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
24 my $update_ref_stdin = $ENV{GIT_CAN_UPDATE_REF_STDIN} // 1;
26 sub _wq_done_wait { # dwaitpid callback (via wq_eof)
28 my ($mrr, $lei) = @$arg;
29 my $f = "$mrr->{dst}/mirror.done";
31 $lei->child_error($?);
32 } elsif (!$mrr->{dry_run} && !unlink($f)) {
33 warn("unlink($f): $!\n") unless $!{ENOENT};
35 if (!$mrr->{dry_run} && $lei->{cmd} ne 'public-inbox-clone') {
36 require PublicInbox::LeiAddExternal;
37 PublicInbox::LeiAddExternal::_finish_add_external(
40 $lei->qerr("# mirrored $mrr->{src} => $mrr->{dst}");
45 # for old installations without manifest.js.gz
47 my ($self, $fallback_manifest) = @_;
48 my $uri = URI->new($self->{src});
49 my $lei = $self->{lei};
50 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
51 my $cmd = $curl->for_uri($lei, $uri, qw(-f --compressed));
52 my $opt = { 0 => $lei->{0}, 2 => $lei->{2} };
53 my $fh = popen_rd($cmd, undef, $opt);
54 my $html = do { local $/; <$fh> } // die "read(curl $uri): $!";
55 close($fh) or return $lei->child_error($?, "@$cmd failed");
57 # we grep with URL below, we don't want Subject/From headers
58 # making us clone random URLs. This assumes remote instances
59 # prior to public-inbox 1.7.0
60 # 5b96edcb1e0d8252 (www: move mirror instructions to /text/, 2021-08-28)
61 my @html = split(/<hr>/, $html);
62 my @urls = ($html[-1] =~ m!\bgit clone --mirror ([a-z\+]+://\S+)!g);
63 if (!@urls && $fallback_manifest) {
65 W: failed to extract URLs from $uri, trying manifest.js.gz...
67 return start_clone_url($self);
69 my $url = $uri->as_string;
70 chop($url) eq '/' or die "BUG: $uri not canonicalized";
72 # since this is for old instances w/o manifest.js.gz, try v1 first
73 return clone_v1($self) if grep(m!\A\Q$url\E/*\z!, @urls);
74 if (my @v2_urls = grep(m!\A\Q$url\E/[0-9]+\z!, @urls)) {
76 my ($n) = (m!/([0-9]+)\z!);
77 $n => [ URI->new($_), '' ]
79 clone_v2_prep($self, \%v2_epochs);
80 delete local $lei->{opt}->{epoch};
85 # filter out common URLs served by WWW (e.g /$MSGID/T/)
86 if (@urls && $url =~ s!/+[^/]+\@[^/]+/.*\z!! &&
87 grep(m!\A\Q$url\E/*\z!, @urls)) {
89 E: confused by scraping <$uri>, did you mean <$url>?
93 E: confused by scraping <$uri>, got ambiguous results:
96 die "E: scraping <$uri> revealed nothing\n";
100 my ($lei, $opt) = @_;
102 $opt->{$_} = $lei->{$_} for (0..2);
103 # we support "-c $key=$val" for arbitrary git config options
104 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
105 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
106 push @cmd, qw(clone --mirror);
107 push @cmd, '-q' if $lei->{opt}->{quiet} ||
108 ($lei->{opt}->{jobs} // 1) > 1;
109 push @cmd, '-v' if $lei->{opt}->{verbose};
110 # XXX any other options to support?
