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 sha1_hex);
22 use POSIX qw(strftime);
24 our $LIVE; # pid => callback
25 our $FGRP_TODO; # objstore -> [ fgrp mirror objects ]
26 our $TODO; # reference => [ non-fgrp mirror objects ]
29 $LIVE && (!$_[0]->{lei}->{child_error} ||
30 $_[0]->{lei}->{opt}->{'keep-going'});
33 sub _wq_done_wait { # dwaitpid callback (via wq_eof)
35 my ($mrr, $lei) = @$arg;
37 $lei->child_error($?);
38 } elsif (!$lei->{child_error}) {
39 if (!$mrr->{dry_run} && $lei->{cmd} ne 'public-inbox-clone') {
40 require PublicInbox::LeiAddExternal;
41 PublicInbox::LeiAddExternal::_finish_add_external(
44 $lei->qerr("# mirrored $mrr->{src} => $mrr->{dst}");
49 # for old installations without manifest.js.gz
51 my ($self, $fallback_manifest) = @_;
52 my $uri = URI->new($self->{src});
53 my $lei = $self->{lei};
54 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
55 my $cmd = $curl->for_uri($lei, $uri, '--compressed');
56 my $opt = { 0 => $lei->{0}, 2 => $lei->{2} };
57 my $fh = popen_rd($cmd, undef, $opt);
58 my $html = do { local $/; <$fh> } // die "read(curl $uri): $!";
59 close($fh) or return $lei->child_error($?, "@$cmd failed");
61 # we grep with URL below, we don't want Subject/From headers
62 # making us clone random URLs. This assumes remote instances
63 # prior to public-inbox 1.7.0
64 # 5b96edcb1e0d8252 (www: move mirror instructions to /text/, 2021-08-28)
65 my @html = split(/<hr>/, $html);
66 my @urls = ($html[-1] =~ m!\bgit clone --mirror ([a-z\+]+://\S+)!g);
67 if (!@urls && $fallback_manifest) {
69 W: failed to extract URLs from $uri, trying manifest.js.gz...
71 return start_clone_url($self);
73 my $url = $uri->as_string;
74 chop($url) eq '/' or die "BUG: $uri not canonicalized";
76 # since this is for old instances w/o manifest.js.gz, try v1 first
77 return clone_v1($self) if grep(m!\A\Q$url\E/*\z!, @urls);
78 if (my @v2_urls = grep(m!\A\Q$url\E/[0-9]+\z!, @urls)) {
80 my ($n) = (m!/([0-9]+)\z!);
81 $n => [ URI->new($_), '' ]
83 clone_v2_prep($self, \%v2_epochs);
84 delete local $lei->{opt}->{epoch};
89 # filter out common URLs served by WWW (e.g /$MSGID/T/)
90 if (@urls && $url =~ s!/+[^/]+\@[^/]+/.*\z!! &&
91 grep(m!\A\Q$url\E/*\z!, @urls)) {
93 E: confused by scraping <$uri>, did you mean <$url>?
97 E: confused by scraping <$uri>, got ambiguous results:
100 die "E: scraping <$uri> revealed nothing\n";
104 my ($lei, $opt) = @_;
106 $opt->{$_} = $lei->{$_} for (0..2);
107 # we support "-c $key=$val" for arbitrary git config options
108 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
109 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
110 push @cmd, qw(clone --mirror);
111 push @cmd, '-q' if $lei->{opt}->{quiet} ||
112 ($lei->{opt}->{jobs} // 1) > 1;
113 push @cmd, '-v' if $lei->{opt}->{verbose};
114 # XXX any other options to support?
