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 IO::Uncompress::Gunzip qw(gunzip $GunzipError);
10 use IO::Compress::Gzip qw(gzip $GzipError);
11 use PublicInbox::Spawn qw(popen_rd spawn run_die);
15 use Fcntl qw(SEEK_SET O_CREAT O_EXCL O_WRONLY);
18 use PublicInbox::Config;
19 use PublicInbox::Inbox;
20 use PublicInbox::LeiCurl;
21 use PublicInbox::OnDestroy;
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 # calls _finish_add_external
36 $lei->lazy_cb('add-external', '_finish_'
37 )->($lei, $mrr->{dst});
39 $lei->qerr("# mirrored $mrr->{src} => $mrr->{dst}");
44 # for old installations without manifest.js.gz
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, '--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
58 my @html = split(/<hr>/, $html);
59 my @urls = ($html[-1] =~ m!\bgit clone --mirror ([a-z\+]+://\S+)!g);
60 my $url = $uri->as_string;
61 chop($url) eq '/' or die "BUG: $uri not canonicalized";
63 # since this is for old instances w/o manifest.js.gz, try v1 first
64 return clone_v1($self) if grep(m!\A\Q$url\E/*\z!, @urls);
65 if (my @v2_urls = grep(m!\A\Q$url\E/[0-9]+\z!, @urls)) {
67 my ($n) = (m!/([0-9]+)\z!);
68 $n => [ URI->new($_), '' ]
70 clone_v2($self, \%v2_epochs);
71 reap_live() while keys(%$LIVE);
75 # filter out common URLs served by WWW (e.g /$MSGID/T/)
76 if (@urls && $url =~ s!/+[^/]+\@[^/]+/.*\z!! &&
77 grep(m!\A\Q$url\E/*\z!, @urls)) {
79 E: confused by scraping <$uri>, did you mean <$url>?
83 E: confused by scraping <$uri>, got ambiguous results:
86 die "E: scraping <$uri> revealed nothing\n";
92 $opt->{$_} = $lei->{$_} for (0..2);
93 # we support "-c $key=$val" for arbitrary git config options
94 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
95 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
96 push @cmd, qw(clone --mirror);
97 push @cmd, '-q' if $lei->{opt}->{quiet} ||
98 ($lei->{opt}->{jobs} // 1) > 1;
99 push @cmd, '-v' if $lei->{opt}->{verbose};
100 # XXX any other options to support?
101 # --reference is tricky with multiple epochs, but handled
102 # automatically if using manifest.js.gz
106 sub ft_rename ($$$) {
107 my ($ft, $dst, $open_mode) = @_;
108 my $fn = $ft->filename;
110 my $mode = @st ? ($st[2] & 07777) : ($open_mode & ~umask);
111 chmod($mode, $ft) or croak "E: chmod $fn: $!";
113 File::Copy::mv($fn, $dst) or croak "E: mv($fn => $ft): $!";
114 $ft->unlink_on_destroy(0);
117 sub _get_txt_start { # non-fatal
118 my ($self, $endpoint, $fini) = @_;
119 my $uri = URI->new($self->{cur_src} // $self->{src});
120 my $lei = $self->{lei};
121 my $path = $uri->path;
122 chop($path) eq '/' or die "BUG: $uri not canonicalized";
123 $uri->path("$path/$endpoint");
124 my $f = (split(m!/!, $endpoint))[-1];
125 my $ft = File::Temp->new(TEMPLATE => "$f-XXXX", TMPDIR => 1);
126 my $opt = { 0 => $lei->{0}, 1 => $lei->{1}, 2 => $lei->{2} };
127 my $cmd = $self->{curl}->for_uri($lei, $uri, qw(--compressed -R -o),
129 my $jobs = $lei->{opt}->{jobs} // 1;
130 reap_live() while keys(%$LIVE) >= $jobs;
131 $lei->qerr("# @$cmd");
132 return if $self->{dry_run};
133 $self->{"-get_txt.$endpoint"} = [ $ft, $cmd, $uri ];
