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;
22 our $LIVE; # pid => callback
24 sub _wq_done_wait { # dwaitpid callback (via wq_eof)
26 my ($mrr, $lei) = @$arg;
27 my $f = "$mrr->{dst}/mirror.done";
29 $lei->child_error($?);
30 } elsif (!$mrr->{dry_run} && !unlink($f)) {
31 warn("unlink($f): $!\n") unless $!{ENOENT};
33 if (!$mrr->{dry_run} && $lei->{cmd} ne 'public-inbox-clone') {
34 # calls _finish_add_external
35 $lei->lazy_cb('add-external', '_finish_'
36 )->($lei, $mrr->{dst});
38 $lei->qerr("# mirrored $mrr->{src} => $mrr->{dst}");
43 # for old installations without manifest.js.gz
45 my ($self, $fallback_manifest) = @_;
46 my $uri = URI->new($self->{src});
47 my $lei = $self->{lei};
48 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
49 my $cmd = $curl->for_uri($lei, $uri, qw(-f --compressed));
50 my $opt = { 0 => $lei->{0}, 2 => $lei->{2} };
51 my $fh = popen_rd($cmd, undef, $opt);
52 my $html = do { local $/; <$fh> } // die "read(curl $uri): $!";
53 close($fh) or return $lei->child_error($?, "@$cmd failed");
55 # we grep with URL below, we don't want Subject/From headers
56 # making us clone random URLs. This assumes remote instances
57 # prior to public-inbox 1.7.0
58 # 5b96edcb1e0d8252 (www: move mirror instructions to /text/, 2021-08-28)
59 my @html = split(/<hr>/, $html);
60 my @urls = ($html[-1] =~ m!\bgit clone --mirror ([a-z\+]+://\S+)!g);
61 if (!@urls && $fallback_manifest) {
63 W: failed to extract URLs from $uri, trying manifest.js.gz...
65 return start_clone_url($self);
67 my $url = $uri->as_string;
68 chop($url) eq '/' or die "BUG: $uri not canonicalized";
70 # since this is for old instances w/o manifest.js.gz, try v1 first
71 return clone_v1($self) if grep(m!\A\Q$url\E/*\z!, @urls);
72 if (my @v2_urls = grep(m!\A\Q$url\E/[0-9]+\z!, @urls)) {
74 my ($n) = (m!/([0-9]+)\z!);
75 $n => [ URI->new($_), '' ]
77 clone_v2_prep($self, \%v2_epochs);
78 delete local $lei->{opt}->{epoch};
83 # filter out common URLs served by WWW (e.g /$MSGID/T/)
84 if (@urls && $url =~ s!/+[^/]+\@[^/]+/.*\z!! &&
85 grep(m!\A\Q$url\E/*\z!, @urls)) {
87 E: confused by scraping <$uri>, did you mean <$url>?
91 E: confused by scraping <$uri>, got ambiguous results:
94 die "E: scraping <$uri> revealed nothing\n";
100 $opt->{$_} = $lei->{$_} for (0..2);
101 # we support "-c $key=$val" for arbitrary git config options
102 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
103 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
104 push @cmd, qw(clone --mirror);
105 push @cmd, '-q' if $lei->{opt}->{quiet} ||
106 ($lei->{opt}->{jobs} // 1) > 1;
107 push @cmd, '-v' if $lei->{opt}->{verbose};
108 # XXX any other options to support?