115 # --reference is tricky with multiple epochs, but handled
116 # automatically if using manifest.js.gz
120 sub ft_rename ($$$;$) {
121 my ($ft, $dst, $open_mode, $fh) = @_;
122 my @st = stat($fh // $dst);
123 my $mode = @st ? ($st[2] & 07777) : ($open_mode & ~umask);
124 chmod($mode, $ft) or croak "E: chmod($ft): $!";
126 File::Copy::mv($ft->filename, $dst) or croak "E: mv($ft => $dst): $!";
127 $ft->unlink_on_destroy(0);
131 my ($self, $jobs) = @_;
132 $jobs //= $self->{-jobs} //= $self->{lei}->{opt}->{jobs} // 1;
133 $jobs = 1 if $jobs < 1;
134 while (keys(%$LIVE) >= $jobs) {
135 my $pid = waitpid(-1, 0) // die "waitpid(-1): $!";
136 if (my $x = delete $LIVE->{$pid}) {
140 warn "reaped unknown PID=$pid ($?)\n";
145 sub _get_txt_start { # non-fatal
146 my ($self, $endpoint, $fini) = @_;
147 my $uri = URI->new($self->{cur_src} // $self->{src});
148 my $lei = $self->{lei};
149 my $path = $uri->path;
150 chop($path) eq '/' or die "BUG: $uri not canonicalized";
151 $uri->path("$path/$endpoint");
152 my $f = (split(m!/!, $endpoint))[-1];
153 my $ft = File::Temp->new(TEMPLATE => "$f-XXXX", TMPDIR => 1);
154 my $opt = { 0 => $lei->{0}, 1 => $lei->{1}, 2 => $lei->{2} };
155 my $cmd = $self->{curl}->for_uri($lei, $uri, qw(--compressed -R -o),
158 $lei->qerr("# @$cmd");
159 return if $self->{dry_run};
160 $self->{"-get_txt.$endpoint"} = [ $ft, $cmd, $uri ];
161 $LIVE->{spawn($cmd, undef, $opt)} =
162 [ \&_get_txt_done, $self, $endpoint, $fini ];
165 sub _get_txt_done { # returns true on error (non-fatal), undef on success
166 my ($self, $endpoint) = @_;
167 my ($fh, $cmd, $uri) = @{delete $self->{"-get_txt.$endpoint"}};
169 $? = 0; # don't influence normal lei exit
170 return warn("$uri missing\n") if ($cerr >> 8) == 22;
171 return warn("# @$cmd failed (non-fatal)\n") if $cerr;
172 seek($fh, SEEK_SET, 0) or die "seek: $!";
173 $self->{"mtime.$endpoint"} = (stat($fh))[9];
175 $self->{"txt.$endpoint"} = <$fh>;
179 sub _write_inbox_config {
181 my $buf = delete($self->{'txt._/text/config/raw'}) // return;
182 my $dst = $self->{cur_dst} // $self->{dst};
183 my $f = "$dst/inbox.config.example";
184 my $mtime = delete $self->{'mtime._/text/config/raw'};
185 if (sysopen(my $fh, $f, O_CREAT|O_EXCL|O_WRONLY)) {
186 print $fh $buf or die "print: $!";
187 chmod(0444 & ~umask, $fh) or die "chmod($f): $!";
188 $fh->flush or die "flush($f): $!";
189 if (defined $mtime) {
190 utime($mtime, $mtime, $fh) or die "utime($f): $!";
192 } elsif (!$!{EEXIST}) {
195 my $cfg = PublicInbox::Config->git_config_dump($f, $self->{lei}->{2});
196 my $ibx = $self->{ibx} = {};
197 for my $sec (grep(/\Apublicinbox\./, @{$cfg->{-section_order}})) {
198 for (qw(address newsgroup nntpmirror)) {
199 $ibx->{$_} = $cfg->{"$sec.$_"};
204 sub set_description ($) {
206 my $dst = $self->{cur_dst} // $self->{dst};
207 chomp(my $orig = PublicInbox::Git::try_cat("$dst/description"));
209 while (defined($d) && ($d =~ m!^\(\$INBOX_DIR/description missing\)! ||
210 $d =~ /^Unnamed repository/ || $d !~ /\S/)) {
211 $d = delete($self->{'txt.description'});
213 $d //= 'mirror of '.($self->{cur_src} // $self->{src});
214 atomic_write($dst, 'description', $d."\n") if $d ne $orig;
217 sub index_cloned_inbox {
218 my ($self, $iv) = @_;
219 my $lei = $self->{lei};
221 # n.b. public-inbox-clone works w/o (SQLite || Xapian)
222 # lei is useless without Xapian + SQLite
223 if ($lei->{cmd} ne 'public-inbox-clone') {
224 require PublicInbox::InboxWritable;
225 require PublicInbox::Admin;
226 my $ibx = delete($self->{ibx}) // {
227 address => [ 'lei@example.com' ],
230 $ibx->{inboxdir} = $self->{cur_dst} // $self->{dst};
231 PublicInbox::Inbox->new($ibx);
232 PublicInbox::InboxWritable->new($ibx);
234 for my $sw ($lei->index_opt) {
235 my ($k) = ($sw =~ /\A([\w-]+)/);
236 $opt->{$k} = $lei->{opt}->{$k};
238 # force synchronous dwaitpid for v2:
239 local $PublicInbox::DS::in_loop = 0;
240 my $cfg = PublicInbox::Config->new(undef, $lei->{2});
241 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
242 local %ENV = (%ENV, %$env) if $env;
243 PublicInbox::Admin::progress_prepare($opt, $lei->{2});
