mrr->{dst}");
44 # for old installations without manifest.js.gz
46 my ($self, $fallback_manifest) = @_;
47 my $uri = URI->new($self->{src});
48 my $lei = $self->{lei};
49 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
50 my $cmd = $curl->for_uri($lei, $uri, qw(-f --compressed));
51 my $opt = { 0 => $lei->{0}, 2 => $lei->{2} };
52 my $fh = popen_rd($cmd, undef, $opt);
53 my $html = do { local $/; <$fh> } // die "read(curl $uri): $!";
54 close($fh) or return $lei->child_error($?, "@$cmd failed");
56 # we grep with URL below, we don't want Subject/From headers
57 # making us clone random URLs. This assumes remote instances
58 # prior to public-inbox 1.7.0
59 # 5b96edcb1e0d8252 (www: move mirror instructions to /text/, 2021-08-28)
60 my @html = split(/<hr>/, $html);
61 my @urls = ($html[-1] =~ m!\bgit clone --mirror ([a-z\+]+://\S+)!g);
62 if (!@urls && $fallback_manifest) {
64 W: failed to extract URLs from $uri, trying manifest.js.gz...
66 return start_clone_url($self);
68 my $url = $uri->as_string;
69 chop($url) eq '/' or die "BUG: $uri not canonicalized";
71 # since this is for old instances w/o manifest.js.gz, try v1 first
72 return clone_v1($self) if grep(m!\A\Q$url\E/*\z!, @urls);
73 if (my @v2_urls = grep(m!\A\Q$url\E/[0-9]+\z!, @urls)) {
75 my ($n) = (m!/([0-9]+)\z!);
76 $n => [ URI->new($_), '' ]
78 clone_v2_prep($self, \%v2_epochs);
79 delete local $lei->{opt}->{epoch};
84 # filter out common URLs served by WWW (e.g /$MSGID/T/)
85 if (@urls && $url =~ s!/+[^/]+\@[^/]+/.*\z!! &&
86 grep(m!\A\Q$url\E/*\z!, @urls)) {
88 E: confused by scraping <$uri>, did you mean <$url>?
92 E: confused by scraping <$uri>, got ambiguous results:
95 die "E: scraping <$uri> revealed nothing\n";
101 $opt->{$_} = $lei->{$_} for (0..2);
102 # we support "-c $key=$val" for arbitrary git config options
103 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
104 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
105 push @cmd, qw(clone --mirror);
106 push @cmd, '-q' if $lei->{opt}->{quiet} ||
107 ($lei->{opt}->{jobs} // 1) > 1;
108 push @cmd, '-v' if $lei->{opt}->{verbose};
109 # XXX any other options to support?
110 # --reference is tricky with multiple epochs, but handled
111 # automatically if using manifest.js.gz
115 sub ft_rename ($$$;$) {
116 my ($ft, $dst, $open_mode, $fh) = @_;
117 my @st = stat($fh // $dst);
118 my $mode = @st ? ($st[2] & 07777) : ($open_mode & ~umask);
119 chmod($mode, $ft) or croak "E: chmod($ft): $!";
121 File::Copy::mv($ft->filename, $dst) or croak "E: mv($ft => $dst): $!";
122 $ft->unlink_on_destroy(0);
126 my ($self, $jobs) = @_;
127 $jobs //= $self->{-jobs} //= $self->{lei}->{opt}->{jobs} // 1;
128 $jobs = 1 if $jobs < 1;
129 while (keys(%$LIVE) >= $jobs) {
130 my $pid = waitpid(-1, 0) // die "waitpid(-1): $!";
131 if (my $x = delete $LIVE->{$pid}) {
135 warn "reaped unknown PID=$pid ($?)\n";
140 sub _get_txt_start { # non-fatal
141 my ($self, $endpoint, $fini) = @_;
142 my $uri = URI->new($self->{cur_src} // $self->{src});
143 my $lei = $self->{lei};
144 my $path = $uri->path;
145 chop($path) eq '/' or die "BUG: $uri not canonicalized";
146 $uri->path("$path/$endpoint");
147 my $f = (split(m!/!, $endpoint))[-1];
148 my $ft = File::Temp->new(TEMPLATE => "$f-XXXX", TMPDIR => 1);
149 my $opt = { 0 => $lei->{0}, 1 => $lei->{1}, 2 => $lei->{2} };
150 my $cmd = $self->{curl}->for_uri($lei, $uri, qw(-f --compressed -R -o),
153 $lei->qerr("# @$cmd");
154 return if $self->{dry_run};
155 $self->{"-get_txt.$endpoint"} = [ $ft, $cmd, $uri ];