109 # --reference is tricky with multiple epochs, but handled
110 # automatically if using manifest.js.gz
114 sub ft_rename ($$$) {
115 my ($ft, $dst, $open_mode) = @_;
117 my $mode = @st ? ($st[2] & 07777) : ($open_mode & ~umask);
118 chmod($mode, $ft) or croak "E: chmod($ft): $!";
120 File::Copy::mv($ft->filename, $dst) or croak "E: mv($ft => $dst): $!";
121 $ft->unlink_on_destroy(0);
124 sub _get_txt_start { # non-fatal
125 my ($self, $endpoint, $fini) = @_;
126 my $uri = URI->new($self->{cur_src} // $self->{src});
127 my $lei = $self->{lei};
128 my $path = $uri->path;
129 chop($path) eq '/' or die "BUG: $uri not canonicalized";
130 $uri->path("$path/$endpoint");
131 my $f = (split(m!/!, $endpoint))[-1];
132 my $ft = File::Temp->new(TEMPLATE => "$f-XXXX", TMPDIR => 1);
133 my $opt = { 0 => $lei->{0}, 1 => $lei->{1}, 2 => $lei->{2} };
134 my $cmd = $self->{curl}->for_uri($lei, $uri, qw(-f --compressed -R -o),
136 my $jobs = $lei->{opt}->{jobs} // 1;
137 reap_live() while keys(%$LIVE) >= $jobs;
138 $lei->qerr("# @$cmd");
139 return if $self->{dry_run};
140 $self->{"-get_txt.$endpoint"} = [ $ft, $cmd, $uri ];
141 $LIVE->{spawn($cmd, undef, $opt)} =
142 [ \&_get_txt_done, $self, $endpoint, $fini ];
145 sub _get_txt_done { # returns true on error (non-fatal), undef on success
146 my ($self, $endpoint) = @_;
147 my ($fh, $cmd, $uri) = @{delete $self->{"-get_txt.$endpoint"}};
149 $? = 0; # don't influence normal lei exit
150 return warn("$uri missing\n") if ($cerr >> 8) == 22;
151 return warn("# @$cmd failed (non-fatal)\n") if $cerr;
152 seek($fh, SEEK_SET, 0) or die "seek: $!";
153 $self->{"mtime.$endpoint"} = (stat($fh))[9];
155 $self->{"txt.$endpoint"} = <$fh>;
159 sub _write_inbox_config {
161 my $buf = delete($self->{'txt._/text/config/raw'}) // return;
162 my $dst = $self->{cur_dst} // $self->{dst};
163 my $f = "$dst/inbox.config.example";
164 open my $fh, '>', $f or die "open($f): $!";
165 print $fh $buf or die "print: $!";
166 chmod(0444 & ~umask, $fh) or die "chmod($f): $!";
167 my $mtime = delete $self->{'mtime._/text/config/raw'};
168 $fh->flush or die "flush($f): $!";
169 if (defined $mtime) {
170 utime($mtime, $mtime, $fh) or die "utime($f): $!";
172 my $cfg = PublicInbox::Config->git_config_dump($f, $self->{lei}->{2});
173 my $ibx = $self->{ibx} = {};
174 for my $sec (grep(/\Apublicinbox\./, @{$cfg->{-section_order}})) {
175 for (qw(address newsgroup nntpmirror)) {
176 $ibx->{$_} = $cfg->{"$sec.$_"};
181 sub set_description ($) {
183 my $dst = $self->{cur_dst} // $self->{dst};
184 my $f = "$dst/description";
185 open my $fh, '+>>', $f or die "open($f): $!";
186 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
187 my $d = do { local $/; <$fh> } // die "read($f): $!";
188 chomp(my $orig = $d);
189 while (defined($d) && ($d =~ m!^\(\$INBOX_DIR/description missing\)! ||
190 $d =~ /^Unnamed repository/ || $d !~ /\S/)) {
191 $d = delete($self->{'txt.description'});
193 $d //= 'mirror of '.($self->{cur_src} // $self->{src});
195 return if $d eq $orig;
196 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
197 truncate($fh, 0) or die "truncate($f): $!";
198 print $fh $d, "\n" or die "print($f): $!";
199 close $fh or die "close($f): $!";
202 sub index_cloned_inbox {
203 my ($self, $iv) = @_;
204 my $lei = $self->{lei};
205 eval { set_description($self) };
208 # n.b. public-inbox-clone works w/o (SQLite || Xapian)
209 # lei is useless without Xapian + SQLite
210 if ($lei->{cmd} ne 'public-inbox-clone') {
211 require PublicInbox::InboxWritable;
212 require PublicInbox::Admin;
213 my $ibx = delete($self->{ibx}) // {
214 address => [ 'lei@example.com' ],
217 $ibx->{inboxdir} = $self->{cur_dst} // $self->{dst};