244 PublicInbox::Admin::index_inbox($ibx, undef, $opt);
246 return if defined $self->{cur_dst}; # one of many repos to clone
250 my ($lei, $cmd, $opt) = @_;
251 $lei->qerr("# @$cmd");
252 waitpid(spawn($cmd, undef, $opt), 0) // die "waitpid: $!";
254 $? = 0; # don't let it influence normal exit
259 my ($self, $cmd, $opt, $fini) = @_;
261 $self->{lei}->qerr("# @$cmd");
262 return if $self->{dry_run};
263 $LIVE->{spawn($cmd, undef, $opt)} = [ \&reap_cmd, $self, $cmd, $fini ]
266 sub fetch_args ($$) {
267 my ($lei, $opt) = @_;
268 my @cmd; # (git --git-dir=...) to be added by caller
269 $opt->{$_} = $lei->{$_} for (0..2);
270 # we support "-c $key=$val" for arbitrary git config options
271 # e.g.: git -c http.proxy=socks5h://127.0.0.1:9050
272 push(@cmd, '-c', $_) for @{$lei->{opt}->{c} // []};
274 push @cmd, '-q' if $lei->{opt}->{quiet} ||
275 ($lei->{opt}->{jobs} // 1) > 1;
276 push @cmd, '-v' if $lei->{opt}->{verbose};
277 push(@cmd, '-p') if $lei->{opt}->{prune};
281 sub upr { # feed `git update-ref --stdin -z' verbosely
282 my ($lei, $w, $op, @rest) = @_; # ($ref, $oid) = @rest
283 $lei->qerr("# $op @rest") if $lei->{opt}->{verbose};
284 print $w "$op ", join("\0", @rest, '') or die "print(w): $!";
289 return if !keep_going($fgrp);
290 my $srcfh = delete $fgrp->{srcfh} or return;
291 my $dstfh = delete $fgrp->{dstfh} or return;
292 seek($srcfh, SEEK_SET, 0) or die "seek(src): $!";
293 seek($dstfh, SEEK_SET, 0) or die "seek(dst): $!";
294 my %src = map { chomp; split(/\0/) } (<$srcfh>);
296 my %dst = map { chomp; split(/\0/) } (<$dstfh>);
298 pipe(my ($r, $w)) or die "pipe: $!";
299 my $cmd = [ 'git', "--git-dir=$fgrp->{cur_dst}",
300 qw(update-ref --stdin -z) ];
301 my $lei = $fgrp->{lei};
302 my $pack = PublicInbox::OnDestroy->new($$, \&pack_dst, $fgrp);
303 start_cmd($fgrp, $cmd, { 0 => $r, 2 => $lei->{2} }, $pack);
304 close $r or die "close(r): $!";
305 return if $fgrp->{dry_run};
306 for my $ref (keys %dst) {
307 my $new = delete $src{$ref};
308 my $old = $dst{$ref};
311 upr($lei, $w, 'update', $ref, $new, $old);
313 upr($lei, $w, 'delete', $ref, $old);
316 while (my ($ref, $oid) = each %src) {
317 upr($lei, $w, 'create', $ref, $oid);
319 close($w) or warn "E: close(update-ref --stdin): $! (need git 1.8.5+)\n";
322 sub pack_dst { # packs lightweight satellite repos
324 pack_refs($fgrp, $fgrp->{cur_dst});
328 my ($self, $git_dir) = @_;
329 my $cmd = [ 'git', "--git-dir=$git_dir", qw(pack-refs --all --prune) ];
330 start_cmd($self, $cmd, { 2 => $self->{lei}->{2} });
336 my $first = $fgrpv->[0] // die 'BUG: no fgrpv->[0]';
337 return if !keep_going($first);
338 pack_refs($first, $first->{-osdir}); # objstore refs always packed
339 for my $fgrp (@$fgrpv) {
340 my $rn = $fgrp->{-remote};
341 my %opt = ( 2 => $fgrp->{lei}->{2} );
343 my $update_ref = PublicInbox::OnDestroy->new($$,
344 \&fgrp_update, $fgrp);
346 my $src = [ 'git', "--git-dir=$fgrp->{-osdir}", 'for-each-ref',
347 "--format=refs/%(refname:lstrip=3)%00%(objectname)",
348 "refs/remotes/$rn/" ];
349 open(my $sfh, '+>', undef) or die "open(src): $!";
350 $fgrp->{srcfh} = $sfh;
351 start_cmd($fgrp, $src, { %opt, 1 => $sfh }, $update_ref);
352 my $dst = [ 'git', "--git-dir=$fgrp->{cur_dst}", 'for-each-ref',
353 '--format=%(refname)%00%(objectname)' ];
354 open(my $dfh, '+>', undef) or die "open(dst): $!";
355 $fgrp->{dstfh} = $dfh;
356 start_cmd($fgrp, $dst, { %opt, 1 => $dfh }, $update_ref);
362 my $todo = $FGRP_TODO;
363 $FGRP_TODO = \'BUG on further use';
364 keys(%$todo) or return;
366 # Rely on the fgrptmp remote groups in the config file rather
367 # than listing all remotes since the remote name list may exceed
368 # system argv limits:
371 my @git = (@{$self->{-torsocks}}, 'git');
372 my $j = $self->{lei}->{opt}->{jobs};
375 local $self->{lei}->{opt}->{jobs} = 1;
376 (fetch_args($self->{lei}, $opt),
377 qw(--no-tags --multiple));
379 push(@fetch, "-j$j") if $j;
380 while (my ($osdir, $fgrpv) = each %$todo) {
381 my $f = "$osdir/config";
382 return if !keep_going($self);
384 # clobber group from previous run atomically
385 my $cmd = ['git', "--git-dir=$osdir", qw(config -f),