156 $LIVE->{spawn($cmd, undef, $opt)} =
157 [ \&_get_txt_done, $self, $endpoint, $fini ];
160 sub _get_txt_done { # returns true on error (non-fatal), undef on success
161 my ($self, $endpoint) = @_;
162 my ($fh, $cmd, $uri) = @{delete $self->{"-get_txt.$endpoint"}};
164 $? = 0; # don't influence normal lei exit
165 return warn("$uri missing\n") if ($cerr >> 8) == 22;
166 return warn("# @$cmd failed (non-fatal)\n") if $cerr;
167 seek($fh, SEEK_SET, 0) or die "seek: $!";
168 $self->{"mtime.$endpoint"} = (stat($fh))[9];
170 $self->{"txt.$endpoint"} = <$fh>;
174 sub _write_inbox_config {
176 my $buf = delete($self->{'txt._/text/config/raw'}) // return;
177 my $dst = $self->{cur_dst} // $self->{dst};
178 my $f = "$dst/inbox.config.example";
179 open my $fh, '>', $f or die "open($f): $!";
180 print $fh $buf or die "print: $!";
181 chmod(0444 & ~umask, $fh) or die "chmod($f): $!";
182 my $mtime = delete $self->{'mtime._/text/config/raw'};
183 $fh->flush or die "flush($f): $!";
184 if (defined $mtime) {
185 utime($mtime, $mtime, $fh) or die "utime($f): $!";
187 my $cfg = PublicInbox::Config->git_config_dump($f, $self->{lei}->{2});
188 my $ibx = $self->{ibx} = {};
189 for my $sec (grep(/\Apublicinbox\./, @{$cfg->{-section_order}})) {
190 for (qw(address newsgroup nntpmirror)) {
191 $ibx->{$_} = $cfg->{"$sec.$_"};
196 sub set_description ($) {
198 my $dst = $self->{cur_dst} // $self->{dst};
199 my $f = "$dst/description";
200 open my $fh, '+>>', $f or die "open($f): $!";
201 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
202 my $d = do { local $/; <$fh> } // die "read($f): $!";
203 chomp(my $orig = $d);
204 while (defined($d) && ($d =~ m!^\(\$INBOX_DIR/description missing\)! ||
205 $d =~ /^Unnamed repository/ || $d !~ /\S/)) {
206 $d = delete($self->{'txt.description'});
208 $d //= 'mirror of '.($self->{cur_src} // $self->{src});
210 return if $d eq $orig;
211 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
212 truncate($fh, 0) or die "truncate($f): $!";
213 print $fh $d, "\n" or die "print($f): $!";
214 close $fh or die "close($f): $!";
217 sub index_cloned_inbox {
218 my ($self, $iv) = @_;
219 my $lei = $self->{lei};
221 # n.b. public-inbox-clone works w/o (SQLite || Xapian)
222 # lei is useless without Xapian + SQLite
223 if ($lei->{cmd} ne 'public-inbox-clone') {
224 require PublicInbox::InboxWritable;
225 require PublicInbox::Admin;
226 my $ibx = delete($self->{ibx}) // {
227 address => [ 'lei@example.com' ],
230 $ibx->{inboxdir} = $self->{cur_dst} // $self->{dst};
231 PublicInbox::Inbox->new($ibx);
232 PublicInbox::InboxWritable->new($ibx);
234 for my $sw ($lei->index_opt) {
235 my ($k) = ($sw =~ /\A([\w-]+)/);
236 $opt->{$k} = $lei->{opt}->{$k};
238 # force synchronous dwaitpid for v2:
239 local $PublicInbox::DS::in_loop = 0;
240 my $cfg = PublicInbox::Config->new(undef, $lei->{2});
241 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
242 local %ENV = (%ENV, %$env) if $env;
243 PublicInbox::Admin::progress_prepare($opt, $lei->{2});
244 PublicInbox::Admin::index_inbox($ibx, undef, $opt);
246 return if defined $self->{cur_dst}; # one of many repos to clone
247 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
251 my ($lei, $cmd, $opt) = @_;
252 $lei->qerr("# @$cmd");
253 waitpid(spawn($cmd, undef, $opt), 0) // die "waitpid: $!";
255 $? = 0; # don't let it influence normal exit
260 my ($self, $cmd, $opt, $fini) = @_;
262 $self->{lei}->qerr("# @$cmd");
263 return if $self->{dry_run};
264 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_cmd, $self, $cmd, $fini ]