218 PublicInbox::Inbox->new($ibx);
219 PublicInbox::InboxWritable->new($ibx);
221 for my $sw ($lei->index_opt) {
222 my ($k) = ($sw =~ /\A([\w-]+)/);
223 $opt->{$k} = $lei->{opt}->{$k};
225 # force synchronous dwaitpid for v2:
226 local $PublicInbox::DS::in_loop = 0;
227 my $cfg = PublicInbox::Config->new(undef, $lei->{2});
228 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
229 local %ENV = (%ENV, %$env) if $env;
230 PublicInbox::Admin::progress_prepare($opt, $lei->{2});
231 PublicInbox::Admin::index_inbox($ibx, undef, $opt);
233 return if defined $self->{cur_dst}; # one of many repos to clone
234 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
238 my ($lei, $cmd, $opt) = @_;
239 $lei->qerr("# @$cmd");
240 waitpid(spawn($cmd, undef, $opt), 0) // die "waitpid: $!";
242 $? = 0; # don't let it influence normal exit
247 my ($self, $cmd, $opt, $fini) = @_;
248 my $jobs = $self->{lei}->{opt}->{jobs} // 1;
249 reap_live() while keys(%$LIVE) >= $jobs;
250 $self->{lei}->qerr("# @$cmd");
251 return if $self->{dry_run};
252 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_clone, $self, $cmd, $fini ]
256 my ($self, $nohang) = @_;
257 my $lei = $self->{lei};
258 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
259 my $uri = URI->new($self->{cur_src} // $self->{src});
260 defined($lei->{opt}->{epoch}) and
261 die "$uri is a v1 inbox, --epoch is not supported\n";
262 my $pfx = $curl->torsocks($lei, $uri) or return;
263 my $dst = $self->{cur_dst} // $self->{dst};
264 my $fini = PublicInbox::OnDestroy->new($$, \&v1_done, $self);
265 my $cmd = [ @$pfx, clone_cmd($lei, my $opt = {}), "$uri", $dst ];
266 my $ref = $self->{-ent} ? $self->{-ent}->{reference} : undef;
267 defined($ref) && -e "$self->{dst}$ref" and
268 push @$cmd, '--reference', "$self->{dst}$ref";
269 start_clone($self, $cmd, $opt, $fini);
271 if (!$self->{-is_epoch} && $lei->{opt}->{'inbox-config'} =~
272 /\A(?:always|v1)\z/s) {
273 _get_txt_start($self, '_/text/config/raw', $fini);
276 my $d = $self->{-ent} ? $self->{-ent}->{description} : undef;
277 $self->{'txt.description'} = $d if defined $d;
278 (!defined($d) && !$nohang) and
279 _get_txt_start($self, 'description', $fini);
281 reap_live() until ($nohang || !keys(%$LIVE)); # for non-manifest clone
284 sub parse_epochs ($$) {
285 my ($opt_epochs, $v2_epochs) = @_; # $epochs "LOW..HIGH"
286 $opt_epochs // return; # undef => all epochs
287 my ($lo, $dotdot, $hi, @extra) = split(/(\.\.)/, $opt_epochs);
288 undef($lo) if ($lo // '') eq '';
289 my $re = qr/\A~?[0-9]+\z/;
290 if (@extra || (($lo // '0') !~ $re) ||
291 (($hi // '0') !~ $re) ||
292 !(grep(defined, $lo, $hi))) {
294 --epoch=$opt_epochs not in the form of `LOW..HIGH', `LOW..', nor `..HIGH'
297 my @n = sort { $a <=> $b } keys %$v2_epochs;
298 for (grep(defined, $lo, $hi)) {
301 "`$_' exceeds maximum available epoch ($n[-1])\n";
303 "`$_' is lower than minimum available epoch ($n[0])\n";
304 } elsif (/\A~([0-9]+)/) {
306 $n[$off] // die "`$_' is out of range\n";
308 } else { die "`$_' not understood\n" }
310 defined($lo) && defined($hi) && $lo > $hi and die
311 "low value (`$lo') exceeds high (`$hi')\n";
312 $lo //= $n[0] if $dotdot;
313 $hi //= $n[-1] if $dotdot;
317 if (defined $v2_epochs->{$_}) {
321 "# epoch $_ is not available (non-fatal, $lo..$hi)\n";
327 sub init_placeholder ($$$) {
328 my ($src, $edst, $ent) = @_;
329 PublicInbox::Import::init_bare($edst);
330 my $f = "$edst/config";
331 open my $fh, '>>', $f or die "open($f): $!";
332 print $fh <<EOM or die "print($f): $!";
335 fetch = +refs/*:refs/*
338 ; This git epoch was created read-only and "public-inbox-fetch"
339 ; will not fetch updates for it unless write permission is added.