386 $f, '--unset-all', "remotes.$grp"];
387 $self->{lei}->qerr("# @$cmd");
388 if (!$self->{dry_run}) {
389 my $pid = spawn($cmd, undef, { 2 => $self->{lei}->{2} });
390 waitpid($pid, 0) // die "waitpid: $!";
391 die "E: @$cmd: \$?=$?" if ($? && ($? >> 8) != 5);
393 # update the config atomically via O_APPEND while
394 # respecting git-config locking
395 sysopen(my $lk, "$f.lock", O_CREAT|O_EXCL|O_WRONLY)
396 or die "open($f.lock): $!";
397 open my $fh, '>>', $f or die "open(>>$f): $!";
399 my $buf = join('', "[remotes]\n",
400 map { "\t$grp = $_->{-remote}\n" } @$fgrpv);
401 print $fh $buf or die "print($f): $!";
402 close $fh or die "close($f): $!";
403 unlink("$f.lock") or die "unlink($f.lock): $!";
406 $cmd = [ @git, "--git-dir=$osdir", @fetch, $grp ];
407 my $end = PublicInbox::OnDestroy->new($$, \&fgrpv_done, $fgrpv);
408 start_cmd($self, $cmd, $opt, $end);
412 # keep this idempotent for future use by public-inbox-fetch
414 my ($self, $uri) = @_;
415 $self->{-ent} // return;
416 my $os = $self->{-objstore} // return;
417 my $fg = $self->{-ent}->{forkgroup} // return;
418 my $dir = "$os/$fg.git";
419 if (!-d $dir && !$self->{dry_run}) {
420 PublicInbox::Import::init_bare($dir);
421 my @cmd = ('git', "--git-dir=$dir", 'config');
422 my $opt = { 2 => $self->{lei}->{2} };
423 for ('repack.useDeltaIslands=true',
424 'pack.island=refs/remotes/([^/]+)/') {
425 run_die([@cmd, split(/=/, $_, 2)], undef, $opt);
428 my $key = $self->{-key} // die 'BUG: no -key';
429 my $rn = substr(sha256_hex($key), 0, 16);
430 if (!-d $self->{cur_dst} && !$self->{dry_run}) {
431 PublicInbox::Import::init_bare($self->{cur_dst});
432 my $f = "$self->{cur_dst}/config";
433 open my $fh, '+>>', $f or die "open:($f): $!";
434 print $fh <<EOM or die "print($f): $!";
435 ; rely on the "$rn" remote in the
436 ; $fg fork group for fetches
437 ; only uncomment the following iff you detach from fork groups
440 ; fetch = +refs/*:refs/*
443 close $fh or die "close($f): $!";
445 if (!$self->{dry_run}) {
446 my $alt = File::Spec->rel2abs("$dir/objects");
447 my $o = "$self->{cur_dst}/objects";
448 my $f = "$o/info/alternates";
449 my $l = File::Spec->abs2rel($alt, File::Spec->rel2abs($o));
450 open my $fh, '+>>', $f or die "open($f): $!";
451 seek($fh, SEEK_SET, 0) or die "seek($f): $!";
452 chomp(my @cur = <$fh>);
453 if (!grep(/\A\Q$l\E\z/, @cur)) {
454 say $fh $l or die "say($f): $!";
456 close $fh or die "close($f): $!";
459 %$self, -osdir => $dir, -remote => $rn, -uri => $uri
464 my ($self, $cmd, $cb, @arg) = @_;
466 $self->{lei}->err("@$cmd failed (\$?=$?) (non-fatal)");
467 $? = 0; # don't let it influence normal exit
469 return if !keep_going($self);
470 my $fh = delete $self->{-show_ref} // die 'BUG: no show-ref output';
471 seek($fh, SEEK_SET, 0) or die "seek(show_ref): $!";
472 $self->{-ent} // die 'BUG: no -ent';
473 my $A = $self->{-ent}->{fingerprint} // die 'BUG: no fingerprint';
474 my $B = sha1_hex(do { local $/; <$fh> } // die("read(show_ref): $!"));
475 return $cb->($self, @arg) if $A ne $B;
476 $self->{lei}->qerr("# $self->{-key} up-to-date");
480 my ($self, $cb, @arg) = @_;
481 # $cb is either resume_fetch or fgrp_enqueue
482 $self->{-ent} // return $cb->($self, @arg);
483 my $new = $self->{-ent}->{fingerprint} // return $cb->($self, @arg);
484 my $key = $self->{-key} // die 'BUG: no -key';
485 if (my $cur_ent = $self->{-local_manifest}->{$key}) {
486 # runs go_fetch->DESTROY run if eq
487 return if $cur_ent->{fingerprint} eq $new;
489 my $dst = $self->{cur_dst} // $self->{dst};
490 my $cmd = ['git', "--git-dir=$dst", 'show-ref'];
491 my $opt = { 2 => $self->{lei}->{2} };
492 open($opt->{1}, '+>', undef) or die "open(tmp): $!";
493 $self->{-show_ref} = $opt->{1};
495 $self->{lei}->qerr("# @$cmd");
496 return if $self->{dry_run};
497 $LIVE->{spawn($cmd, undef, $opt)} = [ \&fp_done, $self, $cmd,
502 my ($self, $uri, $fini) = @_;
503 return if !keep_going($self);
504 my $dst = $self->{cur_dst} // $self->{dst};
505 my @git = ('git', "--git-dir=$dst");
506 my $opt = { 2 => $self->{lei}->{2} };
507 my $rn = 'origin'; # configurable?
