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;
20 use PublicInbox::LeiCurl;
21 use PublicInbox::OnDestroy;
22 use PublicInbox::SHA qw(sha256_hex sha1_hex);
23 use POSIX qw(strftime);
25 our $LIVE; # pid => callback
26 our $FGRP_TODO; # objstore -> [ fgrp mirror objects ]
27 our $TODO; # reference => [ non-fgrp mirror objects ]
28 our @PUH; # post-update hooks
31 $LIVE && (!$_[0]->{lei}->{child_error} ||
32 $_[0]->{lei}->{opt}->{'keep-going'});
35 sub _wq_done_wait { # awaitpid cb (via wq_eof)
36 my ($pid, $mrr, $lei) = @_;
38 $lei->child_error($?);
39 } elsif (!$lei->{child_error}) {
40 if (!$mrr->{dry_run} && $lei->{cmd} ne 'public-inbox-clone') {
41 require PublicInbox::LeiAddExternal;
42 PublicInbox::LeiAddExternal::_finish_add_external(
45 $lei->qerr("# mirrored $mrr->{src} => $mrr->{dst}");
50 # for old installations without manifest.js.gz
52 my ($self, $fallback_manifest) = @_;
53 my $uri = URI->new($self->{src});
54 my $lei = $self->{lei};
55 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
56 my $cmd = $curl->for_uri($lei, $uri, '--compressed');
57 my $opt = { 0 => $lei->{0}, 2 => $lei->{2} };
58 my $fh = popen_rd($cmd, undef, $opt);
59 my $html = do { local $/; <$fh> } // die "read(curl $uri): $!";
60 close($fh) or return $lei->child_error($?, "@$cmd failed");
62 # we grep with URL below, we don't want Subject/From headers
63 # making us clone random URLs. This assumes remote instances
64 # prior to public-inbox 1.7.0
65 # 5b96edcb1e0d8252 (www: move mirror instructions to /text/, 2021-08-28)
66 my @html = split(/<hr>/, $html);
67 my @urls = ($html[-1] =~ m!\bgit clone --mirror ([a-z\+]+://\S+)!g);
68 if (!@urls && $fallback_manifest) {
70 W: failed to extract URLs from $uri, trying manifest.js.gz...
72 return start_clone_url($self);
74 my $url = $uri->as_string;
75 chop($url) eq '/' or die "BUG: $uri not canonicalized";
77 # since this is for old instances w/o manifest.js.gz, try v1 first
78 return clone_v1($self) if grep(m!\A\Q$url\E/*\z!, @urls);
79 if (my @v2_urls = grep(m!\A\Q$url\E/[0-9]+\z!, @urls)) {
81 my ($n) = (m!/([0-9]+)\z!);
82 $n => [ URI->new($_), '' ]
84 clone_v2_prep($self, \%v2_epochs);
85 delete local $lei->{opt}->{epoch};
90 # filter out common URLs served by WWW (e.g /$MSGID/T/)
91 if (@urls && $url =~ s!/+[^/]+\@[^/]+/.*\z!! &&
92 grep(m!\A\Q$url\E/*\z!, @urls)) {
94 E: confused by scraping <$uri>, did you mean <$url>?
98 E: confused by scraping <$uri>, got ambiguous results:
101 die "E: scraping <$uri> revealed nothing\n";
105 my ($lei, $opt) = @_;
107 $opt->{$_} = $lei->{$_} for (0..2);
108 # we support "-c $key=$val" for arbitrary git config options
109 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
110 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
111 push @cmd, qw(clone --mirror);
112 push @cmd, '-q' if $lei->{opt}->{quiet} ||
113 ($lei->{opt}->{jobs} // 1) > 1;
114 push @cmd, '-v' if $lei->{opt}->{verbose};
115 # XXX any other options to support?
116 # --reference is tricky with multiple epochs, but handled
117 # automatically if using manifest.js.gz
121 sub ft_rename ($$$;$) {
122 my ($ft, $dst, $open_mode, $fh) = @_;
123 my @st = stat($fh // $dst);
124 my $mode = @st ? ($st[2] & 07777) : ($open_mode & ~umask);
125 chmod($mode, $ft) or croak "E: chmod($ft): $!";
127 File::Copy::mv($ft->filename, $dst) or croak "E: mv($ft => $dst): $!";
128 $ft->unlink_on_destroy(0);
132 my ($self, $jobs) = @_;
133 $jobs //= $self->{-jobs} //= $self->{lei}->{opt}->{jobs} // 1;
134 $jobs = 1 if $jobs < 1;
135 while (keys(%$LIVE) >= $jobs) {
136 my $pid = waitpid(-1, 0) // die "waitpid(-1): $!";
137 if (my $x = delete $LIVE->{$pid}) {
141 warn "reaped unknown PID=$pid ($?)\n";
146 sub _get_txt_start { # non-fatal
147 my ($self, $endpoint, $fini) = @_;
148 my $uri = URI->new($self->{cur_src} // $self->{src});
149 my $lei = $self->{lei};
150 my $path = $uri->path;
151 chop($path) eq '/' or die "BUG: $uri not canonicalized";
152 $uri->path("$path/$endpoint");
153 my $f = (split(m!/!, $endpoint))[-1];
154 my $ft = File::Temp->new(TEMPLATE => "$f-XXXX", TMPDIR => 1);
155 my $opt = { 0 => $lei->{0}, 1 => $lei->{1}, 2 => $lei->{2} };
156 my $cmd = $self->{curl}->for_uri($lei, $uri, qw(--compressed -R -o),
159 $lei->qerr("# @$cmd");
160 return if $self->{dry_run};
161 $self->{"-get_txt.$endpoint"} = [ $ft, $cmd, $uri ];
162 $LIVE->{spawn($cmd, undef, $opt)} =
163 [ \&_get_txt_done, $self, $endpoint, $fini ];
166 sub _get_txt_done { # returns true on error (non-fatal), undef on success
167 my ($self, $endpoint) = @_;
168 my ($fh, $cmd, $uri) = @{delete $self->{"-get_txt.$endpoint"}};
170 $? = 0; # don't influence normal lei exit
171 return warn("$uri missing\n") if ($cerr >> 8) == 22;