134 $LIVE->{spawn($cmd, undef, $opt)} =
135 [ \&_get_txt_done, $self, $endpoint, $fini ];
138 sub _get_txt_done { # returns true on error (non-fatal), undef on success
139 my ($self, $endpoint) = @_;
140 my ($fh, $cmd, $uri) = @{delete $self->{"-get_txt.$endpoint"}};
142 $? = 0; # don't influence normal lei exit
143 return warn("$uri missing\n") if ($cerr >> 8) == 22;
144 return warn("# @$cmd failed (non-fatal)\n") if $cerr;
145 seek($fh, SEEK_SET, 0) or die "seek: $!";
146 $self->{"mtime.$endpoint"} = (stat($fh))[9];
148 $self->{"txt.$endpoint"} = <$fh>;
152 sub _write_inbox_config {
154 my $buf = delete($self->{'txt._/text/config/raw'}) // return;
155 my $dst = $self->{cur_dst} // $self->{dst};
156 my $f = "$dst/inbox.config.example";
157 open my $fh, '>', $f or die "open($f): $!";
158 print $fh $buf or die "print: $!";
159 chmod(0444 & ~umask, $fh) or die "chmod($f): $!";
160 my $mtime = delete $self->{'mtime._/text/config/raw'};
161 $fh->flush or die "flush($f): $!";
162 if (defined $mtime) {
163 utime($mtime, $mtime, $fh) or die "utime($f): $!";
165 my $cfg = PublicInbox::Config->git_config_dump($f, $self->{lei}->{2});
166 my $ibx = $self->{ibx} = {};
167 for my $sec (grep(/\Apublicinbox\./, @{$cfg->{-section_order}})) {
168 for (qw(address newsgroup nntpmirror)) {
169 $ibx->{$_} = $cfg->{"$sec.$_"};
174 sub set_description ($) {
176 my $dst = $self->{cur_dst} // $self->{dst};
177 my $f = "$dst/description";
178 open my $fh, '+>>', $f or die "open($f): $!";
179 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
180 my $d = do { local $/; <$fh> } // die "read($f): $!";
181 chomp(my $orig = $d);
182 while (defined($d) && ($d =~ m!^\(\$INBOX_DIR/description missing\)! ||
183 $d =~ /^Unnamed repository/ || $d !~ /\S/)) {
184 $d = delete($self->{'txt.description'});
186 $d //= 'mirror of '.($self->{cur_src} // $self->{src});
188 return if $d eq $orig;
189 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
190 truncate($fh, 0) or die "truncate($f): $!";
191 print $fh $d, "\n" or die "print($f): $!";
192 close $fh or die "close($f): $!";
195 sub index_cloned_inbox {
196 my ($self, $iv) = @_;
197 my $lei = $self->{lei};
198 eval { set_description($self) };
201 # n.b. public-inbox-clone works w/o (SQLite || Xapian)
202 # lei is useless without Xapian + SQLite
203 if ($lei->{cmd} ne 'public-inbox-clone') {
204 require PublicInbox::InboxWritable;
205 require PublicInbox::Admin;
206 my $ibx = delete($self->{ibx}) // {
207 address => [ 'lei@example.com' ],
210 $ibx->{inboxdir} = $self->{cur_dst} // $self->{dst};
211 PublicInbox::Inbox->new($ibx);
212 PublicInbox::InboxWritable->new($ibx);
214 for my $sw ($lei->index_opt) {
215 my ($k) = ($sw =~ /\A([\w-]+)/);
216 $opt->{$k} = $lei->{opt}->{$k};
218 # force synchronous dwaitpid for v2:
219 local $PublicInbox::DS::in_loop = 0;
220 my $cfg = PublicInbox::Config->new(undef, $lei->{2});
221 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
222 local %ENV = (%ENV, %$env) if $env;
223 PublicInbox::Admin::progress_prepare($opt, $lei->{2});
224 PublicInbox::Admin::index_inbox($ibx, undef, $opt);