508 for ("url=$uri", "fetch=+refs/*:refs/*", 'mirror=true') {
509 my @kv = split(/=/, $_, 2);
510 $kv[0] = "remote.$rn.$kv[0]";
511 next if $self->{dry_run};
512 run_die([@git, 'config', @kv], undef, $opt);
514 my $cmd = [ @{$self->{-torsocks}}, @git,
515 fetch_args($self->{lei}, $opt), $rn ];
516 push @$cmd, '-P' if $self->{lei}->{prune}; # --prune-tags implied
517 start_cmd($self, $cmd, $opt, $fini);
521 my ($fgrp, $end) = @_; # $end calls fgrp_fetch_all
522 return if !keep_going($fgrp);
523 my $opt = { 2 => $fgrp->{lei}->{2} };
524 # --no-tags is required to avoid conflicts
525 my $u = $fgrp->{-uri} // die 'BUG: no {-uri}';
526 my $rn = $fgrp->{-remote} // die 'BUG: no {-remote}';
527 my @cmd = ('git', "--git-dir=$fgrp->{-osdir}", 'config');
528 for ("url=$u", "fetch=+refs/*:refs/remotes/$rn/*", 'tagopt=--no-tags') {
529 my @kv = split(/=/, $_, 2);
530 $kv[0] = "remote.$rn.$kv[0]";
531 $fgrp->{dry_run} ? $fgrp->{lei}->qerr("# @cmd @kv") :
532 run_die([@cmd, @kv], undef, $opt);
534 push @{$FGRP_TODO->{$fgrp->{-osdir}}}, $fgrp;
538 my ($self, $end) = @_;
539 my $lei = $self->{lei};
540 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
541 my $uri = URI->new($self->{cur_src} // $self->{src});
542 my $path = $uri->path;
543 $path =~ s!/*\z!! and $uri->path($path);
544 defined($lei->{opt}->{epoch}) and
545 die "$uri is a v1 inbox, --epoch is not supported\n";
546 $self->{-torsocks} //= $curl->torsocks($lei, $uri) or return;
547 my $dst = $self->{cur_dst} // $self->{dst};
548 my $fini = PublicInbox::OnDestroy->new($$, \&v1_done, $self);
549 my $resume = -d $dst;
550 if (my $fgrp = forkgroup_prep($self, $uri)) {
551 $fgrp->{-fini} = $fini;
552 $resume ? cmp_fp_do($fgrp, \&fgrp_enqueue, $end)
553 : fgrp_enqueue($fgrp, $end);
555 cmp_fp_do($self, \&resume_fetch, $uri, $fini);
556 } else { # normal clone
557 my $cmd = [ @{$self->{-torsocks}},
558 clone_cmd($lei, my $opt = {}), "$uri", $dst ];
559 if (defined($self->{-ent})) {
560 if (defined(my $ref = $self->{-ent}->{reference})) {
561 -e "$self->{dst}$ref" and
562 push @$cmd, '--reference',
566 start_cmd($self, $cmd, $opt, $fini);
568 if (!$self->{-is_epoch} && $lei->{opt}->{'inbox-config'} =~
569 /\A(?:always|v1)\z/s) {
570 _get_txt_start($self, '_/text/config/raw', $fini);
573 my $d = $self->{-ent} ? $self->{-ent}->{description} : undef;
574 $self->{'txt.description'} = $d if defined $d;
575 (!defined($d) && !$end) and
576 _get_txt_start($self, 'description', $fini);
578 $end or do_reap($self, 1); # for non-manifest clone
581 sub parse_epochs ($$) {
582 my ($opt_epochs, $v2_epochs) = @_; # $epochs "LOW..HIGH"
583 $opt_epochs // return; # undef => all epochs
584 my ($lo, $dotdot, $hi, @extra) = split(/(\.\.)/, $opt_epochs);
585 undef($lo) if ($lo // '') eq '';
586 my $re = qr/\A~?[0-9]+\z/;
587 if (@extra || (($lo // '0') !~ $re) ||
588 (($hi // '0') !~ $re) ||
589 !(grep(defined, $lo, $hi))) {
591 --epoch=$opt_epochs not in the form of `LOW..HIGH', `LOW..', nor `..HIGH'
594 my @n = sort { $a <=> $b } keys %$v2_epochs;
595 for (grep(defined, $lo, $hi)) {
598 "`$_' exceeds maximum available epoch ($n[-1])\n";
600 "`$_' is lower than minimum available epoch ($n[0])\n";
601 } elsif (/\A~([0-9]+)/) {
603 $n[$off] // die "`$_' is out of range\n";
605 } else { die "`$_' not understood\n" }
607 defined($lo) && defined($hi) && $lo > $hi and die
608 "low value (`$lo') exceeds high (`$hi')\n";
609 $lo //= $n[0] if $dotdot;
610 $hi //= $n[-1] if $dotdot;
614 if (defined $v2_epochs->{$_}) {
618 "# epoch $_ is not available (non-fatal, $lo..$hi)\n";
624 sub init_placeholder ($$$) {
625 my ($src, $edst, $ent) = @_;
626 PublicInbox::Import::init_bare($edst);
627 my $f = "$edst/config";
628 open my $fh, '>>', $f or die "open($f): $!";
629 print $fh <<EOM or die "print($f): $!";
632 fetch = +refs/*:refs/*
635 ; This git epoch was created read-only and "public-inbox-fetch"
636 ; will not fetch updates for it unless write permission is added.