172 return warn("# @$cmd failed (non-fatal)\n") if $cerr;
173 seek($fh, SEEK_SET, 0) or die "seek: $!";
174 $self->{"mtime.$endpoint"} = (stat($fh))[9];
176 $self->{"txt.$endpoint"} = <$fh>;
180 sub _write_inbox_config {
182 my $buf = delete($self->{'txt._/text/config/raw'}) // return;
183 my $dst = $self->{cur_dst} // $self->{dst};
184 my $f = "$dst/inbox.config.example";
185 my $mtime = delete $self->{'mtime._/text/config/raw'};
186 if (sysopen(my $fh, $f, O_CREAT|O_EXCL|O_WRONLY)) {
187 print $fh $buf or die "print: $!";
188 chmod(0444 & ~umask, $fh) or die "chmod($f): $!";
189 $fh->flush or die "flush($f): $!";
190 if (defined $mtime) {
191 utime($mtime, $mtime, $fh) or die "utime($f): $!";
193 } elsif (!$!{EEXIST}) {
196 my $cfg = PublicInbox::Config->git_config_dump($f, $self->{lei}->{2});
197 my $ibx = $self->{ibx} = {};
198 for my $sec (grep(/\Apublicinbox\./, @{$cfg->{-section_order}})) {
199 for (qw(address newsgroup nntpmirror)) {
200 $ibx->{$_} = $cfg->{"$sec.$_"};
205 sub set_description ($) {
207 my $dst = $self->{cur_dst} // $self->{dst};
208 chomp(my $orig = PublicInbox::Git::try_cat("$dst/description"));
210 while (defined($d) && ($d =~ m!^\(\$INBOX_DIR/description missing\)! ||
211 $d =~ /^Unnamed repository/ || $d !~ /\S/)) {
212 $d = delete($self->{'txt.description'});
214 $d //= 'mirror of '.($self->{cur_src} // $self->{src});
215 atomic_write($dst, 'description', $d."\n") if $d ne $orig;
218 sub index_cloned_inbox {
219 my ($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 awaitpid 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 $self->{cur_dst}; # one of many repos to clone
251 my ($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 utf8::decode(my $msg = "# @$cmd");
263 $self->{lei}->qerr($msg);
264 return if $self->{dry_run};
265 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_cmd, $self, $cmd, $fini ]
268 sub fetch_args ($$) {
269 my ($lei, $opt) = @_;
270 my @cmd; # (git --git-dir=...) to be added by caller
271 $opt->{$_} = $lei->{$_} for (0..2);
272 # we support "-c $key=$val" for arbitrary git config options
273 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
274 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
276 push @cmd, '-q' if $lei->{opt}->{quiet} ||
277 ($lei->{opt}->{jobs} // 1) > 1;
278 push @cmd, '-v' if $lei->{opt}->{verbose};
279 push(@cmd, '-p') if $lei->{opt}->{prune};
280 PublicInbox::Git::version() >= ((2 << 24) | (29 << 16)) and
281 push(@cmd, '--no-write-fetch-head');
285 sub upr { # feed `git update-ref --stdin -z' verbosely
286 my ($lei, $w, $op, @rest) = @_; # ($ref, $oid) = @rest
287 $lei->qerr("# $op @rest") if $lei->{opt}->{verbose};
288 print $w "$op ", join("\0", @rest, '') or die "print(w): $!";
291 sub start_update_ref {
293 pipe(my ($r, $w)) or die "pipe: $!";
294 my $cmd = [ 'git', "--git-dir=$fgrp->{cur_dst}",
295 qw(update-ref --stdin -z) ];
296 my $pack = PublicInbox::OnDestroy->new($$, \&satellite_done, $fgrp);
297 start_cmd($fgrp, $cmd, { 0 => $r, 2 => $fgrp->{lei}->{2} }, $pack);
298 close $r or die "close(r): $!";
299 $fgrp->{dry_run} ? undef : $w;
302 sub upref_warn { warn "E: close(update-ref --stdin): $! (need git 1.8.5+)\n" }
306 return if !keep_going($fgrp);
307 my $srcfh = delete $fgrp->{srcfh} or return;
308 my $dstfh = delete $fgrp->{dstfh} or return;
309 seek($srcfh, SEEK_SET, 0) or die "seek(src): $!";
310 seek($dstfh, SEEK_SET, 0) or die "seek(dst): $!";
311 my %src = map { chomp; split(/\0/) } (<$srcfh>);
313 my %dst = map { chomp; split(/\0/) } (<$dstfh>);
315 my $w = start_update_ref($fgrp) or return;
316 my $lei = $fgrp->{lei};
318 for my $ref (keys %dst) {
319 my $new = delete $src{$ref};
320 my $old = $dst{$ref};
323 upr($lei, $w, 'update', $ref, $new, $old);
325 upr($lei, $w, 'delete', $ref, $old);
329 # git's ref files backend doesn't allow directory/file conflicts
330 # between `delete' and `create' ops:
331 if ($ndel && scalar(keys %src)) {
332 $fgrp->{-create_refs} = \%src;
334 while (my ($ref, $oid) = each %src) {
335 upr($lei, $w, 'create', $ref, $oid);
338 close($w) or upref_warn();
343 if (my $create = delete $fgrp->{-create_refs}) {
344 my $w = start_update_ref($fgrp) or return;
345 while (my ($ref, $oid) = each %$create) {
346 upr($fgrp->{lei}, $w, 'create', $ref, $oid);
348 close($w) or upref_warn();
350 pack_refs($fgrp, $fgrp->{cur_dst});
356 my ($self, $git_dir) = @_;
357 my $cmd = [ 'git', "--git-dir=$git_dir", qw(pack-refs --all --prune) ];
358 start_cmd($self, $cmd, { 2 => $self->{lei}->{2} });
364 my $first = $fgrpv->[0] // die 'BUG: no fgrpv->[0]';
365 return if !keep_going($first);
366 pack_refs($first, $first->{-osdir}); # objstore refs always packed
367 for my $fgrp (@$fgrpv) {
368 my $rn = $fgrp->{-remote};
369 my %opt = ( 2 => $fgrp->{lei}->{2} );