226 return if defined $self->{cur_dst};
227 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
231 my ($lei, $cmd, $opt) = @_;
232 $lei->qerr("# @$cmd");
233 waitpid(spawn($cmd, undef, $opt), 0) // die "waitpid: $!";
235 $? = 0; # don't let it influence normal exit
240 my ($self, $cmd, $opt, $fini) = @_;
241 my $jobs = $self->{lei}->{opt}->{jobs} // 1;
242 reap_live() while keys(%$LIVE) >= $jobs;
243 $self->{lei}->qerr("# @$cmd");
244 return if $self->{dry_run};
245 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_clone, $self, $cmd, $fini ]
249 my ($self, $nohang) = @_;
250 my $lei = $self->{lei};
251 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
252 my $uri = URI->new($self->{cur_src} // $self->{src});
253 defined($lei->{opt}->{epoch}) and
254 die "$uri is a v1 inbox, --epoch is not supported\n";
255 my $pfx = $curl->torsocks($lei, $uri) or return;
256 my $dst = $self->{cur_dst} // $self->{dst};
257 my $fini = PublicInbox::OnDestroy->new($$, \&v1_done, $self);
258 my $cmd = [ @$pfx, clone_cmd($lei, my $opt = {}), "$uri", $dst ];
259 my $ref = $self->{-ent} ? $self->{-ent}->{reference} : undef;
260 defined($ref) && -e "$self->{dst}$ref" and
261 push @$cmd, '--reference', "$self->{dst}$ref";
262 start_clone($self, $cmd, $opt, $fini);
264 if (!$self->{-is_epoch} && $lei->{opt}->{'inbox-config'} =~
265 /\A(?:always|v1)\z/s) {
266 _get_txt_start($self, '_/text/config/raw', $fini);
269 my $d = $self->{-ent} ? $self->{-ent}->{description} : undef;
270 $self->{'txt.description'} = $d if defined $d;
271 (!defined($d) && !$nohang) and
272 _get_txt_start($self, 'description', $fini);
274 reap_live() until ($nohang || !keys(%$LIVE)); # for non-manifest clone
277 sub parse_epochs ($$) {
278 my ($opt_epochs, $v2_epochs) = @_; # $epochs "LOW..HIGH"
279 $opt_epochs // return; # undef => all epochs
280 my ($lo, $dotdot, $hi, @extra) = split(/(\.\.)/, $opt_epochs);
281 undef($lo) if ($lo // '') eq '';
282 my $re = qr/\A~?[0-9]+\z/;
283 if (@extra || (($lo // '0') !~ $re) ||
284 (($hi // '0') !~ $re) ||
285 !(grep(defined, $lo, $hi))) {
287 --epoch=$opt_epochs not in the form of `LOW..HIGH', `LOW..', nor `..HIGH'
290 my @n = sort { $a <=> $b } keys %$v2_epochs;
291 for (grep(defined, $lo, $hi)) {
294 "`$_' exceeds maximum available epoch ($n[-1])\n";
296 "`$_' is lower than minimum available epoch ($n[0])\n";
297 } elsif (/\A~([0-9]+)/) {
299 $n[$off] // die "`$_' is out of range\n";
301 } else { die "`$_' not understood\n" }
303 defined($lo) && defined($hi) && $lo > $hi and die
304 "low value (`$lo') exceeds high (`$hi')\n";
305 $lo //= $n[0] if $dotdot;
306 $hi //= $n[-1] if $dotdot;
310 if (defined $v2_epochs->{$_}) {
314 "# epoch $_ is not available (non-fatal, $lo..$hi)\n";
320 sub init_placeholder ($$$) {
321 my ($src, $edst, $ent) = @_;
322 PublicInbox::Import::init_bare($edst);
323 my $f = "$edst/config";
324 open my $fh, '>>', $f or die "open($f): $!";
325 print $fh <<EOM or die "print($f): $!";
328 fetch = +refs/*:refs/*
331 ; This git epoch was created read-only and "public-inbox-fetch"
332 ; will not fetch updates for it unless write permission is added.