637 ; Hint: chmod +w $edst
639 if (defined($ent->{owner})) {
640 print $fh <<EOM or die "print($f): $!";
642 owner = $ent->{owner}
645 close $fh or die "close($f): $!";
646 my %map = (head => 'HEAD', description => undef);
647 while (my ($key, $fn) = each %map) {
648 my $val = $ent->{$key} // next;
651 open $fh, '>', $fn or die "open($fn): $!";
652 print $fh $val, "\n" or die "print($fn): $!";
653 close $fh or die "close($fn): $!";
657 sub reap_cmd { # async, called via SIGCHLD
658 my ($self, $cmd) = @_;
660 $? = 0; # don't let it influence normal exit
661 $self->{lei}->child_error($cerr, "@$cmd failed (\$?=$cerr)") if $cerr;
666 return if !keep_going($self);
667 my $fh = delete $self->{-show_ref_up} // die 'BUG: no show-ref output';
668 seek($fh, SEEK_SET, 0) or die "seek(show_ref): $!";
669 $self->{-ent} // die 'BUG: no -ent';
670 my $A = $self->{-ent}->{fingerprint} // die 'BUG: no fingerprint';
671 my $B = sha1_hex(do { local $/; <$fh> } // die("read(show_ref): $!"));
673 $self->{-ent}->{fingerprint} = $B;
674 push @{$self->{chg}->{fp_mismatch}}, $self->{-key};
677 sub atomic_write ($$$) {
678 my ($dn, $bn, $raw) = @_;
679 my $ft = File::Temp->new(DIR => $dn, TEMPLATE => "$bn-XXXX");
680 print $ft $raw or die "print($ft): $!";
681 $ft->flush or die "flush($ft): $!";
682 ft_rename($ft, "$dn/$bn", 0666);
687 my $puh = shift @{$self->{-puh_todo}} // return;
688 my $fini = PublicInbox::OnDestroy->new($$, \&run_next_puh, $self);
689 my $cmd = [ @$puh, ($self->{cur_dst} // $self->{dst}) ];
690 my $opt = +{ map { $_ => $self->{lei}->{$_} } (0..2) };
691 start_cmd($self, $cmd, undef, $opt, $fini);
694 sub run_post_update_hooks {
696 my $puh = $self->{-puh} // return;
697 @{$self->{-puh_todo}} = @$puh;
701 # modifies the to-be-written manifest entry, and sets values from it, too
704 my $key = $self->{-key} // die 'BUG: no -key';
705 my $new = $self->{-ent}->{fingerprint};
706 my $cur = $self->{-local_manifest}->{$key}->{fingerprint} // "\0";
707 my $dst = $self->{cur_dst} // $self->{dst};
708 if (defined($new) && $new ne $cur) {
709 my $cmd = ['git', "--git-dir=$dst", 'show-ref'];
710 my $opt = { 2 => $self->{lei}->{2} };
711 open($opt->{1}, '+>', undef) or die "open(tmp): $!";
712 $self->{-show_ref_up} = $opt->{1};
713 my $done = PublicInbox::OnDestroy->new($$, \&up_fp_done, $self);
714 start_cmd($self, $cmd, $opt, $done);
716 $new = $self->{-ent}->{head};
717 $cur = $self->{-local_manifest}->{$key}->{head} // "\0";
718 if (defined($new) && $new ne $cur) {
719 # n.b. grokmirror writes raw contents to $dst/HEAD w/o locking
720 my $cmd = [ 'git', "--git-dir=$dst" ];
721 if ($new =~ s/\Aref: //) {
722 push @$cmd, qw(symbolic-ref HEAD), $new;
723 } elsif ($new =~ /\A[a-f0-9]{40,}\z/) {
724 push @$cmd, qw(update-ref --no-deref HEAD), $new;
727 warn "W: $key: {head} => `$new' not understood\n";
729 start_cmd($self, $cmd, { 2 => $self->{lei}->{2} }) if $cmd;
731 if (my $symlinks = $self->{-ent}->{symlinks}) {
732 my $top = File::Spec->rel2abs($self->{dst});
733 push @{$self->{-new_symlinks}}, @$symlinks;
734 for my $p (@$symlinks) {
737 my (undef, $dn, $bn) = File::Spec->splitpath($ln);
738 File::Path::mkpath($dn);
739 my $tgt = "$top/$key";
740 $tgt = File::Spec->abs2rel($tgt, $dn);
743 next if readlink($ln) eq $tgt;
744 unlink($ln) or die "unlink($ln): $!";
746 push @{$self->{chg}->{badlink}}, $p;
749 symlink($tgt, $ln) or die "symlink($tgt, $ln): $!";