371 my $update_ref = PublicInbox::OnDestroy->new($$,
372 \&fgrp_update, $fgrp);
374 my $src = [ 'git', "--git-dir=$fgrp->{-osdir}", 'for-each-ref',
375 "--format=refs/%(refname:lstrip=3)%00%(objectname)",
376 "refs/remotes/$rn/" ];
377 open(my $sfh, '+>', undef) or die "open(src): $!";
378 $fgrp->{srcfh} = $sfh;
379 start_cmd($fgrp, $src, { %opt, 1 => $sfh }, $update_ref);
380 my $dst = [ 'git', "--git-dir=$fgrp->{cur_dst}", 'for-each-ref',
381 '--format=%(refname)%00%(objectname)' ];
382 open(my $dfh, '+>', undef) or die "open(dst): $!";
383 $fgrp->{dstfh} = $dfh;
384 start_cmd($fgrp, $dst, { %opt, 1 => $dfh }, $update_ref);
390 my $todo = $FGRP_TODO;
391 $FGRP_TODO = \'BUG on further use';
392 keys(%$todo) or return;
394 # Rely on the fgrptmp remote groups in the config file rather
395 # than listing all remotes since the remote name list may exceed
396 # system argv limits:
399 my @git = (@{$self->{-torsocks}}, 'git');
400 my $j = $self->{lei}->{opt}->{jobs};
403 local $self->{lei}->{opt}->{jobs} = 1;
404 (fetch_args($self->{lei}, $opt),
405 qw(--no-tags --multiple));
407 push(@fetch, "-j$j") if $j;
408 while (my ($osdir, $fgrpv) = each %$todo) {
409 my $f = "$osdir/config";
410 return if !keep_going($self);
412 # clobber group from previous run atomically
413 my $cmd = ['git', "--git-dir=$osdir", qw(config -f),
414 $f, '--unset-all', "remotes.$grp"];
415 $self->{lei}->qerr("# @$cmd");
416 if (!$self->{dry_run}) {
417 my $pid = spawn($cmd, undef, { 2 => $self->{lei}->{2} });
418 waitpid($pid, 0) // die "waitpid: $!";
419 die "E: @$cmd: \$?=$?" if ($? && ($? >> 8) != 5);
421 # update the config atomically via O_APPEND while
422 # respecting git-config locking
423 sysopen(my $lk, "$f.lock", O_CREAT|O_EXCL|O_WRONLY)
424 or die "open($f.lock): $!";
425 open my $fh, '>>', $f or die "open(>>$f): $!";
427 my $buf = join('', "[remotes]\n",
428 map { "\t$grp = $_->{-remote}\n" } @$fgrpv);
429 print $fh $buf or die "print($f): $!";
430 close $fh or die "close($f): $!";
431 unlink("$f.lock") or die "unlink($f.lock): $!";
434 $cmd = [ @git, "--git-dir=$osdir", @fetch, $grp ];
435 my $end = PublicInbox::OnDestroy->new($$, \&fgrpv_done, $fgrpv);
436 start_cmd($self, $cmd, $opt, $end);
440 # keep this idempotent for future use by public-inbox-fetch
442 my ($self, $uri) = @_;
443 $self->{-ent} // return;
444 my $os = $self->{-objstore} // return;
445 my $fg = $self->{-ent}->{forkgroup} // return;
446 my $dir = "$os/$fg.git";
447 if (!-d $dir && !$self->{dry_run}) {
448 PublicInbox::Import::init_bare($dir);
449 my @cmd = ('git', "--git-dir=$dir", 'config');
450 my $opt = { 2 => $self->{lei}->{2} };
451 for ('repack.useDeltaIslands=true',
452 'pack.island=refs/remotes/([^/]+)/') {
453 run_die([@cmd, split(/=/, $_, 2)], undef, $opt);
456 my $key = $self->{-key} // die 'BUG: no -key';
457 my $rn = substr(sha256_hex($key), 0, 16);
458 if (!-d $self->{cur_dst} && !$self->{dry_run}) {
459 PublicInbox::Import::init_bare($self->{cur_dst});
460 my $f = "$self->{cur_dst}/config";
461 open my $fh, '+>>', $f or die "open:($f): $!";
462 print $fh <<EOM or die "print($f): $!";
463 ; rely on the "$rn" remote in the
464 ; $fg fork group for fetches
465 ; only uncomment the following iff you detach from fork groups
468 ; fetch = +refs/*:refs/*
471 close $fh or die "close($f): $!";
473 if (!$self->{dry_run}) {
474 my $alt = File::Spec->rel2abs("$dir/objects");
475 my $o = "$self->{cur_dst}/objects";
476 my $f = "$o/info/alternates";
477 my $l = File::Spec->abs2rel($alt, File::Spec->rel2abs($o));
478 open my $fh, '+>>', $f or die "open($f): $!";
479 seek($fh, SEEK_SET, 0) or die "seek($f): $!";
480 chomp(my @cur = <$fh>);
481 if (!grep(/\A\Q$l\E\z/, @cur)) {
482 say $fh $l or die "say($f): $!";
484 close $fh or die "close($f): $!";
487 %$self, -osdir => $dir, -remote => $rn, -uri => $uri
492 my ($self, $cmd, $cb, @arg) = @_;
494 $self->{lei}->err("@$cmd failed (\$?=$?) (non-fatal)");
495 $? = 0; # don't let it influence normal exit
497 return if !keep_going($self);
498 my $fh = delete $self->{-show_ref} // die 'BUG: no show-ref output';
499 seek($fh, SEEK_SET, 0) or die "seek(show_ref): $!";
500 $self->{-ent} // die 'BUG: no -ent';
501 my $A = $self->{-ent}->{fingerprint} // die 'BUG: no fingerprint';
502 my $B = sha1_hex(do { local $/; <$fh> } // die("read(show_ref): $!"));
503 return $cb->($self, @arg) if $A ne $B;
504 $self->{lei}->qerr("# $self->{-key} up-to-date");
508 my ($self, $cb, @arg) = @_;
509 # $cb is either resume_fetch or fgrp_enqueue
510 $self->{-ent} // return $cb->($self, @arg);
511 my $new = $self->{-ent}->{fingerprint} // return $cb->($self, @arg);
512 my $key = $self->{-key} // die 'BUG: no -key';
513 if (my $cur_ent = $self->{-local_manifest}->{$key}) {