267 sub fetch_args ($$) {
268 my ($lei, $opt) = @_;
269 my @cmd; # (git --git-dir=...) to be added by caller
270 $opt->{$_} = $lei->{$_} for (0..2);
271 # we support "-c $key=$val" for arbitrary git config options
272 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
273 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
275 push @cmd, '-q' if $lei->{opt}->{quiet} ||
276 ($lei->{opt}->{jobs} // 1) > 1;
277 push @cmd, '-v' if $lei->{opt}->{verbose};
281 sub upr { # feed `git update-ref --stdin -z' verbosely
282 my ($lei, $w, $op, $ref, $oid) = @_;
283 $lei->qerr("# $op $ref $oid") if $lei->{opt}->{verbose};
284 print $w "$op $ref\0$oid\0" or die "print(w): $!";
289 my $srcfh = delete $fgrp->{srcfh} or return;
290 my $dstfh = delete $fgrp->{dstfh} or return;
291 seek($srcfh, SEEK_SET, 0) or die "seek(src): $!";
292 seek($dstfh, SEEK_SET, 0) or die "seek(dst): $!";
293 my %src = map { chomp; split(/\0/) } (<$srcfh>);
295 my %dst = map { chomp; split(/\0/) } (<$dstfh>);
297 pipe(my ($r, $w)) or die "pipe: $!";
298 my $cmd = [ 'git', "--git-dir=$fgrp->{cur_dst}",
299 qw(update-ref --stdin -z) ];
300 my $lei = $fgrp->{lei};
301 $lei->qerr("# @$cmd");
302 my $opt = { 0 => $r, 1 => $lei->{1}, 2 => $lei->{2} };
303 my $pid = spawn($cmd, undef, $opt);
304 close $r or die "close(r): $!";
305 for my $ref (keys %dst) {
306 my $new = delete $src{$ref};
307 my $old = $dst{$ref};
309 upr($lei, $w, 'update', $ref, $new) if $new ne $old;
311 upr($lei, $w, 'delete', $ref, $old);
314 while (my ($ref, $oid) = each %src) {
315 upr($lei, $w, 'create', $ref, $oid);
317 close($w) or warn "E: close(update-ref --stdin): $! (need git 1.8.5+)\n";
318 $LIVE->{$pid} = [ \&reap_cmd, $fgrp, $cmd ];
323 my ($self, $git_dir) = @_;
325 my $cmd = [ 'git', "--git-dir=$git_dir", qw(pack-refs --all --prune) ];
326 $self->{lei}->qerr("# @$cmd");
327 my $opt = { 1 => $self->{lei}->{1}, 2 => $self->{lei}->{2} };
328 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_cmd, $self, $cmd ];
333 return if $fgrp->{dry_run} || !$LIVE;
334 my $rn = $fgrp->{-remote};
335 my %opt = map { $_ => $fgrp->{lei}->{$_} } (0..2);
336 pack_refs($fgrp, $fgrp->{-osdir}); # objstore refs always packed
338 my $update_ref = PublicInbox::OnDestroy->new($$, \&fgrp_update, $fgrp);
340 my $src = [ 'git', "--git-dir=$fgrp->{-osdir}", 'for-each-ref',
341 "--format=refs/%(refname:lstrip=3)%00%(objectname)",
342 "refs/remotes/$rn/" ];
344 open($fgrp->{srcfh}, '+>', undef) or die "open(src): $!";
345 $fgrp->{lei}->qerr("# @$src >SRC");
346 my $pid = spawn($src, undef, { %opt, 1 => $fgrp->{srcfh} });
347 $LIVE->{$pid} = [ \&reap_cmd, $fgrp, $src, $update_ref ];
349 my $dst = [ 'git', "--git-dir=$fgrp->{cur_dst}", 'for-each-ref',
350 '--format=%(refname)%00%(objectname)' ];
352 open($fgrp->{dstfh}, '+>', undef) or die "open(dst): $!";
353 $fgrp->{lei}->qerr("# @$dst >DST");
354 $pid = spawn($dst, undef, { %opt, 1 => $fgrp->{dstfh} });
355 $LIVE->{$pid} = [ \&reap_cmd, $fgrp, $dst, $update_ref ];
359 my ($fgrp, $fini) = @_;
360 my $cmd = [ @{$fgrp->{-torsocks}}, 'git', "--git-dir=$fgrp->{-osdir}",
361 fetch_args($fgrp->{lei}, my $opt = {}), '--no-tags',
363 $fgrp->{-fini} = $fini;
365 $fgrp->{lei}->qerr("# @$cmd");
366 return if $fgrp->{dry_run};
367 my $fgrp_fini = PublicInbox::OnDestroy->new($$, \&fgrp_fetched, $fgrp);
368 my $pid = spawn($cmd, undef, $opt);
369 $LIVE->{$pid} = [ \&reap_cmd, $fgrp, $cmd, $fgrp_fini ];
372 # keep this idempotent for future use by public-inbox-fetch
374 my ($self, $uri) = @_;
375 $self->{-ent} // return;