333 ; Hint: chmod +w $edst
335 if (defined($ent->{owner})) {
336 print $fh <<EOM or die "print($f): $!";
338 owner = $ent->{owner}
341 close $fh or die "close($f): $!";
342 my %map = (head => 'HEAD', description => undef);
343 while (my ($key, $fn) = each %map) {
344 my $val = $ent->{$key} // next;
347 open $fh, '>', $fn or die "open($fn): $!";
348 print $fh $val, "\n" or die "print($fn): $!";
349 close $fh or die "close($fn): $!";
353 sub reap_clone { # async, called via SIGCHLD
354 my ($self, $cmd) = @_;
356 $? = 0; # don't let it influence normal exit
358 kill('TERM', keys %$LIVE);
359 $self->{lei}->child_error($cerr, "@$cmd failed");
363 sub v1_done { # called via OnDestroy
365 return if $self->{dry_run} || !$LIVE;
366 _write_inbox_config($self);
367 my $dst = $self->{cur_dst} // $self->{dst};
368 if (defined(my $o = $self->{-ent} ? $self->{-ent}->{owner} : undef)) {
369 run_die([qw(git config -f), "$dst/config", 'gitweb.owner', $o]);
371 my $o = "$dst/objects";
372 if (open(my $fh, '<', "$o/info/alternates")) {
373 chomp(my @l = <$fh>);
374 for (@l) { $_ = File::Spec->abs2rel($_, $o)."\n" }
375 my $f = File::Temp->new(TEMPLATE => '.XXXX', DIR => "$o/info");
377 $f->flush or die "flush($f): $!";
378 rename($f->filename, "$o/info/alternates") or
379 die "rename($f, $o/info/alternates): $!";
380 $f->unlink_on_destroy(0);
382 return if $self->{-is_epoch};
383 write_makefile($dst, 1);
384 index_cloned_inbox($self, 1);
387 sub v2_done { # called via OnDestroy
389 return if $self->{dry_run} || !$LIVE;
390 my $dst = $self->{cur_dst} // $self->{dst};
391 require PublicInbox::Lock;
392 my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
393 my $lck = $lk->lock_for_scope($$);
394 _write_inbox_config($self);
395 require PublicInbox::MultiGit;
396 my $mg = PublicInbox::MultiGit->new($dst, 'all.git', 'git');
397 $mg->fill_alternates;
398 for my $i ($mg->git_epochs) { $mg->epoch_cfg_set($i) }
399 my $entries = delete($self->{-entv}) // [];
401 my ($edst, $ent) = splice(@$entries);
402 if (defined(my $o = $ent->{owner})) {
403 run_die [qw(git config -f), "$edst/config",
406 my $d = $ent->{description} // next;
407 my $fn = "$edst/description";
408 open my $fh, '>', $fn or die "open($fn): $!";
409 print $fh $d, "\n" or die "print($fn): $!";
410 close $fh or die "close($fn): $!";
412 for my $edst (@{delete($self->{-read_only}) // []}) {
413 my @st = stat($edst) or die "stat($edst): $!";
414 chmod($st[2] & 0555, $edst) or die "chmod(a-w, $edst): $!";
416 write_makefile($dst, 2);
418 index_cloned_inbox($self, 2);
422 my $pid = waitpid(-1, 0) // die "waitpid(-1): $!";
423 if (my $x = delete $LIVE->{$pid}) {
427 warn "reaped unknown PID=$pid ($?)\n";
431 sub clone_v2 ($$;$) {
432 my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
433 my $lei = $self->{lei};
434 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
435 my $first_uri = (map { $_->[0] } values %$v2_epochs)[0];
436 my $pfx = $curl->torsocks($lei, $first_uri) or return;
437 my $dst = $self->{cur_dst} // $self->{dst};
438 my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
439 my $task = $m ? bless { %$self }, __PACKAGE__ : $self;
440 delete $task->{todo}; # $self->{todo} still exists
441 my (@src_edst, @skip, $desc);
442 for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
443 my ($uri, $key) = @{$v2_epochs->{$nr}};
444 my $src = $uri->as_string;
446 $src =~ m!/([0-9]+)(?:\.git)?\z! or die <<"";
447 failed to extract epoch number from $src
449 $1 + 0 == $nr or die "BUG: <$uri> miskeyed $1 != $nr";
450 $edst .= "/git/$nr.git";