752 if (defined(my $t = $self->{-ent}->{modified})) {
753 my ($dn, $bn) = ("$dst/info/web", 'last-modified');
754 my $orig = PublicInbox::Git::try_cat("$dn/$bn");
755 $t = strftime('%F %T', gmtime($t))." +0000\n";
756 File::Path::mkpath($dn);
757 atomic_write($dn, $bn, $t) if $orig ne $t;
760 $new = $self->{-ent}->{owner} // return;
761 $cur = $self->{-local_manifest}->{$key}->{owner} // "\0";
762 return if $cur eq $new;
763 my $cmd = [ qw(git config -f), "$dst/config", 'gitweb.owner', $new ];
764 start_cmd($self, $cmd, { 2 => $self->{lei}->{2} });
767 sub v1_done { # called via OnDestroy
769 return if $self->{dry_run} || !keep_going($self);
770 _write_inbox_config($self);
771 my $dst = $self->{cur_dst} // $self->{dst};
772 update_ent($self) if $self->{-ent};
773 my $o = "$dst/objects";
774 if (open(my $fh, '<', my $fn = "$o/info/alternates")) {;
775 my $base = File::Spec->rel2abs($o);
780 $_ = File::Spec->abs2rel($_, $base);
781 $ft //= File::Temp->new(TEMPLATE => '.XXXX',
785 print $ft @l or die "print($ft): $!";
786 $ft->flush or die "flush($ft): $!";
787 ft_rename($ft, $fn, 0666, $fh);
790 eval { set_description($self) };
792 run_post_update_hooks($self);
793 return if ($self->{-is_epoch} ||
794 $self->{lei}->{opt}->{'inbox-config'} ne 'always');
795 write_makefile($dst, 1);
796 index_cloned_inbox($self, 1);
799 sub v2_done { # called via OnDestroy
801 return if $self->{dry_run} || !keep_going($self);
802 my $dst = $self->{cur_dst} // $self->{dst};
803 require PublicInbox::Lock;
804 my $lk = bless { lock_path => "$dst/inbox.lock" }, 'PublicInbox::Lock';
805 my $lck = $lk->lock_for_scope($$);
806 _write_inbox_config($self);
807 require PublicInbox::MultiGit;
808 my $mg = PublicInbox::MultiGit->new($dst, 'all.git', 'git');
809 $mg->fill_alternates;
810 for my $i ($mg->git_epochs) { $mg->epoch_cfg_set($i) }
811 for my $edst (@{delete($self->{-read_only}) // []}) {
812 my @st = stat($edst) or die "stat($edst): $!";
813 chmod($st[2] & 0555, $edst) or die "chmod(a-w, $edst): $!";
815 write_makefile($dst, 2);
817 eval { set_description($self) };
819 index_cloned_inbox($self, 2);
822 sub clone_v2_prep ($$;$) {
823 my ($self, $v2_epochs, $m) = @_; # $m => manifest.js.gz hashref
824 my $lei = $self->{lei};
825 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
826 my $first_uri = (map { $_->[0] } values %$v2_epochs)[0];
827 $self->{-torsocks} //= $curl->torsocks($lei, $first_uri) or return;
828 my $dst = $self->{cur_dst} // $self->{dst};
829 my $want = parse_epochs($lei->{opt}->{epoch}, $v2_epochs);
830 my $task = $m ? bless { %$self }, __PACKAGE__ : $self;
832 my $fini = PublicInbox::OnDestroy->new($$, \&v2_done, $task);
833 for my $nr (sort { $a <=> $b } keys %$v2_epochs) {
834 my ($uri, $key) = @{$v2_epochs->{$nr}};
835 my $src = $uri->as_string;
837 $src =~ m!/([0-9]+)(?:\.git)?\z! or die <<"";
838 failed to extract epoch number from $src
840 $1 + 0 == $nr or die "BUG: <$uri> miskeyed $1 != $nr";
841 $edst .= "/git/$nr.git";
845 die("BUG: `$key' not in manifest.js.gz");
846 if (defined(my $d = $ent->{description})) {
847 $d =~ s/ \[epoch [0-9]+\]\z//s;
851 if (!$want || $want->{$nr}) {
852 my $etask = bless { %$task, -key => $key }, __PACKAGE__;
853 $etask->{-ent} = $ent; # may have {reference}
854 $etask->{cur_src} = $src;
855 $etask->{cur_dst} = $edst;
856 $etask->{-is_epoch} = $fini;
857 my $ref = $ent->{reference} // '';
858 push @{$TODO->{$ref}}, $etask;