376 my $os = $self->{-objstore} // return;
377 my $fg = $self->{-ent}->{forkgroup} // return;
378 my $dir = "$os/$fg.git";
379 my @cmd = ('git', "--git-dir=$dir", 'config');
380 my $opt = +{ map { $_ => $self->{lei}->{$_} } (0..2) };
382 PublicInbox::Import::init_bare($dir);
383 for ('repack.useDeltaIslands=true',
384 'pack.island=refs/remotes/([^/]+)/') {
385 run_die([@cmd, split(/=/, $_, 2)], undef, $opt);
388 my $key = $self->{-key} // die 'BUG: no -key';
389 my ($bn) = ($key =~ m{/([a-z0-9_,;=!\+\{\}\|][^/]*)(?:\.git)?\z}i);
390 my $rn = "$bn-".substr(sha256_hex($key), 0, 16);
391 # --no-tags is required to avoid conflicts
392 for ("url=$uri", "fetch=+refs/*:refs/remotes/$rn/*",
393 'tagopt=--no-tags') {
394 my @kv = split(/=/, $_, 2);
395 $kv[0] = "remote.$rn.$kv[0]";
396 run_die([@cmd, @kv], undef, $opt);
398 $self->{-do_pack_refs} = 1; # likely coderepo
399 if (!-d $self->{cur_dst}) {
400 my $alt = File::Spec->rel2abs("$dir/objects");
401 PublicInbox::Import::init_bare($self->{cur_dst});
402 my $o = "$self->{cur_dst}/objects";
403 my $f = "$o/info/alternates";
404 my $l = File::Spec->abs2rel($alt, File::Spec->rel2abs($o));
405 open my $fh, '+>>', $f or die "open($f): $!";
406 seek($fh, SEEK_SET, 0) or die "seek($f): $!";
407 chomp(my @cur = <$fh>);
408 if (!grep(/\A\Q$l\E\z/, @cur)) {
409 say $fh $l or die "say($f): $!";
411 close $fh or die "close($f): $!";
412 $f = "$self->{cur_dst}/config";
413 open $fh, '+>>', $f or die "open:($f): $!";
414 print $fh <<EOM or die "print($f): $!";
415 ; rely on the "$rn" remote in the
416 ; $fg fork group for fetches
417 ; only uncomment the following iff you detach from fork groups
420 ; fetch = +refs/*:refs/*
423 close $fh or die "close($f): $!";
425 bless { %$self, -osdir => $dir, -remote => $rn }, __PACKAGE__;
429 my ($self, $nohang) = @_;
430 my $lei = $self->{lei};
431 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
432 my $uri = URI->new($self->{cur_src} // $self->{src});
433 defined($lei->{opt}->{epoch}) and
434 die "$uri is a v1 inbox, --epoch is not supported\n";
435 $self->{-torsocks} //= $curl->torsocks($lei, $uri) or return;
436 my $dst = $self->{cur_dst} // $self->{dst};
437 my $fini = PublicInbox::OnDestroy->new($$, \&v1_done, $self);
438 my $cmd = [ @{$self->{-torsocks}}, clone_cmd($lei, my $opt = {}),
440 my $fgrp = forkgroup_prep($self, $uri);
441 if (!defined($fgrp) && defined($self->{-ent})) {
442 if (defined(my $ref = $self->{-ent}->{reference})) {
443 -e "$self->{dst}$ref" and
444 push @$cmd, '--reference', "$self->{dst}$ref";
447 $fgrp ? fgrp_fetch($fgrp, $fini) :
448 start_clone($self, $cmd, $opt, $fini);
450 if (!$self->{-is_epoch} && $lei->{opt}->{'inbox-config'} =~
451 /\A(?:always|v1)\z/s) {
452 _get_txt_start($self, '_/text/config/raw', $fini);
455 my $d = $self->{-ent} ? $self->{-ent}->{description} : undef;
456 $self->{'txt.description'} = $d if defined $d;
457 (!defined($d) && !$nohang) and
458 _get_txt_start($self, 'description', $fini);
460 $nohang or do_reap($self, 1); # for non-manifest clone
463 sub parse_epochs ($$) {
464 my ($opt_epochs, $v2_epochs) = @_; # $epochs "LOW..HIGH"
465 $opt_epochs // return; # undef => all epochs
466 my ($lo, $dotdot, $hi, @extra) = split(/(\.\.)/, $opt_epochs);
467 undef($lo) if ($lo // '') eq '';
468 my $re = qr/\A~?[0-9]+\z/;
469 if (@extra || (($lo // '0') !~ $re) ||
470 (($hi // '0') !~ $re) ||
471 !(grep(defined, $lo, $hi))) {
473 --epoch=$opt_epochs not in the form of `LOW..HIGH', `LOW..', nor `..HIGH'
476 my @n = sort { $a <=> $b } keys %$v2_epochs;