111 # --reference is tricky with multiple epochs, but handled
112 # automatically if using manifest.js.gz
116 sub ft_rename ($$$) {
117 my ($ft, $dst, $open_mode) = @_;
119 my $mode = @st ? ($st[2] & 07777) : ($open_mode & ~umask);
120 chmod($mode, $ft) or croak "E: chmod($ft): $!";
122 File::Copy::mv($ft->filename, $dst) or croak "E: mv($ft => $dst): $!";
123 $ft->unlink_on_destroy(0);
127 my ($self, $jobs) = @_;
128 $jobs //= $self->{-jobs} //= $self->{lei}->{opt}->{jobs} // 1;
129 $jobs = 1 if $jobs < 1;
130 while (keys(%$LIVE) >= $jobs) {
131 my $pid = waitpid(-1, 0) // die "waitpid(-1): $!";
132 if (my $x = delete $LIVE->{$pid}) {
136 warn "reaped unknown PID=$pid ($?)\n";
141 sub _get_txt_start { # non-fatal
142 my ($self, $endpoint, $fini) = @_;
143 my $uri = URI->new($self->{cur_src} // $self->{src});
144 my $lei = $self->{lei};
145 my $path = $uri->path;
146 chop($path) eq '/' or die "BUG: $uri not canonicalized";
147 $uri->path("$path/$endpoint");
148 my $f = (split(m!/!, $endpoint))[-1];
149 my $ft = File::Temp->new(TEMPLATE => "$f-XXXX", TMPDIR => 1);
150 my $opt = { 0 => $lei->{0}, 1 => $lei->{1}, 2 => $lei->{2} };
151 my $cmd = $self->{curl}->for_uri($lei, $uri, qw(-f --compressed -R -o),
154 $lei->qerr("# @$cmd");
155 return if $self->{dry_run};
156 $self->{"-get_txt.$endpoint"} = [ $ft, $cmd, $uri ];
157 $LIVE->{spawn($cmd, undef, $opt)} =
158 [ \&_get_txt_done, $self, $endpoint, $fini ];
161 sub _get_txt_done { # returns true on error (non-fatal), undef on success
162 my ($self, $endpoint) = @_;
163 my ($fh, $cmd, $uri) = @{delete $self->{"-get_txt.$endpoint"}};
165 $? = 0; # don't influence normal lei exit
166 return warn("$uri missing\n") if ($cerr >> 8) == 22;
167 return warn("# @$cmd failed (non-fatal)\n") if $cerr;
168 seek($fh, SEEK_SET, 0) or die "seek: $!";
169 $self->{"mtime.$endpoint"} = (stat($fh))[9];
171 $self->{"txt.$endpoint"} = <$fh>;
175 sub _write_inbox_config {
177 my $buf = delete($self->{'txt._/text/config/raw'}) // return;
178 my $dst = $self->{cur_dst} // $self->{dst};
179 my $f = "$dst/inbox.config.example";
180 open my $fh, '>', $f or die "open($f): $!";
181 print $fh $buf or die "print: $!";
182 chmod(0444 & ~umask, $fh) or die "chmod($f): $!";
183 my $mtime = delete $self->{'mtime._/text/config/raw'};
184 $fh->flush or die "flush($f): $!";
185 if (defined $mtime) {
186 utime($mtime, $mtime, $fh) or die "utime($f): $!";
188 my $cfg = PublicInbox::Config->git_config_dump($f, $self->{lei}->{2});
189 my $ibx = $self->{ibx} = {};
190 for my $sec (grep(/\Apublicinbox\./, @{$cfg->{-section_order}})) {
191 for (qw(address newsgroup nntpmirror)) {
192 $ibx->{$_} = $cfg->{"$sec.$_"};
197 sub set_description ($) {
199 my $dst = $self->{cur_dst} // $self->{dst};
200 my $f = "$dst/description";
201 open my $fh, '+>>', $f or die "open($f): $!";
202 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
203 my $d = do { local $/; <$fh> } // die "read($f): $!";
204 chomp(my $orig = $d);
205 while (defined($d) && ($d =~ m!^\(\$INBOX_DIR/description missing\)! ||
206 $d =~ /^Unnamed repository/ || $d !~ /\S/)) {
207 $d = delete($self->{'txt.description'});
209 $d //= 'mirror of '.($self->{cur_src} // $self->{src});
211 return if $d eq $orig;
212 seek($fh, 0, SEEK_SET) or die "seek($f): $!";
213 truncate($fh, 0) or die "truncate($f): $!";
214 print $fh $d, "\n" or die "print($f): $!";
215 close $fh or die "close($f): $!";
218 sub index_cloned_inbox {
219 my ($self, $iv) = @_;
220 my $lei = $self->{lei};
222 # n.b. public-inbox-clone works w/o (SQLite || Xapian)
223 # lei is useless without Xapian + SQLite
224 if ($lei->{cmd} ne 'public-inbox-clone') {
225 require PublicInbox::InboxWritable;
226 require PublicInbox::Admin;
227 my $ibx = delete($self->{ibx}) // {
228 address => [ 'lei@example.com' ],
231 $ibx->{inboxdir} = $self->{cur_dst} // $self->{dst};
232 PublicInbox::Inbox->new($ibx);