514 # runs go_fetch->DESTROY run if eq
515 return if $cur_ent->{fingerprint} eq $new;
517 my $dst = $self->{cur_dst} // $self->{dst};
518 my $cmd = ['git', "--git-dir=$dst", 'show-ref'];
519 my $opt = { 2 => $self->{lei}->{2} };
520 open($opt->{1}, '+>', undef) or die "open(tmp): $!";
521 $self->{-show_ref} = $opt->{1};
523 $self->{lei}->qerr("# @$cmd");
524 $LIVE->{spawn($cmd, undef, $opt)} = [ \&fp_done, $self, $cmd,
529 my ($self, $uri, $fini) = @_;
530 return if !keep_going($self);
531 my $dst = $self->{cur_dst} // $self->{dst};
532 my @git = ('git', "--git-dir=$dst");
533 my $opt = { 2 => $self->{lei}->{2} };
534 my $rn = 'random'.int(rand(1 << 30));
535 for ("url=$uri", "fetch=+refs/*:refs/*", 'mirror=true') {
536 push @git, '-c', "remote.$rn.$_";
538 my $cmd = [ @{$self->{-torsocks}}, @git,
539 fetch_args($self->{lei}, $opt), $rn ];
540 push @$cmd, '-P' if $self->{lei}->{prune}; # --prune-tags implied
541 my $run_puh = PublicInbox::OnDestroy->new($$, \&run_puh, $self, $fini);
542 ++$self->{chg}->{nr_chg};
543 start_cmd($self, $cmd, $opt, $run_puh);
547 my ($fgrp, $end) = @_; # $end calls fgrp_fetch_all
548 return if !keep_going($fgrp);
549 my $opt = { 2 => $fgrp->{lei}->{2} };
550 # --no-tags is required to avoid conflicts
551 my $u = $fgrp->{-uri} // die 'BUG: no {-uri}';
552 my $rn = $fgrp->{-remote} // die 'BUG: no {-remote}';
553 my @cmd = ('git', "--git-dir=$fgrp->{-osdir}", 'config');
554 for ("url=$u", "fetch=+refs/*:refs/remotes/$rn/*", 'tagopt=--no-tags') {
555 my @kv = split(/=/, $_, 2);
556 $kv[0] = "remote.$rn.$kv[0]";
557 $fgrp->{dry_run} ? $fgrp->{lei}->qerr("# @cmd @kv") :
558 run_die([@cmd, @kv], undef, $opt);
560 ++$fgrp->{chg}->{nr_chg};
561 push @{$FGRP_TODO->{$fgrp->{-osdir}}}, $fgrp;
565 my ($self, $end) = @_;
566 my $lei = $self->{lei};
567 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
568 my $uri = URI->new($self->{cur_src} // $self->{src});
569 my $path = $uri->path;
570 $path =~ s!/*\z!! and $uri->path($path);
571 defined($lei->{opt}->{epoch}) and
572 die "$uri is a v1 inbox, --epoch is not supported\n";
573 $self->{-torsocks} //= $curl->torsocks($lei, $uri) or return;
574 my $dst = $self->{cur_dst} // $self->{dst};
575 my $fini = PublicInbox::OnDestroy->new($$, \&v1_done, $self);
576 my $resume = -d $dst;
577 if (my $fgrp = forkgroup_prep($self, $uri)) {
578 $fgrp->{-fini} = $fini;
579 $resume ? cmp_fp_do($fgrp, \&fgrp_enqueue, $end)
580 : fgrp_enqueue($fgrp, $end);
582 cmp_fp_do($self, \&resume_fetch, $uri, $fini);
583 } else { # normal clone
584 my $cmd = [ @{$self->{-torsocks}},
585 clone_cmd($lei, my $opt = {}), "$uri", $dst ];
586 if (defined($self->{-ent})) {
587 if (defined(my $ref = $self->{-ent}->{reference})) {
588 -e "$self->{dst}$ref" and
589 push @$cmd, '--reference',
593 ++$self->{chg}->{nr_chg};
594 start_cmd($self, $cmd, $opt, PublicInbox::OnDestroy->new($$,
595 \&run_puh, $self, $fini));
597 if (!$self->{-is_epoch} && $lei->{opt}->{'inbox-config'} =~
598 /\A(?:always|v1)\z/s) {
599 _get_txt_start($self, '_/text/config/raw', $fini);
602 my $d = $self->{-ent} ? $self->{-ent}->{description} : undef;
603 $self->{'txt.description'} = $d if defined $d;
604 (!defined($d) && !$end) and
605 _get_txt_start($self, 'description', $fini);
607 $end or do_reap($self, 1); # for non-manifest clone
610 sub parse_epochs ($$) {
611 my ($opt_epochs, $v2_epochs) = @_; # $epochs "LOW..HIGH"
612 $opt_epochs // return; # undef => all epochs
613 my ($lo, $dotdot, $hi, @extra) = split(/(\.\.)/, $opt_epochs);
614 undef($lo) if ($lo // '') eq '';
615 my $re = qr/\A~?[0-9]+\z/;
616 if (@extra || (($lo // '0') !~ $re) ||
617 (($hi // '0') !~ $re) ||
618 !(grep(defined, $lo, $hi))) {
620 --epoch=$opt_epochs not in the form of `LOW..HIGH', `LOW..', nor `..HIGH'
623 my @n = sort { $a <=> $b } keys %$v2_epochs;
624 for (grep(defined, $lo, $hi)) {
627 "`$_' exceeds maximum available epoch ($n[-1])\n";
629 "`$_' is lower than minimum available epoch ($n[0])\n";
630 } elsif (/\A~([0-9]+)/) {
632 $n[$off] // die "`$_' is out of range\n";
634 } else { die "`$_' not understood\n" }
636 defined($lo) && defined($hi) && $lo > $hi and die
637 "low value (`$lo') exceeds high (`$hi')\n";
638 $lo //= $n[0] if $dotdot;
639 $hi //= $n[-1] if $dotdot;
643 if (defined $v2_epochs->{$_}) {
647 "# epoch $_ is not available (non-fatal, $lo..$hi)\n";
653 sub init_placeholder ($$$) {
654 my ($src, $edst, $ent) = @_;
655 PublicInbox::Import::init_bare($edst);
656 my $f = "$edst/config";
657 open my $fh, '>>', $f or die "open($f): $!";
658 print $fh <<EOM or die "print($f): $!";
661 fetch = +refs/*:refs/*
664 ; This git epoch was created read-only and "public-inbox-fetch"
665 ; will not fetch updates for it unless write permission is added.