477 for (grep(defined, $lo, $hi)) {
480 "`$_' exceeds maximum available epoch ($n[-1])\n";
482 "`$_' is lower than minimum available epoch ($n[0])\n";
483 } elsif (/\A~([0-9]+)/) {
485 $n[$off] // die "`$_' is out of range\n";
487 } else { die "`$_' not understood\n" }
489 defined($lo) && defined($hi) && $lo > $hi and die
490 "low value (`$lo') exceeds high (`$hi')\n";
491 $lo //= $n[0] if $dotdot;
492 $hi //= $n[-1] if $dotdot;
496 if (defined $v2_epochs->{$_}) {
500 "# epoch $_ is not available (non-fatal, $lo..$hi)\n";
506 sub init_placeholder ($$$) {
507 my ($src, $edst, $ent) = @_;
508 PublicInbox::Import::init_bare($edst);
509 my $f = "$edst/config";
510 open my $fh, '>>', $f or die "open($f): $!";
511 print $fh <<EOM or die "print($f): $!";
514 fetch = +refs/*:refs/*
517 ; This git epoch was created read-only and "public-inbox-fetch"
518 ; will not fetch updates for it unless write permission is added.
519 ; Hint: chmod +w $edst
521 if (defined($ent->{owner})) {
522 print $fh <<EOM or die "print($f): $!";
524 owner = $ent->{owner}
527 close $fh or die "close($f): $!";
528 my %map = (head => 'HEAD', description => undef);
529 while (my ($key, $fn) = each %map) {
530 my $val = $ent->{$key} // next;
533 open $fh, '>', $fn or die "open($fn): $!";
534 print $fh $val, "\n" or die "print($fn): $!";
535 close $fh or die "close($fn): $!";
539 sub reap_cmd { # async, called via SIGCHLD
540 my ($self, $cmd) = @_;
542 $? = 0; # don't let it influence normal exit
544 kill('TERM', keys %$LIVE);
545 $self->{lei}->child_error($cerr, "@$cmd failed");
549 sub v1_done { # called via OnDestroy
551 return if $self->{dry_run} || !$LIVE;
552 _write_inbox_config($self);
553 my $dst = $self->{cur_dst} // $self->{dst};
554 if (defined(my $o = $self->{-ent} ? $self->{-ent}->{owner} : undef)) {
555 run_die([qw(git config -f), "$dst/config", 'gitweb.owner', $o]);
557 my $o = "$dst/objects";
558 if (open(my $fh, '<', my $fn = "$o/info/alternates")) {;
559 my $base = File::Spec->rel2abs($o);
564 $_ = File::Spec->abs2rel($_, $base);
565 $ft //= File::Temp->new(TEMPLATE => '.XXXX',
569 print $ft @l or die "print($ft): $!";
570 $ft->flush or die "flush($ft): $!";
571 ft_rename($ft, $fn, 0666, $fh);
574 pack_refs($self, $dst) if delete $self->{-do_pack_refs};
575 eval { set_description($self) };
577 return if ($self->{-is_epoch} ||
578 $self->{lei}->{opt}->{'inbox-config'} ne 'always');
579 write_makefile($dst, 1);
580 index_cloned_inbox($self, 1);
583 sub v2_done { # called via OnDestroy
585 return if $self->{dry_run} || !$LIVE;
586 my $dst = $self->{cur_dst} // $self->{dst};
587 require PublicInbox::Lock;
588 my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
589 my $lck = $lk->lock_for_scope($$);
590 _write_inbox_config($self);
591 require PublicInbox::MultiGit;
592 my $mg = PublicInbox::MultiGit->new($dst, 'all.git', 'git');
593 $mg->fill_alternates;
594 for my $i ($mg->git_epochs) { $mg->epoch_cfg_set($i) }
595 for my $edst (@{delete($self->{-read_only}) // []}) {
596 my @st = stat($edst) or die "stat($edst): $!";
597 chmod($st[2] & 0555, $edst) or die "chmod(a-w, $edst): $!";
599 write_makefile($dst, 2);
601 eval { set_description($self) };
603 index_cloned_inbox($self, 2);
606 sub clone_v2_prep ($$;$) {
607 my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
608 my $lei = $self->{lei};
609 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
610 my $first_uri = (map { $_->[0] } values %$v2_epochs)[0];
611 $self->{-torsocks} //= $curl->torsocks($lei, $first_uri) or return;