340 ; Hint: chmod +w $edst
342 if (defined($ent->{owner})) {
343 print $fh <<EOM or die "print($f): $!";
345 owner = $ent->{owner}
348 close $fh or die "close($f): $!";
349 my %map = (head => 'HEAD', description => undef);
350 while (my ($key, $fn) = each %map) {
351 my $val = $ent->{$key} // next;
354 open $fh, '>', $fn or die "open($fn): $!";
355 print $fh $val, "\n" or die "print($fn): $!";
356 close $fh or die "close($fn): $!";
360 sub reap_clone { # async, called via SIGCHLD
361 my ($self, $cmd) = @_;
363 $? = 0; # don't let it influence normal exit
365 kill('TERM', keys %$LIVE);
366 $self->{lei}->child_error($cerr, "@$cmd failed");
370 sub v1_done { # called via OnDestroy
372 return if $self->{dry_run} || !$LIVE;
373 _write_inbox_config($self);
374 my $dst = $self->{cur_dst} // $self->{dst};
375 if (defined(my $o = $self->{-ent} ? $self->{-ent}->{owner} : undef)) {
376 run_die([qw(git config -f), "$dst/config", 'gitweb.owner', $o]);
378 my $o = "$dst/objects";
379 if (open(my $fh, '<', "$o/info/alternates")) {
380 chomp(my @l = <$fh>);
381 for (@l) { $_ = File::Spec->abs2rel($_, $o)."\n" }
382 my $f = File::Temp->new(TEMPLATE => '.XXXX', DIR => "$o/info");
384 $f->flush or die "flush($f): $!";
385 rename($f->filename, "$o/info/alternates") or
386 die "rename($f, $o/info/alternates): $!";
387 $f->unlink_on_destroy(0);
389 return if $self->{-is_epoch};
390 write_makefile($dst, 1);
391 index_cloned_inbox($self, 1);
394 sub v2_done { # called via OnDestroy
396 return if $self->{dry_run} || !$LIVE;
397 my $dst = $self->{cur_dst} // $self->{dst};
398 require PublicInbox::Lock;
399 my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
400 my $lck = $lk->lock_for_scope($$);
401 _write_inbox_config($self);
402 require PublicInbox::MultiGit;
403 my $mg = PublicInbox::MultiGit->new($dst, 'all.git', 'git');
404 $mg->fill_alternates;
405 for my $i ($mg->git_epochs) { $mg->epoch_cfg_set($i) }
406 for my $edst (@{delete($self->{-read_only}) // []}) {
407 my @st = stat($edst) or die "stat($edst): $!";
408 chmod($st[2] & 0555, $edst) or die "chmod(a-w, $edst): $!";
410 write_makefile($dst, 2);
412 index_cloned_inbox($self, 2);
416 my $pid = waitpid(-1, 0) // die "waitpid(-1): $!";
417 if (my $x = delete $LIVE->{$pid}) {
421 warn "reaped unknown PID=$pid ($?)\n";
425 sub clone_v2_prep ($$;$) {
426 my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
427 my $lei = $self->{lei};
428 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
429 my $first_uri = (map { $_->[0] } values %$v2_epochs)[0];
430 my $pfx = $curl->torsocks($lei, $first_uri) or return;
431 my $dst = $self->{cur_dst} // $self->{dst};
432 my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
433 my $task = $m ? bless { %$self }, __PACKAGE__ : $self;
434 delete $task->{todo}; # $self->{todo} still exists
435 my (@src_edst, @skip, $desc, @entv);
436 for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
437 my ($uri, $key) = @{$v2_epochs->{$nr}};
438 my $src = $uri->as_string;
440 $src =~ m!/([0-9]+)(?:\.git)?\z! or die <<"";
441 failed to extract epoch number from $src
443 $1 + 0 == $nr or die "BUG: <$uri> miskeyed $1 != $nr";
444 $edst .= "/git/$nr.git";