233 PublicInbox::InboxWritable->new($ibx);
235 for my $sw ($lei->index_opt) {
236 my ($k) = ($sw =~ /\A([\w-]+)/);
237 $opt->{$k} = $lei->{opt}->{$k};
239 # force synchronous dwaitpid for v2:
240 local $PublicInbox::DS::in_loop = 0;
241 my $cfg = PublicInbox::Config->new(undef, $lei->{2});
242 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
243 local %ENV = (%ENV, %$env) if $env;
244 PublicInbox::Admin::progress_prepare($opt, $lei->{2});
245 PublicInbox::Admin::index_inbox($ibx, undef, $opt);
247 return if defined $self->{cur_dst}; # one of many repos to clone
248 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
252 my ($lei, $cmd, $opt) = @_;
253 $lei->qerr("# @$cmd");
254 waitpid(spawn($cmd, undef, $opt), 0) // die "waitpid: $!";
256 $? = 0; # don't let it influence normal exit
261 my ($self, $cmd, $opt, $fini) = @_;
263 $self->{lei}->qerr("# @$cmd");
264 return if $self->{dry_run};
265 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_cmd, $self, $cmd, $fini ]
268 sub fetch_args ($$) {
269 my ($lei, $opt) = @_;
270 my @cmd; # (git --git-dir=...) to be added by caller
271 $opt->{$_} = $lei->{$_} for (0..2);
272 # we support "-c $key=$val" for arbitrary git config options
273 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
274 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
276 push @cmd, '-q' if $lei->{opt}->{quiet} ||
277 ($lei->{opt}->{jobs} // 1) > 1;
278 push @cmd, '-v' if $lei->{opt}->{verbose};
282 sub fgrp_update_old ($) { # for git <1.8.5
284 my $cmd = [ @{$fgrp->{-torsocks}}, 'git', "--git-dir=$fgrp->{cur_dst}",
285 fetch_args($fgrp->{lei}, my $opt = {}) ];
286 $fgrp->{lei}->qerr("# @$cmd");
288 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_cmd, $fgrp, $cmd ];
291 sub upr { # feed `git update-ref --stdin -z' verbosely
292 my ($fgrp, $w, $op, $ref, $oid) = @_;
293 $fgrp->{lei}->qerr("# $op $ref $oid");
294 print $w "$op $ref\0$oid\0" or die "print(w): $!";
299 my $srcfh = delete $fgrp->{srcfh} or return;
300 my $dstfh = delete $fgrp->{dstfh} or return;
301 seek($srcfh, SEEK_SET, 0) or die "seek(src): $!";
302 seek($dstfh, SEEK_SET, 0) or die "seek(dst): $!";
303 my %src = map { chomp; split(/\0/) } (<$srcfh>);
305 my %dst = map { chomp; split(/\0/) } (<$dstfh>);
307 pipe(my ($r, $w)) or die "pipe: $!";
308 my $cmd = [ 'git', "--git-dir=$fgrp->{cur_dst}",
309 qw(update-ref --stdin -z) ];
310 $fgrp->{lei}->qerr("# @$cmd");
311 my $opt = { 0 => $r, 1 => $fgrp->{lei}->{1}, 2 => $fgrp->{lei}->{2} };
312 my $pid = spawn($cmd, undef, $opt);
313 close $r or die "close(r): $!";
314 for my $ref (keys %dst) {
315 my $new = delete $src{$ref};
316 my $old = $dst{$ref};
318 upr($fgrp, $w, 'update', $ref, $new) if $new ne $old;
320 upr($fgrp, $w, 'delete', $ref, $old);
323 while (my ($ref, $oid) = each %src) {
324 upr($fgrp, $w, 'create', $ref, $oid);
326 if (close($w)) { # git >= 1.8.5
327 $LIVE->{$pid} = [ \&reap_cmd, $fgrp, $cmd ];
329 } else { # git <1.8.5 w/o update-ref --stdin
330 warn "E: close(update-ref --stdin): $!\n";
331 $update_ref_stdin = 0;
332 waitpid($pid, 0) // die "waitpid(update-ref --stdin): $!";
333 fgrp_update_old($fgrp);
338 my ($self, $git_dir) = @_;
340 my $cmd = [ 'git', "--git-dir=$git_dir", qw(pack-refs --all --prune) ];
341 $self->{lei}->qerr("# @$cmd");
342 my $opt = { 1 => $self->{lei}->{1}, 2 => $self->{lei}->{2} };
343 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_cmd, $self, $cmd ];
348 return if $fgrp->{dry_run} || !$LIVE;
349 my $rn = $fgrp->{-remote};
350 my %opt = map { $_ => $fgrp->{lei}->{$_} } (0..2);
351 pack_refs($fgrp, $fgrp->{-osdir}); # objstore refs always packed
353 $update_ref_stdin or return fgrp_update_old($fgrp);
355 my $update_ref = PublicInbox::OnDestroy->new($$, \&fgrp_update, $fgrp);
357 my $src = [ 'git', "--git-dir=$fgrp->{-osdir}", 'for-each-ref',
358 "--format=refs/%(refname:lstrip=3)%00%(objectname)",