612 my $dst = $self->{cur_dst} // $self->{dst};
613 my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
614 my $task = $m ? bless { %$self }, __PACKAGE__ : $self;
615 delete $task->{todo}; # $self->{todo} still exists
617 my $fini = PublicInbox::OnDestroy->new($$, \&v2_done, $task);
618 for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
619 my ($uri, $key) = @{$v2_epochs->{$nr}};
620 my $src = $uri->as_string;
622 $src =~ m!/([0-9]+)(?:\.git)?\z! or die <<"";
623 failed to extract epoch number from $src
625 $1 + 0 == $nr or die "BUG: <$uri> miskeyed $1 != $nr";
626 $edst .= "/git/$nr.git";
630 die("BUG: `$key' not in manifest.js.gz");
631 if (defined(my $d = $ent->{description})) {
632 $d =~ s/ \[epoch [0-9]+\]\z//s;
636 if (!$want || $want->{$nr}) {
637 my $etask = bless { %$task, -key => $key }, __PACKAGE__;
638 $etask->{-ent} = $ent; # may have {reference}
639 $etask->{cur_src} = $src;
640 $etask->{cur_dst} = $edst;
641 $etask->{-is_epoch} = $fini;
642 my $ref = $ent->{reference} // '';
643 push @{$self->{todo}->{$ref}}, $etask;
644 $self->{any_want}->{$key} = 1;
645 } else { # create a placeholder so users only need to chmod +w
646 init_placeholder($src, $edst, $ent);
647 push @{$task->{-read_only}}, $edst;
651 # filter out the epochs we skipped
652 $self->{-culled_manifest} = 1 if $m && delete(@$m{@skip});
654 (!$self->{dry_run} && !-d $dst) and File::Path::mkpath($dst);
656 $lei->{opt}->{'inbox-config'} =~ /\A(?:always|v2)\z/s and
657 _get_txt_start($task, '_/text/config/raw', $fini);
659 defined($desc) ? ($task->{'txt.description'} = $desc) :
660 _get_txt_start($task, 'description', $fini);
663 sub decode_manifest ($$$) {
664 my ($fh, $fn, $uri) = @_;
666 my $gz = do { local $/; <$fh> } // die "slurp($fn): $!";
667 gunzip(\$gz => \$js, MultiStream => 1) or
668 die "gunzip($uri): $GunzipError\n";
669 my $m = eval { PublicInbox::Config->json->decode($js) };
670 die "$uri: error decoding `$js': $@\n" if $@;
671 ref($m) eq 'HASH' or die "$uri unknown type: ".ref($m);
675 sub multi_inbox ($$$) {
676 my ($self, $path, $m) = @_;
677 my $incl = $self->{lei}->{opt}->{include};
678 my $excl = $self->{lei}->{opt}->{exclude};
680 # assuming everything not v2 is v1, for now
681 my @v1 = sort grep(!m!.+/git/[0-9]+\.git\z!, keys %$m);
682 my @v2_epochs = sort grep(m!.+/git/[0-9]+\.git\z!, keys %$m);
686 m!\A(/.+)/git/[0-9]+\.git\z! or die "BUG: $_";
687 push @{$v2->{$1}}, $_;
689 my $n = scalar(keys %$v2) + scalar(@v1);
690 my @orig = defined($incl // $excl) ? (keys %$v2, @v1) : ();
692 my $re = '(?:'.join('\\z|', map {
693 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
695 my @gone = delete @$v2{grep(!/$re/, keys %$v2)};
696 delete @$m{map { @$_ } @gone} and $self->{-culled_manifest} = 1;
697 delete @$m{grep(!/$re/, @v1)} and $self->{-culled_manifest} = 1;
698 @v1 = grep(/$re/, @v1);
701 my $re = '(?:'.join('\\z|', map {
702 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
704 my @gone = delete @$v2{grep(/$re/, keys %$v2)};
705 delete @$m{map { @$_ } @gone} and $self->{-culled_manifest} = 1;
706 delete @$m{grep(/$re/, @v1)} and $self->{-culled_manifest} = 1;
707 @v1 = grep(!/$re/, @v1);
709 my $ret; # { v1 => [ ... ], v2 => { "/$inbox_name" => [ epochs ] }}
710 $ret->{v1} = \@v1 if @v1;
711 $ret->{v2} = $v2 if keys %$v2;
712 $ret //= @orig ? "Nothing to clone, available repositories:\n\t".
713 join("\n\t", sort @orig)
714 : "Nothing available to clone\n";
717 # PSGI mount prefixes and manifest.js.gz prefixes don't always align...