451 my $ent = $m->{$key} // die "BUG: `$key' not in manifest.js.gz";
452 if (defined(my $d = $ent->{description})) {
453 $d =~ s/ \[epoch [0-9]+\]\z//s;
456 if (!$want || $want->{$nr}) {
457 push @src_edst, $src, $edst;
458 push @{$task->{-entv}}, $edst, $ent;
459 $self->{any_want}->{$key} = 1;
460 } else { # create a placeholder so users only need to chmod +w
461 init_placeholder($src, $edst, $ent);
462 push @{$task->{-read_only}}, $edst;
466 # filter out the epochs we skipped
467 $self->{-culled_manifest} = 1 if delete(@$m{@skip});
469 (!$self->{dry_run} && !-d $dst) and File::Path::mkpath($dst);
471 my $fini = PublicInbox::OnDestroy->new($$, \&v2_done, $task);
473 $lei->{opt}->{'inbox-config'} =~ /\A(?:always|v2)\z/s and
474 _get_txt_start($task, '_/text/config/raw', $fini);
476 defined($desc) ? ($task->{'txt.description'} = $desc) :
477 _get_txt_start($task, 'description', $fini);
478 if (my $todo = $self->{todo}) { # manifest clone, deal with references
479 my $entv = delete $task->{-entv};
481 my ($edst, $ent) = splice(@$entv, 0, 2);
482 my $etask = bless { %$task }, __PACKAGE__;
483 $etask->{-ent} = $ent; # may have {reference}
484 $etask->{cur_src} = shift @src_edst //
485 die 'BUG: no cur_src';
486 $etask->{cur_dst} = shift @src_edst //
487 die 'BUG: no cur_dst';
488 $etask->{cur_dst} eq $edst or
489 die "BUG: `$etask->{cur_dst}' != `$edst'";
490 $etask->{-is_epoch} = $fini;
491 push @{$todo->{($ent->{reference} // '')}}, $etask;
494 my @cmd = clone_cmd($lei, my $opt = {});
495 while (@src_edst && !$lei->{child_error}) {
496 my $cmd = [ @$pfx, @cmd, splice(@src_edst, 0, 2) ];
497 start_clone($self, $cmd, $opt, $fini);
502 sub decode_manifest ($$$) {
503 my ($fh, $fn, $uri) = @_;
505 my $gz = do { local $/; <$fh> } // die "slurp($fn): $!";
506 gunzip(\$gz => \$js, MultiStream => 1) or
507 die "gunzip($uri): $GunzipError\n";
508 my $m = eval { PublicInbox::Config->json->decode($js) };
509 die "$uri: error decoding `$js': $@\n" if $@;
510 ref($m) eq 'HASH' or die "$uri unknown type: ".ref($m);
514 sub multi_inbox ($$$) {
515 my ($self, $path, $m) = @_;
516 my $incl = $self->{lei}->{opt}->{include};
517 my $excl = $self->{lei}->{opt}->{exclude};
519 # assuming everything not v2 is v1, for now
520 my @v1 = sort grep(!m!.+/git/[0-9]+\.git\z!, keys %$m);
521 my @v2_epochs = sort grep(m!.+/git/[0-9]+\.git\z!, keys %$m);
525 m!\A(/.+)/git/[0-9]+\.git\z! or die "BUG: $_";
526 push @{$v2->{$1}}, $_;
528 my $n = scalar(keys %$v2) + scalar(@v1);
529 my @orig = defined($incl // $excl) ? (keys %$v2, @v1) : ();
531 my $re = '(?:'.join('\\z|', map {
532 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
534 my @gone = delete @$v2{grep(!/$re/, keys %$v2)};
535 delete @$m{map { @$_ } @gone} and $self->{-culled_manifest} = 1;
536 delete @$m{grep(!/$re/, @v1)} and $self->{-culled_manifest} = 1;
537 @v1 = grep(/$re/, @v1);
540 my $re = '(?:'.join('\\z|', map {
541 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
543 my @gone = delete @$v2{grep(/$re/, keys %$v2)};
544 delete @$m{map { @$_ } @gone} and $self->{-culled_manifest} = 1;
545 delete @$m{grep(/$re/, @v1)} and $self->{-culled_manifest} = 1;
546 @v1 = grep(!/$re/, @v1);
548 my $ret; # { v1 => [ ... ], v2 => { "/$inbox_name" => [ epochs ] }}
549 $ret->{v1} = \@v1 if @v1;
550 $ret->{v2} = $v2 if keys %$v2;
551 $ret //= @orig ? "Nothing to clone, available repositories:\n\t".
552 join("\n\t", sort @orig)
553 : "Nothing available to clone\n";
556 # PSGI mount prefixes and manifest.js.gz prefixes don't always align...