448 die("BUG: `$key' not in manifest.js.gz");
449 if (defined(my $d = $ent->{description})) {
450 $d =~ s/ \[epoch [0-9]+\]\z//s;
454 if (!$want || $want->{$nr}) {
455 push @src_edst, $src, $edst;
456 push @entv, $edst, $ent;
457 $self->{any_want}->{$key} = 1;
458 } else { # create a placeholder so users only need to chmod +w
459 init_placeholder($src, $edst, $ent);
460 push @{$task->{-read_only}}, $edst;
464 # filter out the epochs we skipped
465 $self->{-culled_manifest} = 1 if $m && delete(@$m{@skip});
467 (!$self->{dry_run} && !-d $dst) and File::Path::mkpath($dst);
469 my $fini = PublicInbox::OnDestroy->new($$, \&v2_done, $task);
471 $lei->{opt}->{'inbox-config'} =~ /\A(?:always|v2)\z/s and
472 _get_txt_start($task, '_/text/config/raw', $fini);
474 defined($desc) ? ($task->{'txt.description'} = $desc) :
475 _get_txt_start($task, 'description', $fini);
477 my ($edst, $ent) = splice(@entv, 0, 2);
478 my $etask = bless { %$task }, __PACKAGE__;
479 $etask->{-ent} = $ent; # may have {reference}
480 $etask->{cur_src} = shift @src_edst // die 'BUG: no cur_src';
481 $etask->{cur_dst} = shift @src_edst // die 'BUG: no cur_dst';
482 $etask->{cur_dst} eq $edst or
483 die "BUG: `$etask->{cur_dst}' != `$edst'";
484 $etask->{-is_epoch} = $fini;
485 push @{$self->{todo}->{($ent->{reference} // '')}}, $etask;
489 sub decode_manifest ($$$) {
490 my ($fh, $fn, $uri) = @_;
492 my $gz = do { local $/; <$fh> } // die "slurp($fn): $!";
493 gunzip(\$gz => \$js, MultiStream => 1) or
494 die "gunzip($uri): $GunzipError\n";
495 my $m = eval { PublicInbox::Config->json->decode($js) };
496 die "$uri: error decoding `$js': $@\n" if $@;
497 ref($m) eq 'HASH' or die "$uri unknown type: ".ref($m);
501 sub multi_inbox ($$$) {
502 my ($self, $path, $m) = @_;
503 my $incl = $self->{lei}->{opt}->{include};
504 my $excl = $self->{lei}->{opt}->{exclude};
506 # assuming everything not v2 is v1, for now
507 my @v1 = sort grep(!m!.+/git/[0-9]+\.git\z!, keys %$m);
508 my @v2_epochs = sort grep(m!.+/git/[0-9]+\.git\z!, keys %$m);
512 m!\A(/.+)/git/[0-9]+\.git\z! or die "BUG: $_";
513 push @{$v2->{$1}}, $_;
515 my $n = scalar(keys %$v2) + scalar(@v1);
516 my @orig = defined($incl // $excl) ? (keys %$v2, @v1) : ();
518 my $re = '(?:'.join('\\z|', map {
519 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
521 my @gone = delete @$v2{grep(!/$re/, keys %$v2)};
522 delete @$m{map { @$_ } @gone} and $self->{-culled_manifest} = 1;
523 delete @$m{grep(!/$re/, @v1)} and $self->{-culled_manifest} = 1;
524 @v1 = grep(/$re/, @v1);
527 my $re = '(?:'.join('\\z|', map {
528 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
530 my @gone = delete @$v2{grep(/$re/, keys %$v2)};
531 delete @$m{map { @$_ } @gone} and $self->{-culled_manifest} = 1;
532 delete @$m{grep(/$re/, @v1)} and $self->{-culled_manifest} = 1;
533 @v1 = grep(!/$re/, @v1);
535 my $ret; # { v1 => [ ... ], v2 => { "/$inbox_name" => [ epochs ] }}
536 $ret->{v1} = \@v1 if @v1;
537 $ret->{v2} = $v2 if keys %$v2;
538 $ret //= @orig ? "Nothing to clone, available repositories:\n\t".
539 join("\n\t", sort @orig)
540 : "Nothing available to clone\n";
543 # PSGI mount prefixes and manifest.js.gz prefixes don't always align...