359 "refs/remotes/$rn/" ];
361 open($fgrp->{srcfh}, '+>', undef) or die "open(src): $!";
362 $fgrp->{lei}->qerr("# @$src >SRC");
363 my $pid = spawn($src, undef, { %opt, 1 => $fgrp->{srcfh} });
364 $LIVE->{$pid} = [ \&reap_cmd, $fgrp, $src, $update_ref ];
366 my $dst = [ 'git', "--git-dir=$fgrp->{cur_dst}", 'for-each-ref',
367 '--format=%(refname)%00%(objectname)' ];
369 open($fgrp->{dstfh}, '+>', undef) or die "open(dst): $!";
370 $fgrp->{lei}->qerr("# @$dst >DST");
371 $pid = spawn($dst, undef, { %opt, 1 => $fgrp->{dstfh} });
372 $LIVE->{$pid} = [ \&reap_cmd, $fgrp, $dst, $update_ref ];
376 my ($fgrp, $fini) = @_;
377 my $cmd = [ @{$fgrp->{-torsocks}}, 'git', "--git-dir=$fgrp->{-osdir}",
378 fetch_args($fgrp->{lei}, my $opt = {}), '--no-tags',
380 $fgrp->{-fini} = $fini;
382 $fgrp->{lei}->qerr("# @$cmd");
383 return if $fgrp->{dry_run};
384 my $fgrp_fini = PublicInbox::OnDestroy->new($$, \&fgrp_fetched, $fgrp);
385 my $pid = spawn($cmd, undef, $opt);
386 $LIVE->{$pid} = [ \&reap_cmd, $fgrp, $cmd, $fgrp_fini ];
389 # keep this idempotent for future use by public-inbox-fetch
391 my ($self, $uri) = @_;
392 $self->{-ent} // return;
393 my $os = $self->{-objstore} // return;
394 my $fg = $self->{-ent}->{forkgroup} // return;
395 my $dir = "$os/$fg.git";
396 my @cmd = ('git', "--git-dir=$dir", 'config');
397 my $opt = +{ map { $_ => $self->{lei}->{$_} } (0..2) };
399 PublicInbox::Import::init_bare($dir);
400 for ('repack.useDeltaIslands=true',
401 'pack.island=refs/remotes/([^/]+)/') {
402 run_die([@cmd, split(/=/, $_, 2)], undef, $opt);
405 my $key = $self->{-key} // die 'BUG: no -key';
406 my ($bn) = ($key =~ m{/([a-z0-9_,;=!\+\{\}\|][^/]*)(?:\.git)?\z}i);
407 my $rn = "$bn-".substr(sha256_hex($key), 0, 16);
408 # --no-tags is required to avoid conflicts
409 for ("url=$uri", "fetch=+refs/*:refs/remotes/$rn/*",
410 'tagopt=--no-tags') {
411 my @kv = split(/=/, $_, 2);
412 $kv[0] = "remote.$rn.$kv[0]";
413 run_die([@cmd, @kv], undef, $opt);
415 $self->{-do_pack_refs} = 1; # likely coderepo
416 if (!-d $self->{cur_dst}) {
417 my $alt = File::Spec->rel2abs("$dir/objects");
418 PublicInbox::Import::init_bare($self->{cur_dst});
419 my $o = "$self->{cur_dst}/objects";
420 my $f = "$o/info/alternates";
421 my $l = File::Spec->abs2rel($alt, File::Spec->rel2abs($o));
422 open my $fh, '+>>', $f or die "open($f): $!";
423 seek($fh, SEEK_SET, 0) or die "seek($f): $!";
424 chomp(my @cur = <$fh>);
425 if (!grep(/\A\Q$l\E\z/, @cur)) {
426 say $fh $l or die "say($f): $!";
428 close $fh or die "close($f): $!";
429 @cmd = ('git', "--git-dir=$self->{cur_dst}",
430 qw(remote add --mirror=fetch origin), "$uri");
431 my $pid = spawn(\@cmd, undef, $opt);
432 waitpid($pid, 0) // die "waitpid(@cmd): $!";
433 die "E: @cmd: \$?=$?" if ($? && ($? >> 8) != 3);
435 bless { %$self, -osdir => $dir, -remote => $rn }, __PACKAGE__;
439 my ($self, $nohang) = @_;
440 my $lei = $self->{lei};
441 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
442 my $uri = URI->new($self->{cur_src} // $self->{src});
443 defined($lei->{opt}->{epoch}) and
444 die "$uri is a v1 inbox, --epoch is not supported\n";
445 $self->{-torsocks} //= $curl->torsocks($lei, $uri) or return;
446 my $dst = $self->{cur_dst} // $self->{dst};
447 my $fini = PublicInbox::OnDestroy->new($$, \&v1_done, $self);
448 my $cmd = [ @{$self->{-torsocks}}, clone_cmd($lei, my $opt = {}),
450 my $fgrp = forkgroup_prep($self, $uri);
451 if (!defined($fgrp) && defined($self->{-ent})) {
452 if (defined(my $ref = $self->{-ent}->{reference})) {
453 -e "$self->{dst}$ref" and
454 push @$cmd, '--reference', "$self->{dst}$ref";
457 $fgrp ? fgrp_fetch($fgrp, $fini) :
458 start_clone($self, $cmd, $opt, $fini);
460 if (!$self->{-is_epoch} && $lei->{opt}->{'inbox-config'} =~
461 /\A(?:always|v1)\z/s) {
462 _get_txt_start($self, '_/text/config/raw', $fini);