666 ; Hint: chmod +w $edst
668 if (defined($ent->{owner})) {
669 print $fh <<EOM or die "print($f): $!";
671 owner = $ent->{owner}
674 close $fh or die "close($f): $!";
675 my %map = (head => 'HEAD', description => undef);
676 while (my ($key, $fn) = each %map) {
677 my $val = $ent->{$key} // next;
680 open $fh, '>', $fn or die "open($fn): $!";
681 print $fh $val, "\n" or die "print($fn): $!";
682 close $fh or die "close($fn): $!";
686 sub reap_cmd { # async, called via SIGCHLD
687 my ($self, $cmd) = @_;
689 $? = 0; # don't let it influence normal exit
690 $self->{lei}->child_error($cerr, "@$cmd failed (\$?=$cerr)") if $cerr;
695 return if !keep_going($self);
696 my $fh = delete $self->{-show_ref_up} // die 'BUG: no show-ref output';
697 seek($fh, SEEK_SET, 0) or die "seek(show_ref): $!";
698 $self->{-ent} // die 'BUG: no -ent';
699 my $A = $self->{-ent}->{fingerprint} // die 'BUG: no fingerprint';
700 my $B = sha1_hex(do { local $/; <$fh> } // die("read(show_ref): $!"));
702 $self->{-ent}->{fingerprint} = $B;
703 push @{$self->{chg}->{fp_mismatch}}, $self->{-key};
706 sub atomic_write ($$$) {
707 my ($dn, $bn, $raw) = @_;
708 my $ft = File::Temp->new(DIR => $dn, TEMPLATE => "$bn-XXXX");
709 print $ft $raw or die "print($ft): $!";
710 $ft->flush or die "flush($ft): $!";
711 ft_rename($ft, "$dn/$bn", 0666);
716 my $puh = shift @{$self->{-puh_todo}} // return delete($self->{-fini});
717 my $fini = PublicInbox::OnDestroy->new($$, \&run_next_puh, $self);
718 my $cmd = [ @$puh, ($self->{cur_dst} // $self->{dst}) ];
719 my $opt = +{ map { $_ => $self->{lei}->{$_} } (0..2) };
720 start_cmd($self, $cmd, undef, $opt, $fini);
724 my ($self, $fini) = @_;
725 $self->{-fini} = $fini;
726 @{$self->{-puh_todo}} = @PUH;
730 # modifies the to-be-written manifest entry, and sets values from it, too
733 my $key = $self->{-key} // die 'BUG: no -key';
734 my $new = $self->{-ent}->{fingerprint};
735 my $cur = $self->{-local_manifest}->{$key}->{fingerprint} // "\0";
736 my $dst = $self->{cur_dst} // $self->{dst};
737 if (defined($new) && $new ne $cur) {
738 my $cmd = ['git', "--git-dir=$dst", 'show-ref'];
739 my $opt = { 2 => $self->{lei}->{2} };
740 open($opt->{1}, '+>', undef) or die "open(tmp): $!";
741 $self->{-show_ref_up} = $opt->{1};
742 my $done = PublicInbox::OnDestroy->new($$, \&up_fp_done, $self);
743 start_cmd($self, $cmd, $opt, $done);
745 $new = $self->{-ent}->{head};
746 $cur = $self->{-local_manifest}->{$key}->{head} // "\0";
747 if (defined($new) && $new ne $cur) {
748 # n.b. grokmirror writes raw contents to $dst/HEAD w/o locking
749 my $cmd = [ 'git', "--git-dir=$dst" ];
750 if ($new =~ s/\Aref: //) {
751 push @$cmd, qw(symbolic-ref HEAD), $new;
752 } elsif ($new =~ /\A[a-f0-9]{40,}\z/) {
753 push @$cmd, qw(update-ref --no-deref HEAD), $new;
756 warn "W: $key: {head} => `$new' not understood\n";
758 start_cmd($self, $cmd, { 2 => $self->{lei}->{2} }) if $cmd;
760 if (my $symlinks = $self->{-ent}->{symlinks}) {
761 my $top = File::Spec->rel2abs($self->{dst});
762 push @{$self->{-new_symlinks}}, @$symlinks;
763 for my $p (@$symlinks) {
766 my (undef, $dn, $bn) = File::Spec->splitpath($ln);
767 File::Path::mkpath($dn);
768 my $tgt = "$top/$key";
769 $tgt = File::Spec->abs2rel($tgt, $dn);
772 next if readlink($ln) eq $tgt;
773 unlink($ln) or die "unlink($ln): $!";
775 push @{$self->{chg}->{badlink}}, $p;
778 symlink($tgt, $ln) or die "symlink($tgt, $ln): $!";
779 ++$self->{chg}->{nr_chg};
782 if (defined(my $t = $self->{-ent}->{modified})) {
783 my ($dn, $bn) = ("$dst/info/web", 'last-modified');
784 my $orig = PublicInbox::Git::try_cat("$dn/$bn");
785 $t = strftime('%F %T', gmtime($t))." +0000\n";
786 File::Path::mkpath($dn);
787 atomic_write($dn, $bn, $t) if $orig ne $t;
790 $new = $self->{-ent}->{owner} // return;
791 $cur = $self->{-local_manifest}->{$key}->{owner} // "\0";
792 return if $cur eq $new;
793 my $cmd = [ qw(git config -f), "$dst/config", 'gitweb.owner', $new ];
794 start_cmd($self, $cmd, { 2 => $self->{lei}->{2} });
797 sub v1_done { # called via OnDestroy
799 return if $self->{dry_run} || !keep_going($self);
800 _write_inbox_config($self);
801 my $dst = $self->{cur_dst} // $self->{dst};
802 update_ent($self) if $self->{-ent};
803 my $o = "$dst/objects";
804 if (open(my $fh, '<', my $fn = "$o/info/alternates")) {;
805 my $base = File::Spec->rel2abs($o);
810 $_ = File::Spec->abs2rel($_, $base);
811 $ft //= File::Temp->new(TEMPLATE => '.XXXX',
815 print $ft @l or die "print($ft): $!";
816 $ft->flush or die "flush($ft): $!";
817 ft_rename($ft, $fn, 0666, $fh);
820 eval { set_description($self) };
822 return if ($self->{-is_epoch} ||
823 $self->{lei}->{opt}->{'inbox-config'} ne 'always');
824 write_makefile($dst, 1);
825 index_cloned_inbox($self, 1);
828 sub v2_done { # called via OnDestroy