558 until (grep(m!\A\Q$$path\E/git/[0-9]+\.git\z!,
559 @v2_epochs) == @v2_epochs) {
560 $$path =~ s!\A(/[^/]+)/!/! or last;
564 while (!defined($m->{$$path}) && $$path =~ s!\A(/[^/]+)/!/!) {
568 ($path_pfx, $n, $ret);
573 my $todo = delete $self->{todo};
574 my $nodep = delete $todo->{''};
576 # do not download unwanted deps
577 my $any_want = delete $self->{any_want};
578 my @unwanted = grep { !$any_want->{$_} } keys %$todo;
579 my @nodep = delete(@$todo{@unwanted});
580 push(@$nodep, @$_) for @nodep;
582 # handle no-dependency repos, first
585 return if $self->{lei}->{child_error};
587 # resolve references, deepest, first:
588 while (scalar keys %$todo) {
589 for my $x (keys %$todo) {
591 # resolve multi-level references
592 while (defined(my $nxt = $m->{$x}->{reference})) {
593 exists($todo->{$nxt}) or last;
594 die <<EOM if ++$nr > 1000;
595 E: dependency loop detected (`$x' => `$nxt')
599 my $y = delete $todo->{$x} // next; # already done
602 return if $self->{lei}->{child_error};
604 last; # restart %$todo iteration
609 # FIXME: this gets confused by single inbox instance w/ global manifest.js.gz
612 my $uri = URI->new($self->{src});
613 my $lei = $self->{lei};
614 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
615 my $path = $uri->path;
616 chop($path) eq '/' or die "BUG: $uri not canonicalized";
617 $uri->path($path . '/manifest.js.gz');
618 my $ft = File::Temp->new(TEMPLATE => '.manifest-XXXX',
619 UNLINK => 1, TMPDIR => 1, SUFFIX => '.tmp');
620 my $fn = $ft->filename;
621 my $cmd = $curl->for_uri($lei, $uri, '-R', '-o', $fn);
622 my %opt = map { $_ => $lei->{$_} } (0..2);
623 my $cerr = run_reap($lei, $cmd, \%opt);
626 return try_scrape($self) if ($cerr >> 8) == 22; # 404 missing
627 return $lei->child_error($cerr, "@$cmd failed");
629 my $m = eval { decode_manifest($ft, $fn, $uri) };
632 return try_scrape($self);
634 my ($path_pfx, $n, $multi) = multi_inbox($self, \$path, $m);
635 return $lei->child_error(1, $multi) if !ref($multi);
636 my $v2 = delete $multi->{v2};
637 local $self->{todo} = {};
639 for my $name (sort keys %$v2) {
640 my $epochs = delete $v2->{$name};
641 my %v2_epochs = map {
642 $uri->path($n > 1 ? $path_pfx.$path.$_
644 my ($e) = ("$uri" =~ m!/([0-9]+)\.git\z!);
645 $e // die "no [0-9]+\.git in `$uri'";
646 $e => [ $uri->clone, $_ ];
648 ("$uri" =~ m!\A(.+/)git/[0-9]+\.git\z!) or
649 die "BUG: `$uri' !~ m!/git/[0-9]+.git!";
650 local $self->{cur_src} = $1;
651 local $self->{cur_dst} = $self->{dst};
652 if ($n > 1 && $uri->path =~ m!\A\Q$path_pfx$path\E/(.+)/
653 git/[0-9]+\.git\z!x) {
654 $self->{cur_dst} .= "/$1";
656 index($self->{cur_dst}, "\n") >= 0 and die <<EOM;
657 E: `$self->{cur_dst}' must not contain newline
659 clone_v2($self, \%v2_epochs, $m);
660 return if $self->{lei}->{child_error};
663 if (my $v1 = delete $multi->{v1}) {
664 my $p = $path_pfx.$path;
665 chop($p) if substr($p, -1, 1) eq '/';
667 for my $name (@$v1) {
668 my $task = bless { %$self }, __PACKAGE__;
669 $task->{-ent} = $m->{$name} //
670 die("BUG: no `$name' in manifest");
671 $task->{cur_src} = "$uri";
672 $task->{cur_dst} = $task->{dst};
674 $task->{cur_dst} .= $name;
675 $task->{cur_src} .= $name;
677 index($task->{cur_dst}, "\n") >= 0 and die <<EOM;
678 E: `$task->{cur_dst}' must not contain newline
680 $task->{cur_src} .= '/';
681 my $dep = $task->{-ent}->{reference} // '';
682 push @{$self->{todo}->{$dep}}, $task;
683 $self->{any_want}->{$name} = 1;