465 my $d = $self->{-ent} ? $self->{-ent}->{description} : undef;
466 $self->{'txt.description'} = $d if defined $d;
467 (!defined($d) && !$nohang) and
468 _get_txt_start($self, 'description', $fini);
470 $nohang or do_reap($self, 1); # for non-manifest clone
473 sub parse_epochs ($$) {
474 my ($opt_epochs, $v2_epochs) = @_; # $epochs "LOW..HIGH"
475 $opt_epochs // return; # undef => all epochs
476 my ($lo, $dotdot, $hi, @extra) = split(/(\.\.)/, $opt_epochs);
477 undef($lo) if ($lo // '') eq '';
478 my $re = qr/\A~?[0-9]+\z/;
479 if (@extra || (($lo // '0') !~ $re) ||
480 (($hi // '0') !~ $re) ||
481 !(grep(defined, $lo, $hi))) {
483 --epoch=$opt_epochs not in the form of `LOW..HIGH', `LOW..', nor `..HIGH'
486 my @n = sort { $a <=> $b } keys %$v2_epochs;
487 for (grep(defined, $lo, $hi)) {
490 "`$_' exceeds maximum available epoch ($n[-1])\n";
492 "`$_' is lower than minimum available epoch ($n[0])\n";
493 } elsif (/\A~([0-9]+)/) {
495 $n[$off] // die "`$_' is out of range\n";
497 } else { die "`$_' not understood\n" }
499 defined($lo) && defined($hi) && $lo > $hi and die
500 "low value (`$lo') exceeds high (`$hi')\n";
501 $lo //= $n[0] if $dotdot;
502 $hi //= $n[-1] if $dotdot;
506 if (defined $v2_epochs->{$_}) {
510 "# epoch $_ is not available (non-fatal, $lo..$hi)\n";
516 sub init_placeholder ($$$) {
517 my ($src, $edst, $ent) = @_;
518 PublicInbox::Import::init_bare($edst);
519 my $f = "$edst/config";
520 open my $fh, '>>', $f or die "open($f): $!";
521 print $fh <<EOM or die "print($f): $!";
524 fetch = +refs/*:refs/*
527 ; This git epoch was created read-only and "public-inbox-fetch"
528 ; will not fetch updates for it unless write permission is added.
529 ; Hint: chmod +w $edst
531 if (defined($ent->{owner})) {
532 print $fh <<EOM or die "print($f): $!";
534 owner = $ent->{owner}
537 close $fh or die "close($f): $!";
538 my %map = (head => 'HEAD', description => undef);
539 while (my ($key, $fn) = each %map) {
540 my $val = $ent->{$key} // next;
543 open $fh, '>', $fn or die "open($fn): $!";
544 print $fh $val, "\n" or die "print($fn): $!";
545 close $fh or die "close($fn): $!";
549 sub reap_cmd { # async, called via SIGCHLD
550 my ($self, $cmd) = @_;
552 $? = 0; # don't let it influence normal exit
554 kill('TERM', keys %$LIVE);
555 $self->{lei}->child_error($cerr, "@$cmd failed");
559 sub v1_done { # called via OnDestroy
561 return if $self->{dry_run} || !$LIVE;
562 _write_inbox_config($self);
563 my $dst = $self->{cur_dst} // $self->{dst};
564 if (defined(my $o = $self->{-ent} ? $self->{-ent}->{owner} : undef)) {
565 run_die([qw(git config -f), "$dst/config", 'gitweb.owner', $o]);
567 my $o = "$dst/objects";
568 if (open(my $fh, '<', "$o/info/alternates")) {
569 my $base = File::Spec->rel2abs($o);
570 chomp(my @l = <$fh>);
572 $_ = File::Spec->abs2rel($_, $base) if m!\A/!;
575 my $f = File::Temp->new(TEMPLATE => '.XXXX', DIR => "$o/info");
577 $f->flush or die "flush($f): $!";
578 rename($f->filename, "$o/info/alternates") or
579 die "rename($f, $o/info/alternates): $!";
580 $f->unlink_on_destroy(0);
582 pack_refs($self, $dst) if delete $self->{-do_pack_refs};
583 eval { set_description($self) };
585 return if ($self->{-is_epoch} ||
586 $self->{lei}->{opt}->{'inbox-config'} ne 'always');
587 write_makefile($dst, 1);
588 index_cloned_inbox($self, 1);
591 sub v2_done { # called via OnDestroy
593 return if $self->{dry_run} || !$LIVE;
594 my $dst = $self->{cur_dst} // $self->{dst};
595 require PublicInbox::Lock;
596 my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
597 my $lck = $lk->lock_for_scope($$);
598 _write_inbox_config($self);
599 require PublicInbox::MultiGit;
600 my $mg = PublicInbox::MultiGit->new($dst, 'all.git', 'git');
601 $mg->fill_alternates;
602 for my $i ($mg->git_epochs) { $mg->epoch_cfg_set($i) }
603 for my $edst (@{delete($self->{-read_only}) // []}) {