719 until (grep(m!\A\Q$$path\E/git/[0-9]+\.git\z!,
720 @v2_epochs) == @v2_epochs) {
721 $$path =~ s!\A(/[^/]+)/!/! or last;
725 while (!defined($m->{$$path}) && $$path =~ s!\A(/[^/]+)/!/!) {
729 ($path_pfx, $n, $ret);
734 my $todo = delete $self->{todo};
735 my $nodep = delete $todo->{''};
737 # do not download unwanted deps
738 my $any_want = delete $self->{any_want};
739 my @unwanted = grep { !$any_want->{$_} } keys %$todo;
740 my @nodep = delete(@$todo{@unwanted});
741 push(@$nodep, @$_) for @nodep;
743 # handle no-dependency repos, first
746 return if $self->{lei}->{child_error};
748 # resolve references, deepest, first:
749 while (scalar keys %$todo) {
750 for my $x (keys %$todo) {
752 # resolve multi-level references
753 while ($m && defined($nxt = $m->{$x}->{reference})) {
754 exists($todo->{$nxt}) or last;
755 die <<EOM if ++$nr > 1000;
756 E: dependency loop detected (`$x' => `$nxt')
760 my $y = delete $todo->{$x} // next; # already done
763 return if $self->{lei}->{child_error};
765 last; # restart %$todo iteration
771 sub dump_manifest ($$) {
773 # write the smaller manifest if epochs were skipped so
774 # users won't have to delete manifest if they +w an
775 # epoch they no longer want to skip
776 my $json = PublicInbox::Config->json->encode($m);
777 my $mtime = (stat($ft))[9];
778 seek($ft, SEEK_SET, 0) or die "seek($ft): $!";
779 truncate($ft, 0) or die "truncate($ft): $!";
780 gzip(\$json => $ft) or die "gzip($ft): $GzipError";
781 $ft->flush or die "flush($ft): $!";
782 utime($mtime, $mtime, "$ft") or die "utime(..., $ft): $!";
785 # FIXME: this gets confused by single inbox instance w/ global manifest.js.gz
788 my $uri = URI->new($self->{src});
789 my $lei = $self->{lei};
790 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
791 my $path = $uri->path;
792 chop($path) eq '/' or die "BUG: $uri not canonicalized";
793 $uri->path($path . '/manifest.js.gz');
794 my $ft = File::Temp->new(TEMPLATE => '.manifest-XXXX',
795 UNLINK => 1, TMPDIR => 1, SUFFIX => '.tmp');
796 my $cmd = $curl->for_uri($lei, $uri, qw(-f -R -o), $ft->filename);
797 my %opt = map { $_ => $lei->{$_} } (0..2);
798 my $cerr = run_reap($lei, $cmd, \%opt);
800 return try_scrape($self) if ($cerr >> 8) == 22; # 404 missing
801 return $lei->child_error($cerr, "@$cmd failed");
803 my $m = eval { decode_manifest($ft, $ft, $uri) };
806 return try_scrape($self);
808 my ($path_pfx, $n, $multi) = multi_inbox($self, \$path, $m);
809 return $lei->child_error(1, $multi) if !ref($multi);
810 my $v2 = delete $multi->{v2};
811 local $self->{todo} = {};
813 for my $name (sort keys %$v2) {
814 my $epochs = delete $v2->{$name};
815 my %v2_epochs = map {
816 $uri->path($n > 1 ? $path_pfx.$path.$_
818 my ($e) = ("$uri" =~ m!/([0-9]+)\.git\z!);
819 $e // die "no [0-9]+\.git in `$uri'";
820 $e => [ $uri->clone, $_ ];
822 ("$uri" =~ m!\A(.+/)git/[0-9]+\.git\z!) or
823 die "BUG: `$uri' !~ m!/git/[0-9]+.git!";
824 local $self->{cur_src} = $1;
825 local $self->{cur_dst} = $self->{dst};
826 if ($n > 1 && $uri->path =~ m!\A\Q$path_pfx$path\E/(.+)/
827 git/[0-9]+\.git\z!x) {
828 $self->{cur_dst} .= "/$1";
830 index($self->{cur_dst}, "\n") >= 0 and die <<EOM;
831 E: `$self->{cur_dst}' must not contain newline
833 clone_v2_prep($self, \%v2_epochs, $m);
834 return if $self->{lei}->{child_error};
837 if (my $v1 = delete $multi->{v1}) {
838 my $p = $path_pfx.$path;
839 chop($p) if substr($p, -1, 1) eq '/';
841 for my $name (@$v1) {
842 my $task = bless { %$self }, __PACKAGE__;
843 $task->{-ent} = $m->{$name} //
844 die("BUG: no `$name' in manifest");
845 $task->{cur_src} = "$uri";
846 $task->{cur_dst} = $task->{dst};
847 $task->{-key} = $name;