830 return if $self->{dry_run} || !keep_going($self);
831 my $dst = $self->{cur_dst} // $self->{dst};
832 require PublicInbox::Lock;
833 my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
834 my $lck = $lk->lock_for_scope($$);
835 _write_inbox_config($self);
836 require PublicInbox::MultiGit;
837 my $mg = PublicInbox::MultiGit->new($dst, 'all.git', 'git');
838 $mg->fill_alternates;
839 for my $i ($mg->git_epochs) { $mg->epoch_cfg_set($i) }
840 for my $edst (@{delete($self->{-read_only}) // []}) {
841 my @st = stat($edst) or die "stat($edst): $!";
842 chmod($st[2] & 0555, $edst) or die "chmod(a-w, $edst): $!";
844 write_makefile($dst, 2);
846 eval { set_description($self) };
848 index_cloned_inbox($self, 2);
851 sub clone_v2_prep ($$;$) {
852 my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
853 my $lei = $self->{lei};
854 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
855 my $first_uri = (map { $_->[0] } values %$v2_epochs)[0];
856 $self->{-torsocks} //= $curl->torsocks($lei, $first_uri) or return;
857 my $dst = $self->{cur_dst} // $self->{dst};
858 my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
859 my $task = $m ? bless { %$self }, __PACKAGE__ : $self;
861 my $fini = PublicInbox::OnDestroy->new($$, \&v2_done, $task);
862 for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
863 my ($uri, $key) = @{$v2_epochs->{$nr}};
864 my $src = $uri->as_string;
866 $src =~ m!/([0-9]+)(?:\.git)?\z! or die <<"";
867 failed to extract epoch number from $src
869 $1 + 0 == $nr or die "BUG: <$uri> miskeyed $1 != $nr";
870 $edst .= "/git/$nr.git";
874 die("BUG: `$key' not in manifest.js.gz");
875 if (defined(my $d = $ent->{description})) {
876 $d =~ s/ \[epoch [0-9]+\]\z//s;
880 if (!$want || $want->{$nr}) {
881 my $etask = bless { %$task, -key => $key }, __PACKAGE__;
882 $etask->{-ent} = $ent; # may have {reference}
883 $etask->{cur_src} = $src;
884 $etask->{cur_dst} = $edst;
885 $etask->{-is_epoch} = $fini;
886 my $ref = $ent->{reference} // '';
887 push @{$TODO->{$ref}}, $etask;
888 $self->{any_want}->{$key} = 1;
889 } else { # create a placeholder so users only need to chmod +w
890 init_placeholder($src, $edst, $ent);
891 push @{$task->{-read_only}}, $edst;
895 # filter out the epochs we skipped
896 $self->{chg}->{manifest} = 1 if $m && delete(@$m{@skip});
898 (!$self->{dry_run} && !-d $dst) and File::Path::mkpath($dst);
900 $lei->{opt}->{'inbox-config'} =~ /\A(?:always|v2)\z/s and
901 _get_txt_start($task, '_/text/config/raw', $fini);
903 defined($desc) ? ($task->{'txt.description'} = $desc) :
904 _get_txt_start($task, 'description', $fini);
907 sub decode_manifest ($$$) {
908 my ($fh, $fn, $uri) = @_;
910 my $gz = do { local $/; <$fh> } // die "slurp($fn): $!";
911 gunzip(\$gz => \$js, MultiStream => 1) or
912 die "gunzip($uri): $GunzipError\n";
913 my $m = eval { PublicInbox::Config->json->decode($js) };
914 die "$uri: error decoding `$js': $@\n" if $@;
915 ref($m) eq 'HASH' or die "$uri unknown type: ".ref($m);
919 sub load_current_manifest ($) {
921 my $fn = $self->{-manifest} // return;
922 if (open(my $fh, '<', $fn)) {
923 decode_manifest($fh, $fn, $fn);
924 } elsif ($!{ENOENT}) { # non-fatal, we can just do it slowly
925 warn "open($fn): $!\n" if !$self->{-initial_clone};
928 die "open($fn): $!\n";
932 sub multi_inbox ($$$) {
933 my ($self, $path, $m) = @_;
934 my $incl = $self->{lei}->{opt}->{include};
935 my $excl = $self->{lei}->{opt}->{exclude};
937 # assuming everything not v2 is v1, for now
938 my @v1 = sort grep(!m!.+/git/[0-9]+\.git\z!, keys %$m);
939 my @v2_epochs = sort grep(m!.+/git/[0-9]+\.git\z!, keys %$m);
943 m!\A(/.+)/git/[0-9]+\.git\z! or die "BUG: $_";
944 push @{$v2->{$1}}, $_;
946 my $n = scalar(keys %$v2) + scalar(@v1);
947 my @orig = defined($incl // $excl) ? (keys %$v2, @v1) : ();
949 my $re = '(?:'.join('\\z|', map {
950 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
952 my @gone = delete @$v2{grep(!/$re/, keys %$v2)};
953 delete @$m{map { @$_ } @gone} and $self->{chg}->{manifest} = 1;
954 delete @$m{grep(!/$re/, @v1)} and $self->{chg}->{manifest} = 1;
955 @v1 = grep(/$re/, @v1);
958 my $re = '(?:'.join('\\z|', map {
959 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
961 my @gone = delete @$v2{grep(/$re/, keys %$v2)};
962 delete @$m{map { @$_ } @gone} and $self->{chg}->{manifest} = 1;
963 delete @$m{grep(/$re/, @v1)} and $self->{chg}->{manifest} = 1;
964 @v1 = grep(!/$re/, @v1);
966 my $ret; # { v1 => [ ... ], v2 => { "/$inbox_name" => [ epochs ] }}
967 $ret->{v1} = \@v1 if @v1;
968 $ret->{v2} = $v2 if keys %$v2;
969 $ret //= @orig ? "Nothing to clone, available repositories:\n\t".
970 join("\n\t", sort @orig)
971 : "Nothing available to clone\n";
974 # PSGI mount prefixes and manifest.js.gz prefixes don't always align...