604 my @st = stat($edst) or die "stat($edst): $!";
605 chmod($st[2] & 0555, $edst) or die "chmod(a-w, $edst): $!";
607 write_makefile($dst, 2);
609 eval { set_description($self) };
611 index_cloned_inbox($self, 2);
614 sub clone_v2_prep ($$;$) {
615 my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
616 my $lei = $self->{lei};
617 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
618 my $first_uri = (map { $_->[0] } values %$v2_epochs)[0];
619 $self->{-torsocks} //= $curl->torsocks($lei, $first_uri) or return;
620 my $dst = $self->{cur_dst} // $self->{dst};
621 my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
622 my $task = $m ? bless { %$self }, __PACKAGE__ : $self;
623 delete $task->{todo}; # $self->{todo} still exists
625 my $fini = PublicInbox::OnDestroy->new($$, \&v2_done, $task);
626 for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
627 my ($uri, $key) = @{$v2_epochs->{$nr}};
628 my $src = $uri->as_string;
630 $src =~ m!/([0-9]+)(?:\.git)?\z! or die <<"";
631 failed to extract epoch number from $src
633 $1 + 0 == $nr or die "BUG: <$uri> miskeyed $1 != $nr";
634 $edst .= "/git/$nr.git";
638 die("BUG: `$key' not in manifest.js.gz");
639 if (defined(my $d = $ent->{description})) {
640 $d =~ s/ \[epoch [0-9]+\]\z//s;
644 if (!$want || $want->{$nr}) {
645 my $etask = bless { %$task, -key => $key }, __PACKAGE__;
646 $etask->{-ent} = $ent; # may have {reference}
647 $etask->{cur_src} = $src;
648 $etask->{cur_dst} = $edst;
649 $etask->{-is_epoch} = $fini;
650 my $ref = $ent->{reference} // '';
651 push @{$self->{todo}->{$ref}}, $etask;
652 $self->{any_want}->{$key} = 1;
653 } else { # create a placeholder so users only need to chmod +w
654 init_placeholder($src, $edst, $ent);
655 push @{$task->{-read_only}}, $edst;
659 # filter out the epochs we skipped
660 $self->{-culled_manifest} = 1 if $m && delete(@$m{@skip});
662 (!$self->{dry_run} && !-d $dst) and File::Path::mkpath($dst);
664 $lei->{opt}->{'inbox-config'} =~ /\A(?:always|v2)\z/s and
665 _get_txt_start($task, '_/text/config/raw', $fini);
667 defined($desc) ? ($task->{'txt.description'} = $desc) :
668 _get_txt_start($task, 'description', $fini);
671 sub decode_manifest ($$$) {
672 my ($fh, $fn, $uri) = @_;
674 my $gz = do { local $/; <$fh> } // die "slurp($fn): $!";
675 gunzip(\$gz => \$js, MultiStream => 1) or
676 die "gunzip($uri): $GunzipError\n";
677 my $m = eval { PublicInbox::Config->json->decode($js) };
678 die "$uri: error decoding `$js': $@\n" if $@;
679 ref($m) eq 'HASH' or die "$uri unknown type: ".ref($m);
683 sub multi_inbox ($$$) {
684 my ($self, $path, $m) = @_;
685 my $incl = $self->{lei}->{opt}->{include};
686 my $excl = $self->{lei}->{opt}->{exclude};
688 # assuming everything not v2 is v1, for now
689 my @v1 = sort grep(!m!.+/git/[0-9]+\.git\z!, keys %$m);
690 my @v2_epochs = sort grep(m!.+/git/[0-9]+\.git\z!, keys %$m);
694 m!\A(/.+)/git/[0-9]+\.git\z! or die "BUG: $_";
695 push @{$v2->{$1}}, $_;
697 my $n = scalar(keys %$v2) + scalar(@v1);
698 my @orig = defined($incl // $excl) ? (keys %$v2, @v1) : ();
700 my $re = '(?:'.join('\\z|', map {
701 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
703 my @gone = delete @$v2{grep(!/$re/, keys %$v2)};
704 delete @$m{map { @$_ } @gone} and $self->{-culled_manifest} = 1;
705 delete @$m{grep(!/$re/, @v1)} and $self->{-culled_manifest} = 1;
706 @v1 = grep(/$re/, @v1);
709 my $re = '(?:'.join('\\z|', map {
710 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
712 my @gone = delete @$v2{grep(/$re/, keys %$v2)};
713 delete @$m{map { @$_ } @gone} and $self->{-culled_manifest} = 1;
714 delete @$m{grep(/$re/, @v1)} and $self->{-culled_manifest} = 1;
715 @v1 = grep(!/$re/, @v1);
717 my $ret; # { v1 => [ ... ], v2 => { "/$inbox_name" => [ epochs ] }}
718 $ret->{v1} = \@v1 if @v1;
719 $ret->{v2} = $v2 if keys %$v2;
720 $ret //= @orig ? "Nothing to clone, available repositories:\n\t".