686 delete local $lei->{opt}->{epoch} if defined($v2);
687 clone_all($self, $m);
688 reap_live() while keys(%$LIVE);
689 return if $self->{lei}->{child_error} || $self->{dry_run};
691 if (delete $self->{-culled_manifest}) { # set by clone_v2/-I/--exclude
692 # write the smaller manifest if epochs were skipped so
693 # users won't have to delete manifest if they +w an
694 # epoch they no longer want to skip
695 my $json = PublicInbox::Config->json->encode($m);
696 my $mtime = (stat($fn))[9];
697 gzip(\$json => $fn) or die "gzip: $GzipError";
698 utime($mtime, $mtime, $fn) or die "utime(..., $fn): $!";
700 ft_rename($ft, "$self->{dst}/manifest.js.gz", 0666);
701 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
704 sub start_clone_url {
706 return try_manifest($self) if $self->{src} =~ m!\Ahttps?://!;
707 die "TODO: non-HTTP/HTTPS clone of $self->{src} not supported, yet";
710 sub do_mirror { # via wq_io_do or public-inbox-clone
712 my $lei = $self->{lei};
713 umask($lei->{client_umask}) if defined $lei->{client_umask};
715 my $ic = $lei->{opt}->{'inbox-config'} //= 'always';
716 $ic =~ /\A(?:v1|v2|always|never)\z/s or die <<"";
717 --inbox-config must be one of `always', `v2', `v1', or `never'
719 my $iv = $lei->{opt}->{'inbox-version'};
722 return clone_v1($self) if $iv == 1;
723 return try_scrape($self) if $iv == 2;
724 die "bad --inbox-version=$iv\n";
726 return start_clone_url($self) if $self->{src} =~ m!://!;
727 die "TODO: cloning local directories not supported, yet";
729 $lei->fail($@) if $@;
733 my ($cls, $lei, $src, $dst) = @_;
734 my $self = bless { src => $src, dst => $dst }, $cls;
736 my ($op_c, $ops) = $lei->workers_start($self, 1);
738 $self->wq_io_do('do_mirror', []);
740 $lei->wait_wq_events($op_c, $ops);
743 sub ipc_atfork_child {
745 $self->{lei}->_lei_atfork_child;
746 $self->SUPER::ipc_atfork_child;
750 my ($dir, $ibx_ver) = @_;
751 my $f = "$dir/Makefile";
752 if (sysopen my $fh, $f, O_CREAT|O_EXCL|O_WRONLY) {
753 print $fh <<EOM or die "print($f) $!";
754 # This is a v$ibx_ver public-inbox, see the public-inbox-v$ibx_ver-format(5)
755 # manpage for more information on the format. This Makefile is
756 # intended as a familiar wrapper for users unfamiliar with
757 # public-inbox-* commands.
759 # See the respective manpages for public-inbox-fetch(1),
760 # public-inbox-index(1), etc for more information on
761 # some of the commands used by this Makefile.
763 # This Makefile will not be modified nor read by public-inbox,
764 # so you may edit it freely with your own convenience targets
765 # and notes. public-inbox-fetch will recreate it if removed.
767 print $fh <<'EOM' or die "print($f): $!";
768 # the default target:
770 @echo Common targets:
771 @echo ' make fetch - fetch from remote git repostorie(s)'
772 @echo ' make update - fetch and update index '
774 @echo Rarely needed targets:
775 @echo ' make reindex - may be needed for new features/bugfixes'
776 @echo ' make compact - rewrite Xapian storage to save space'
781 @if ! public-inbox-fetch --exit-code; \
784 test $$c -eq 127 && exit 0; \
786 elif test -f msgmap.sqlite3 || test -f public-inbox/msgmap.sqlite3; \
788 public-inbox-index; \
790 echo 'public-inbox index not initialized'; \
791 echo 'see public-inbox-index(1) man page'; \
794 public-inbox-index --reindex
798 .PHONY : help fetch update reindex compact
800 close $fh or die "close($f): $!";
802 die "open($f): $!" unless $!{EEXIST};