976 until (grep(m!\A\Q$$path\E/git/[0-9]+\.git\z!,
977 @v2_epochs) == @v2_epochs) {
978 $$path =~ s!\A(/[^/]+)/!/! or last;
982 while (!defined($m->{$$path}) && $$path =~ s!\A(/[^/]+)/!/!) {
986 ($path_pfx, $n, $ret);
992 $TODO = \'BUG on further use';
993 my $end = PublicInbox::OnDestroy->new($$, \&fgrp_fetch_all, $self);
995 my $nodep = delete $todo->{''};
997 # do not download unwanted deps
998 my $any_want = delete $self->{any_want};
999 my @unwanted = grep { !$any_want->{$_} } keys %$todo;
1000 my @nodep = delete(@$todo{@unwanted});
1001 push(@$nodep, @$_) for @nodep;
1003 # handle no-dependency repos, first
1006 return if !keep_going($self);
1009 # resolve references, deepest, first:
1010 while (scalar keys %$todo) {
1011 for my $x (keys %$todo) {
1013 # resolve multi-level references
1014 while ($m && defined($nxt = $m->{$x}->{reference})) {
1015 exists($todo->{$nxt}) or last;
1017 $m->{$x}->{reference} = undef;
1018 $m->{$nxt}->{reference} = undef;
1020 E: dependency loop detected (`$x' => `$nxt'), breaking
1025 my $y = delete $todo->{$x} // next; # already done
1028 return if !keep_going($self);
1030 last; # restart %$todo iteration
1034 # $end->DESTROY will call fgrp_fetch_all once all references
1035 # in $LIVE are gone, and do_reap will eventually drain $LIVE
1040 sub dump_manifest ($$) {
1042 # write the smaller manifest if epochs were skipped so
1043 # users won't have to delete manifest if they +w an
1044 # epoch they no longer want to skip
1045 my $json = PublicInbox::Config->json->encode($m);
1046 my $mtime = (stat($ft))[9];
1047 seek($ft, SEEK_SET, 0) or die "seek($ft): $!";
1048 truncate($ft, 0) or die "truncate($ft): $!";
1049 gzip(\$json => $ft) or die "gzip($ft): $GzipError";
1050 $ft->flush or die "flush($ft): $!";
1051 utime($mtime, $mtime, "$ft") or die "utime(..., $ft): $!";
1054 sub dump_project_list ($$) {
1055 my ($self, $m) = @_;
1056 my $f = $self->{'-project-list'} // return;
1057 my $old = PublicInbox::Git::try_cat($f);
1060 open my $dh, '<', '.' or die "open(.): $!";
1061 chdir($self->{dst}) or die "chdir($self->{dst}): $!";
1062 my @local = grep { -e $_ ? ($new{$_} = undef) : 1 } split(/\n/s, $old);
1063 chdir($dh) or die "chdir(restore): $!";
1065 $new{substr($_, 1)} = 1 for keys %$m; # drop leading '/'
1066 my @list = sort keys %new;
1067 my @remote = grep { !defined($new{$_}) } @list;
1068 my %lnk = map { substr($_, 1) => undef } @{$self->{-new_symlinks}};
1069 @remote = grep { !exists($lnk{$_}) } @remote;
1071 warn <<EOM if @remote;
1072 The following local repositories are ignored/gone from $self->{src}:
1074 warn "\t", $_, "\n" for @remote;
1075 warn <<EOM if @local;
1076 The following repos in $f no longer exist on the filesystem:
1078 warn "\t", $_, "\n" for @local;
1080 my (undef, $dn, $bn) = File::Spec->splitpath($f);
1081 $self->{chg}->{nr_chg} += scalar(@remote) + scalar(@local);
1082 my $new = join("\n", @list, '');
1083 atomic_write($dn, $bn, $new) if $new ne $old;
1086 # FIXME: this gets confused by single inbox instance w/ global manifest.js.gz
1089 my $uri = URI->new($self->{src});
1090 my $lei = $self->{lei};
1091 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
1092 $self->{-torsocks} //= $curl->torsocks($lei, $uri) or return;
1093 my $path = $uri->path;
1094 chop($path) eq '/' or die "BUG: $uri not canonicalized";
1095 $uri->path($path . '/manifest.js.gz');
1096 my $manifest = $self->{-manifest} // "$self->{dst}/manifest.js.gz";
1097 my %opt = (UNLINK => 1, SUFFIX => '.tmp', TMPDIR => 1);
1098 if (!$self->{dry_run} && $manifest =~ m!\A(.+?)/[^/]+\z! and -d $1) {
1099 $opt{DIR} = $1; # allows fast rename(2) w/o EXDEV
1100 delete $opt{TMPDIR};
1102 my $ft = File::Temp->new(TEMPLATE => '.manifest-XXXX', %opt);
1103 my $cmd = $curl->for_uri($lei, $uri, qw(-R -o), $ft->filename);
1104 push(@$cmd, '-z', $manifest) if -f $manifest;
1105 my $mf_url = "$uri";
1106 %opt = map { $_ => $lei->{$_} } (0..2);
1107 my $cerr = run_reap($lei, $cmd, \%opt);
1109 return try_scrape($self) if ($cerr >> 8) == 22; # 404 missing
1110 return $lei->child_error($cerr, "@$cmd failed");
1113 # bail out if curl -z/--timecond hit 304 Not Modified, $ft will be empty
1114 if (-f $manifest && !-s $ft) {
1115 $lei->child_error(127 << 8) if $lei->{opt}->{'exit-code'};
1116 return $lei->qerr("# $manifest unchanged");
1119 my $m = eval { decode_manifest($ft, $ft, $uri) };
1122 return try_scrape($self);
1124 local $self->{chg} = {};
1125 local $self->{-local_manifest} = load_current_manifest($self);
1126 local $self->{-new_symlinks} = [];
1127 my ($path_pfx, $n, $multi) = multi_inbox($self, \$path, $m);
1128 return $lei->child_error(1, $multi) if !ref($multi);
1129 my $v2 = delete $multi->{v2};
1131 for my $name (sort keys %$v2) {
1132 my $epochs = delete $v2->{$name};
1133 my %v2_epochs = map {
1134 $uri->path($n > 1 ? $path_pfx.$path.$_
1136 my ($e) = ("$uri" =~ m!/([0-9]+)\.git\z!);
1137 $e // die "no [0-9]+\.git in `$uri'";
1138 $e => [ $uri->clone, $_ ];
1140 ("$uri" =~ m!\A(.+/)git/[0-9]+\.git\z!) or
1141 die "BUG: `$uri' !~ m!/git/[0-9]+.git!";
1142 local $self->{cur_src} = $1;
1143 local $self->{cur_dst} = $self->{dst};
1144 if ($n > 1 && $uri->path =~ m!\A\Q$path_pfx$path\E/(.+)/
1145 git/[0-9]+\.git\z!x) {
1146 $self->{cur_dst} .= "/$1";
1148 index($self->{cur_dst}, "\n") >= 0 and die <<EOM;
1149 E: `$self->{cur_dst}' must not contain newline