545 until (grep(m!\A\Q$$path\E/git/[0-9]+\.git\z!,
546 @v2_epochs) == @v2_epochs) {
547 $$path =~ s!\A(/[^/]+)/!/! or last;
551 while (!defined($m->{$$path}) && $$path =~ s!\A(/[^/]+)/!/!) {
555 ($path_pfx, $n, $ret);
560 my $todo = delete $self->{todo};
561 my $nodep = delete $todo->{''};
563 # do not download unwanted deps
564 my $any_want = delete $self->{any_want};
565 my @unwanted = grep { !$any_want->{$_} } keys %$todo;
566 my @nodep = delete(@$todo{@unwanted});
567 push(@$nodep, @$_) for @nodep;
569 # handle no-dependency repos, first
572 return if $self->{lei}->{child_error};
574 # resolve references, deepest, first:
575 while (scalar keys %$todo) {
576 for my $x (keys %$todo) {
578 # resolve multi-level references
579 while ($m && defined($nxt = $m->{$x}->{reference})) {
580 exists($todo->{$nxt}) or last;
581 die <<EOM if ++$nr > 1000;
582 E: dependency loop detected (`$x' => `$nxt')
586 my $y = delete $todo->{$x} // next; # already done
589 return if $self->{lei}->{child_error};
591 last; # restart %$todo iteration
594 reap_live() while keys(%$LIVE);
597 # FIXME: this gets confused by single inbox instance w/ global manifest.js.gz
600 my $uri = URI->new($self->{src});
601 my $lei = $self->{lei};
602 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
603 my $path = $uri->path;
604 chop($path) eq '/' or die "BUG: $uri not canonicalized";
605 $uri->path($path . '/manifest.js.gz');
606 my $ft = File::Temp->new(TEMPLATE => '.manifest-XXXX',
607 UNLINK => 1, TMPDIR => 1, SUFFIX => '.tmp');
608 my $cmd = $curl->for_uri($lei, $uri, qw(-f -R -o), $ft->filename);
609 my %opt = map { $_ => $lei->{$_} } (0..2);
610 my $cerr = run_reap($lei, $cmd, \%opt);
612 return try_scrape($self) if ($cerr >> 8) == 22; # 404 missing
613 return $lei->child_error($cerr, "@$cmd failed");
615 my $m = eval { decode_manifest($ft, $ft, $uri) };
618 return try_scrape($self);
620 my ($path_pfx, $n, $multi) = multi_inbox($self, \$path, $m);
621 return $lei->child_error(1, $multi) if !ref($multi);
622 my $v2 = delete $multi->{v2};
623 local $self->{todo} = {};
625 for my $name (sort keys %$v2) {
626 my $epochs = delete $v2->{$name};
627 my %v2_epochs = map {
628 $uri->path($n > 1 ? $path_pfx.$path.$_
630 my ($e) = ("$uri" =~ m!/([0-9]+)\.git\z!);
631 $e // die "no [0-9]+\.git in `$uri'";
632 $e => [ $uri->clone, $_ ];
634 ("$uri" =~ m!\A(.+/)git/[0-9]+\.git\z!) or
635 die "BUG: `$uri' !~ m!/git/[0-9]+.git!";
636 local $self->{cur_src} = $1;
637 local $self->{cur_dst} = $self->{dst};
638 if ($n > 1 && $uri->path =~ m!\A\Q$path_pfx$path\E/(.+)/
639 git/[0-9]+\.git\z!x) {
640 $self->{cur_dst} .= "/$1";
642 index($self->{cur_dst}, "\n") >= 0 and die <<EOM;
643 E: `$self->{cur_dst}' must not contain newline
645 clone_v2_prep($self, \%v2_epochs, $m);
646 return if $self->{lei}->{child_error};
649 if (my $v1 = delete $multi->{v1}) {
650 my $p = $path_pfx.$path;
651 chop($p) if substr($p, -1, 1) eq '/';
653 for my $name (@$v1) {
654 my $task = bless { %$self }, __PACKAGE__;
655 $task->{-ent} = $m->{$name} //
656 die("BUG: no `$name' in manifest");
657 $task->{cur_src} = "$uri";
658 $task->{cur_dst} = $task->{dst};
660 $task->{cur_dst} .= $name;
661 $task->{cur_src} .= $name;
663 index($task->{cur_dst}, "\n") >= 0 and die <<EOM;
664 E: `$task->{cur_dst}' must not contain newline
666 $task->{cur_src} .= '/';
667 my $dep = $task->{-ent}->{reference} // '';
668 push @{$self->{todo}->{$dep}}, $task; # for clone_all
669 $self->{any_want}->{$name} = 1;
672 delete local $lei->{opt}->{epoch} if defined($v2);