721 join("\n\t", sort @orig)
722 : "Nothing available to clone\n";
725 # PSGI mount prefixes and manifest.js.gz prefixes don't always align...
727 until (grep(m!\A\Q$$path\E/git/[0-9]+\.git\z!,
728 @v2_epochs) == @v2_epochs) {
729 $$path =~ s!\A(/[^/]+)/!/! or last;
733 while (!defined($m->{$$path}) && $$path =~ s!\A(/[^/]+)/!/!) {
737 ($path_pfx, $n, $ret);
742 my $todo = delete $self->{todo};
743 my $nodep = delete $todo->{''};
745 # do not download unwanted deps
746 my $any_want = delete $self->{any_want};
747 my @unwanted = grep { !$any_want->{$_} } keys %$todo;
748 my @nodep = delete(@$todo{@unwanted});
749 push(@$nodep, @$_) for @nodep;
751 # handle no-dependency repos, first
754 return if $self->{lei}->{child_error};
756 # resolve references, deepest, first:
757 while (scalar keys %$todo) {
758 for my $x (keys %$todo) {
760 # resolve multi-level references
761 while ($m && defined($nxt = $m->{$x}->{reference})) {
762 exists($todo->{$nxt}) or last;
763 die <<EOM if ++$nr > 1000;
764 E: dependency loop detected (`$x' => `$nxt')
768 my $y = delete $todo->{$x} // next; # already done
771 return if $self->{lei}->{child_error};
773 last; # restart %$todo iteration
779 sub dump_manifest ($$) {
781 # write the smaller manifest if epochs were skipped so
782 # users won't have to delete manifest if they +w an
783 # epoch they no longer want to skip
784 my $json = PublicInbox::Config->json->encode($m);
785 my $mtime = (stat($ft))[9];
786 seek($ft, SEEK_SET, 0) or die "seek($ft): $!";
787 truncate($ft, 0) or die "truncate($ft): $!";
788 gzip(\$json => $ft) or die "gzip($ft): $GzipError";
789 $ft->flush or die "flush($ft): $!";
790 utime($mtime, $mtime, "$ft") or die "utime(..., $ft): $!";
793 # FIXME: this gets confused by single inbox instance w/ global manifest.js.gz
796 my $uri = URI->new($self->{src});
797 my $lei = $self->{lei};
798 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
799 my $path = $uri->path;
800 chop($path) eq '/' or die "BUG: $uri not canonicalized";
801 $uri->path($path . '/manifest.js.gz');
802 my $ft = File::Temp->new(TEMPLATE => '.manifest-XXXX',
803 UNLINK => 1, TMPDIR => 1, SUFFIX => '.tmp');
804 my $cmd = $curl->for_uri($lei, $uri, qw(-f -R -o), $ft->filename);
805 my %opt = map { $_ => $lei->{$_} } (0..2);
806 my $cerr = run_reap($lei, $cmd, \%opt);
808 return try_scrape($self) if ($cerr >> 8) == 22; # 404 missing
809 return $lei->child_error($cerr, "@$cmd failed");
811 my $m = eval { decode_manifest($ft, $ft, $uri) };
814 return try_scrape($self);
816 my ($path_pfx, $n, $multi) = multi_inbox($self, \$path, $m);
817 return $lei->child_error(1, $multi) if !ref($multi);
818 my $v2 = delete $multi->{v2};
819 local $self->{todo} = {};
821 for my $name (sort keys %$v2) {
822 my $epochs = delete $v2->{$name};
823 my %v2_epochs = map {
824 $uri->path($n > 1 ? $path_pfx.$path.$_
826 my ($e) = ("$uri" =~ m!/([0-9]+)\.git\z!);
827 $e // die "no [0-9]+\.git in `$uri'";
828 $e => [ $uri->clone, $_ ];
830 ("$uri" =~ m!\A(.+/)git/[0-9]+\.git\z!) or
831 die "BUG: `$uri' !~ m!/git/[0-9]+.git!";
832 local $self->{cur_src} = $1;
833 local $self->{cur_dst} = $self->{dst};
834 if ($n > 1 && $uri->path =~ m!\A\Q$path_pfx$path\E/(.+)/
835 git/[0-9]+\.git\z!x) {
836 $self->{cur_dst} .= "/$1";
838 index($self->{cur_dst}, "\n") >= 0 and die <<EOM;
839 E: `$self->{cur_dst}' must not contain newline
841 clone_v2_prep($self, \%v2_epochs, $m);
842 return if $self->{lei}->{child_error};
845 if (my $v1 = delete $multi->{v1}) {
846 my $p = $path_pfx.$path;
847 chop($p) if substr($p, -1, 1) eq '/';
849 for my $name (@$v1) {
850 my $task = bless { %$self }, __PACKAGE__;
851 $task->{-ent} = $m->{$name} //
852 die("BUG: no `$name' in manifest");
853 $task->{cur_src} = "$uri";