1151 clone_v2_prep($self, \%v2_epochs, $m);
1152 return if !keep_going($self);
1155 if (my $v1 = delete $multi->{v1}) {
1156 my $p = $path_pfx.$path;
1157 chop($p) if substr($p, -1, 1) eq '/';
1159 for my $name (@$v1) {
1160 my $task = bless { %$self }, __PACKAGE__;
1161 $task->{-ent} = $m->{$name} //
1162 die("BUG: no `$name' in manifest");
1163 $task->{cur_src} = "$uri";
1164 $task->{cur_dst} = $task->{dst};
1165 $task->{-key} = $name;
1167 $task->{cur_dst} .= $name;
1168 $task->{cur_src} .= $name;
1170 index($task->{cur_dst}, "\n") >= 0 and die <<EOM;
1171 E: `$task->{cur_dst}' must not contain newline
1173 $task->{cur_src} .= '/';
1174 my $dep = $task->{-ent}->{reference} // '';
1175 push @{$TODO->{$dep}}, $task; # for clone_all
1176 $self->{any_want}->{$name} = 1;
1179 delete local $lei->{opt}->{epoch} if defined($v2);
1180 clone_all($self, $m);
1181 return if $self->{dry_run} || !keep_going($self);
1183 # set by clone_v2_prep/-I/--exclude
1184 my $mis = delete $self->{chg}->{fp_mismatch};
1186 my $t = (stat($ft))[9];
1187 $t = strftime('%F %k:%M:%S %z', localtime($t));
1189 W: Fingerprints for the following repositories do not match
1191 W: These repositories may have updated since $t:
1193 warn "\t", $_, "\n" for @$mis;
1194 warn <<EOM if !$self->{lei}->{opt}->{prune};
1195 W: The above fingerprints may never match without --prune
1198 dump_manifest($m => $ft) if delete($self->{chg}->{manifest}) || $mis;
1199 my $bad = delete $self->{chg}->{badlink};
1200 warn(<<EOM, map { ("\t", $_, "\n") } @$bad) if $bad;
1201 W: The following exist and have not been converted to symlinks
1203 dump_project_list($self, $m);
1204 ft_rename($ft, $manifest, 0666);
1205 !$self->{chg}->{nr_chg} && $lei->{opt}->{'exit-code'} and
1206 $lei->child_error(127 << 8);
1209 sub start_clone_url {
1211 return try_manifest($self) if $self->{src} =~ m!\Ahttps?://!;
1212 die "TODO: non-HTTP/HTTPS clone of $self->{src} not supported, yet";
1215 sub do_mirror { # via wq_io_do or public-inbox-clone
1217 my $lei = $self->{lei};
1218 $self->{dry_run} = 1 if $lei->{opt}->{'dry-run'};
1219 umask($lei->{client_umask}) if defined $lei->{client_umask};
1220 $self->{-initial_clone} = 1 if !-d $self->{dst};
1222 if (defined(my $puh = $lei->{opt}->{'post-update-hook'})) {
1223 require Text::ParseWords;
1224 @PUH = map { [ Text::ParseWords::shellwords($_) ] } @$puh;
1227 my $ic = $lei->{opt}->{'inbox-config'} //= 'always';
1228 $ic =~ /\A(?:v1|v2|always|never)\z/s or die <<"";
1229 --inbox-config must be one of `always', `v2', `v1', or `never'
1231 # we support these switches with '' (empty string).
1232 # defaults match example conf distributed with grokmirror
1233 my @pairs = qw(objstore objstore manifest manifest.js.gz
1234 project-list projects.list);
1236 my ($k, $default) = splice(@pairs, 0, 2);
1237 my $v = $lei->{opt}->{$k} // next;
1238 $v = $default if $v eq '';
1239 $v = "$self->{dst}/$v" if $v !~ m!\A\.{0,2}/!;
1240 $self->{"-$k"} = $v;
1245 local $FGRP_TODO = {};
1246 my $iv = $lei->{opt}->{'inbox-version'} //
1247 return start_clone_url($self);
1248 return clone_v1($self) if $iv == 1;
1249 die "bad --inbox-version=$iv\n" if $iv != 2;
1250 die <<EOM if $self->{src} !~ m!://!;
1251 cloning local v2 inboxes not supported
1253 try_scrape($self, 1);
1255 $lei->fail($@) if $@;
1259 my ($cls, $lei, $src, $dst) = @_;
1260 my $self = bless { src => $src, dst => $dst }, $cls;
1261 $lei->request_umask;
1262 my ($op_c, $ops) = $lei->workers_start($self, 1);
1263 $lei->{wq1} = $self;
1264 $self->wq_io_do('do_mirror', []);
1266 $lei->wait_wq_events($op_c, $ops);
1269 sub ipc_atfork_child {
1271 $self->{lei}->_lei_atfork_child;
1272 $self->SUPER::ipc_atfork_child;
1275 sub write_makefile {
1276 my ($dir, $ibx_ver) = @_;
1277 my $f = "$dir/Makefile";
1278 if (sysopen my $fh, $f, O_CREAT|O_EXCL|O_WRONLY) {
1279 print $fh <<EOM or die "print($f) $!";
1280 # This is a v$ibx_ver public-inbox, see the public-inbox-v$ibx_ver-format(5)
1281 # manpage for more information on the format. This Makefile is
1282 # intended as a familiar wrapper for users unfamiliar with
1283 # public-inbox-* commands.
1285 # See the respective manpages for public-inbox-fetch(1),
1286 # public-inbox-index(1), etc for more information on
1287 # some of the commands used by this Makefile.
1289 # This Makefile will not be modified nor read by public-inbox,
1290 # so you may edit it freely with your own convenience targets
1291 # and notes. public-inbox-fetch will recreate it if removed.
1293 print $fh <<'EOM' or die "print($f): $!";
1294 # the default target:
1296 @echo Common targets:
1297 @echo ' make fetch - fetch from remote git repostorie(s)'
1298 @echo ' make update - fetch and update index '
1300 @echo Rarely needed targets:
1301 @echo ' make reindex - may be needed for new features/bugfixes'
1302 @echo ' make compact - rewrite Xapian storage to save space'
1303 @echo ' make index - initial index after clone
1308 @if ! public-inbox-fetch --exit-code; \
1311 test $$c -eq 127 && exit 0; \
1313 elif test -f msgmap.sqlite3 || test -f public-inbox/msgmap.sqlite3; \
1315 public-inbox-index; \
1317 echo 'public-inbox index not initialized'; \
1318 echo 'see public-inbox-index(1) man page'; \
1323 public-inbox-index --reindex
1325 public-inbox-compact
1327 .PHONY : help fetch update index reindex compact
1329 close $fh or die "close($f): $!";
1331 die "open($f): $!" unless $!{EEXIST};