849 $task->{cur_dst} .= $name;
850 $task->{cur_src} .= $name;
852 index($task->{cur_dst}, "\n") >= 0 and die <<EOM;
853 E: `$task->{cur_dst}' must not contain newline
855 $task->{cur_src} .= '/';
856 my $dep = $task->{-ent}->{reference} // '';
857 push @{$self->{todo}->{$dep}}, $task; # for clone_all
858 $self->{any_want}->{$name} = 1;
861 delete local $lei->{opt}->{epoch} if defined($v2);
862 clone_all($self, $m);
863 return if $self->{lei}->{child_error} || $self->{dry_run};
865 # set by clone_v2_prep/-I/--exclude
866 dump_manifest($m => $ft) if delete $self->{-culled_manifest};
867 ft_rename($ft, "$self->{dst}/manifest.js.gz", 0666);
868 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
871 sub start_clone_url {
873 return try_manifest($self) if $self->{src} =~ m!\Ahttps?://!;
874 die "TODO: non-HTTP/HTTPS clone of $self->{src} not supported, yet";
877 sub do_mirror { # via wq_io_do or public-inbox-clone
879 my $lei = $self->{lei};
880 umask($lei->{client_umask}) if defined $lei->{client_umask};
882 my $ic = $lei->{opt}->{'inbox-config'} //= 'always';
883 $ic =~ /\A(?:v1|v2|always|never)\z/s or die <<"";
884 --inbox-config must be one of `always', `v2', `v1', or `never'
886 if (defined(my $os = $lei->{opt}->{objstore})) {
887 $os = 'objstore' if $os eq ''; # --objstore w/o args
888 $os = "$self->{dst}/$os" if $os !~ m!\A/!;
889 $self->{-objstore} = $os;
892 my $iv = $lei->{opt}->{'inbox-version'} //
893 return start_clone_url($self);
894 return clone_v1($self) if $iv == 1;
895 die "bad --inbox-version=$iv\n" if $iv != 2;
896 die <<EOM if $self->{src} !~ m!://!;
897 cloning local v2 inboxes not supported
899 try_scrape($self, 1);
901 $lei->fail($@) if $@;
905 my ($cls, $lei, $src, $dst) = @_;
906 my $self = bless { src => $src, dst => $dst }, $cls;
908 my ($op_c, $ops) = $lei->workers_start($self, 1);
910 $self->wq_io_do('do_mirror', []);
912 $lei->wait_wq_events($op_c, $ops);
915 sub ipc_atfork_child {
917 $self->{lei}->_lei_atfork_child;
918 $self->SUPER::ipc_atfork_child;
922 my ($dir, $ibx_ver) = @_;
923 my $f = "$dir/Makefile";
924 if (sysopen my $fh, $f, O_CREAT|O_EXCL|O_WRONLY) {
925 print $fh <<EOM or die "print($f) $!";
926 # This is a v$ibx_ver public-inbox, see the public-inbox-v$ibx_ver-format(5)
927 # manpage for more information on the format. This Makefile is
928 # intended as a familiar wrapper for users unfamiliar with
929 # public-inbox-* commands.
931 # See the respective manpages for public-inbox-fetch(1),
932 # public-inbox-index(1), etc for more information on
933 # some of the commands used by this Makefile.
935 # This Makefile will not be modified nor read by public-inbox,
936 # so you may edit it freely with your own convenience targets
937 # and notes. public-inbox-fetch will recreate it if removed.
939 print $fh <<'EOM' or die "print($f): $!";
940 # the default target:
942 @echo Common targets:
943 @echo ' make fetch - fetch from remote git repostorie(s)'
944 @echo ' make update - fetch and update index '
946 @echo Rarely needed targets:
947 @echo ' make reindex - may be needed for new features/bugfixes'
948 @echo ' make compact - rewrite Xapian storage to save space'
949 @echo ' make index - initial index after clone
954 @if ! public-inbox-fetch --exit-code; \
957 test $$c -eq 127 && exit 0; \
959 elif test -f msgmap.sqlite3 || test -f public-inbox/msgmap.sqlite3; \
961 public-inbox-index; \
963 echo 'public-inbox index not initialized'; \
964 echo 'see public-inbox-index(1) man page'; \
969 public-inbox-index --reindex
973 .PHONY : help fetch update index reindex compact
975 close $fh or die "close($f): $!";
977 die "open($f): $!" unless $!{EEXIST};