854 $task->{cur_dst} = $task->{dst};
855 $task->{-key} = $name;
857 $task->{cur_dst} .= $name;
858 $task->{cur_src} .= $name;
860 index($task->{cur_dst}, "\n") >= 0 and die <<EOM;
861 E: `$task->{cur_dst}' must not contain newline
863 $task->{cur_src} .= '/';
864 my $dep = $task->{-ent}->{reference} // '';
865 push @{$self->{todo}->{$dep}}, $task; # for clone_all
866 $self->{any_want}->{$name} = 1;
869 delete local $lei->{opt}->{epoch} if defined($v2);
870 clone_all($self, $m);
871 return if $self->{lei}->{child_error} || $self->{dry_run};
873 # set by clone_v2_prep/-I/--exclude
874 dump_manifest($m => $ft) if delete $self->{-culled_manifest};
875 ft_rename($ft, "$self->{dst}/manifest.js.gz", 0666);
876 open my $x, '>', "$self->{dst}/mirror.done"; # for _wq_done_wait
879 sub start_clone_url {
881 return try_manifest($self) if $self->{src} =~ m!\Ahttps?://!;
882 die "TODO: non-HTTP/HTTPS clone of $self->{src} not supported, yet";
885 sub do_mirror { # via wq_io_do or public-inbox-clone
887 my $lei = $self->{lei};
888 umask($lei->{client_umask}) if defined $lei->{client_umask};
890 my $ic = $lei->{opt}->{'inbox-config'} //= 'always';
891 $ic =~ /\A(?:v1|v2|always|never)\z/s or die <<"";
892 --inbox-config must be one of `always', `v2', `v1', or `never'
894 if (defined(my $os = $lei->{opt}->{objstore})) {
895 $os = 'objstore' if $os eq ''; # --objstore w/o args
896 $os = "$self->{dst}/$os" if $os !~ m!\A/!;
897 $self->{-objstore} = $os;
900 my $iv = $lei->{opt}->{'inbox-version'} //
901 return start_clone_url($self);
902 return clone_v1($self) if $iv == 1;
903 die "bad --inbox-version=$iv\n" if $iv != 2;
904 die <<EOM if $self->{src} !~ m!://!;
905 cloning local v2 inboxes not supported
907 try_scrape($self, 1);
909 $lei->fail($@) if $@;
913 my ($cls, $lei, $src, $dst) = @_;
914 my $self = bless { src => $src, dst => $dst }, $cls;
916 my ($op_c, $ops) = $lei->workers_start($self, 1);
918 $self->wq_io_do('do_mirror', []);
920 $lei->wait_wq_events($op_c, $ops);
923 sub ipc_atfork_child {
925 $self->{lei}->_lei_atfork_child;
926 $self->SUPER::ipc_atfork_child;
930 my ($dir, $ibx_ver) = @_;
931 my $f = "$dir/Makefile";
932 if (sysopen my $fh, $f, O_CREAT|O_EXCL|O_WRONLY) {
933 print $fh <<EOM or die "print($f) $!";
934 # This is a v$ibx_ver public-inbox, see the public-inbox-v$ibx_ver-format(5)
935 # manpage for more information on the format. This Makefile is
936 # intended as a familiar wrapper for users unfamiliar with
937 # public-inbox-* commands.
939 # See the respective manpages for public-inbox-fetch(1),
940 # public-inbox-index(1), etc for more information on
941 # some of the commands used by this Makefile.
943 # This Makefile will not be modified nor read by public-inbox,
944 # so you may edit it freely with your own convenience targets
945 # and notes. public-inbox-fetch will recreate it if removed.
947 print $fh <<'EOM' or die "print($f): $!";
948 # the default target:
950 @echo Common targets:
951 @echo ' make fetch - fetch from remote git repostorie(s)'
952 @echo ' make update - fetch and update index '
954 @echo Rarely needed targets:
955 @echo ' make reindex - may be needed for new features/bugfixes'
956 @echo ' make compact - rewrite Xapian storage to save space'
957 @echo ' make index - initial index after clone
962 @if ! public-inbox-fetch --exit-code; \
965 test $$c -eq 127 && exit 0; \
967 elif test -f msgmap.sqlite3 || test -f public-inbox/msgmap.sqlite3; \
969 public-inbox-index; \
971 echo 'public-inbox index not initialized'; \
972 echo 'see public-inbox-index(1) man page'; \
977 public-inbox-index --reindex
981 .PHONY : help fetch update index reindex compact
983 close $fh or die "close($f): $!";
985 die "open($f): $!" unless $!{EEXIST};