673 clone_all($self, $m);
674 return if $self->{lei}->{child_error} || $self->{dry_run};
676 # set by clone_v2_prep/-I/--exclude
677 if (delete $self->{-culled_manifest}) {
678 # write the smaller manifest if epochs were skipped so
679 # users won't have to delete manifest if they +w an
680 # epoch they no longer want to skip
681 my $json = PublicInbox::Config->json->encode($m);
682 my $mtime = (stat($ft))[9];
683 seek($ft, SEEK_SET, 0) or die "seek($ft): $!";
684 truncate($ft, 0) or die "truncate($ft): $!";
685 gzip(\$json => $ft) or die "gzip($ft): $GzipError";
686 $ft->flush or die "flush($ft): $!";
687 utime($mtime, $mtime, "$ft") or die "utime(..., $ft): $!";
689 ft_rename($ft, "$self->{dst}/manifest.js.gz", 0666);
690 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
693 sub start_clone_url {
695 return try_manifest($self) if $self->{src} =~ m!\Ahttps?://!;
696 die "TODO: non-HTTP/HTTPS clone of $self->{src} not supported, yet";
699 sub do_mirror { # via wq_io_do or public-inbox-clone
701 my $lei = $self->{lei};
702 umask($lei->{client_umask}) if defined $lei->{client_umask};
704 my $ic = $lei->{opt}->{'inbox-config'} //= 'always';
705 $ic =~ /\A(?:v1|v2|always|never)\z/s or die <<"";
706 --inbox-config must be one of `always', `v2', `v1', or `never'
709 my $iv = $lei->{opt}->{'inbox-version'} //
710 return start_clone_url($self);
711 return clone_v1($self) if $iv == 1;
712 die "bad --inbox-version=$iv\n" if $iv != 2;
713 die <<EOM if $self->{src} !~ m!://!;
714 cloning local v2 inboxes not supported
716 try_scrape($self, 1);
718 $lei->fail($@) if $@;
722 my ($cls, $lei, $src, $dst) = @_;
723 my $self = bless { src => $src, dst => $dst }, $cls;
725 my ($op_c, $ops) = $lei->workers_start($self, 1);
727 $self->wq_io_do('do_mirror', []);
729 $lei->wait_wq_events($op_c, $ops);
732 sub ipc_atfork_child {
734 $self->{lei}->_lei_atfork_child;
735 $self->SUPER::ipc_atfork_child;
739 my ($dir, $ibx_ver) = @_;
740 my $f = "$dir/Makefile";
741 if (sysopen my $fh, $f, O_CREAT|O_EXCL|O_WRONLY) {
742 print $fh <<EOM or die "print($f) $!";
743 # This is a v$ibx_ver public-inbox, see the public-inbox-v$ibx_ver-format(5)
744 # manpage for more information on the format. This Makefile is
745 # intended as a familiar wrapper for users unfamiliar with
746 # public-inbox-* commands.
748 # See the respective manpages for public-inbox-fetch(1),
749 # public-inbox-index(1), etc for more information on
750 # some of the commands used by this Makefile.
752 # This Makefile will not be modified nor read by public-inbox,
753 # so you may edit it freely with your own convenience targets
754 # and notes. public-inbox-fetch will recreate it if removed.
756 print $fh <<'EOM' or die "print($f): $!";
757 # the default target:
759 @echo Common targets:
760 @echo ' make fetch - fetch from remote git repostorie(s)'
761 @echo ' make update - fetch and update index '
763 @echo Rarely needed targets:
764 @echo ' make reindex - may be needed for new features/bugfixes'
765 @echo ' make compact - rewrite Xapian storage to save space'
766 @echo ' make index - initial index after clone
771 @if ! public-inbox-fetch --exit-code; \
774 test $$c -eq 127 && exit 0; \
776 elif test -f msgmap.sqlite3 || test -f public-inbox/msgmap.sqlite3; \
778 public-inbox-index; \
780 echo 'public-inbox index not initialized'; \
781 echo 'see public-inbox-index(1) man page'; \
786 public-inbox-index --reindex
790 .PHONY : help fetch update index reindex compact
792 close $fh or die "close($f): $!";
794 die "open($f): $!" unless $!{EEXIST};