859 $self->{any_want}->{$key} = 1;
860 } else { # create a placeholder so users only need to chmod +w
861 init_placeholder($src, $edst, $ent);
862 push @{$task->{-read_only}}, $edst;
866 # filter out the epochs we skipped
867 $self->{chg}->{manifest} = 1 if $m && delete(@$m{@skip});
869 (!$self->{dry_run} && !-d $dst) and File::Path::mkpath($dst);
871 $lei->{opt}->{'inbox-config'} =~ /\A(?:always|v2)\z/s and
872 _get_txt_start($task, '_/text/config/raw', $fini);
874 defined($desc) ? ($task->{'txt.description'} = $desc) :
875 _get_txt_start($task, 'description', $fini);
878 sub decode_manifest ($$$) {
879 my ($fh, $fn, $uri) = @_;
881 my $gz = do { local $/; <$fh> } // die "slurp($fn): $!";
882 gunzip(\$gz => \$js, MultiStream => 1) or
883 die "gunzip($uri): $GunzipError\n";
884 my $m = eval { PublicInbox::Config->json->decode($js) };
885 die "$uri: error decoding `$js': $@\n" if $@;
886 ref($m) eq 'HASH' or die "$uri unknown type: ".ref($m);
890 sub load_current_manifest ($) {
892 my $fn = $self->{-manifest} // return;
893 if (open(my $fh, '<', $fn)) {
894 decode_manifest($fh, $fn, $fn);
895 } elsif ($!{ENOENT}) { # non-fatal, we can just do it slowly
896 warn "open($fn): $!\n" if !$self->{-initial_clone};
899 die "open($fn): $!\n";
903 sub multi_inbox ($$$) {
904 my ($self, $path, $m) = @_;
905 my $incl = $self->{lei}->{opt}->{include};
906 my $excl = $self->{lei}->{opt}->{exclude};
908 # assuming everything not v2 is v1, for now
909 my @v1 = sort grep(!m!.+/git/[0-9]+\.git\z!, keys %$m);
910 my @v2_epochs = sort grep(m!.+/git/[0-9]+\.git\z!, keys %$m);
914 m!\A(/.+)/git/[0-9]+\.git\z! or die "BUG: $_";
915 push @{$v2->{$1}}, $_;
917 my $n = scalar(keys %$v2) + scalar(@v1);
918 my @orig = defined($incl // $excl) ? (keys %$v2, @v1) : ();
920 my $re = '(?:'.join('\\z|', map {
921 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
923 my @gone = delete @$v2{grep(!/$re/, keys %$v2)};
924 delete @$m{map { @$_ } @gone} and $self->{chg}->{manifest} = 1;
925 delete @$m{grep(!/$re/, @v1)} and $self->{chg}->{manifest} = 1;
926 @v1 = grep(/$re/, @v1);
929 my $re = '(?:'.join('\\z|', map {
930 $self->{lei}->glob2re($_) // qr/\A\Q$_\E/
932 my @gone = delete @$v2{grep(/$re/, keys %$v2)};
933 delete @$m{map { @$_ } @gone} and $self->{chg}->{manifest} = 1;
934 delete @$m{grep(/$re/, @v1)} and $self->{chg}->{manifest} = 1;
935 @v1 = grep(!/$re/, @v1);
937 my $ret; # { v1 => [ ... ], v2 => { "/$inbox_name" => [ epochs ] }}
938 $ret->{v1} = \@v1 if @v1;
939 $ret->{v2} = $v2 if keys %$v2;
940 $ret //= @orig ? "Nothing to clone, available repositories:\n\t".
941 join("\n\t", sort @orig)
942 : "Nothing available to clone\n";
945 # PSGI mount prefixes and manifest.js.gz prefixes don't always align...
947 until (grep(m!\A\Q$$path\E/git/[0-9]+\.git\z!,
948 @v2_epochs) == @v2_epochs) {
949 $$path =~ s!\A(/[^/]+)/!/! or last;
953 while (!defined($m->{$$path}) && $$path =~ s!\A(/[^/]+)/!/!) {
957 ($path_pfx, $n, $ret);
963 $TODO = \'BUG on further use';
964 my $end = PublicInbox::OnDestroy->new($$, \&fgrp_fetch_all, $self);
966 my $nodep = delete $todo->{''};
968 # do not download unwanted deps
969 my $any_want = delete $self->{any_want};
970 my @unwanted = grep { !$any_want->{$_} } keys %$todo;
971 my @nodep = delete(@$todo{@unwanted});
972 push(@$nodep, @$_) for @nodep;
974 # handle no-dependency repos, first
977 return if !keep_going($self);
980 # resolve references, deepest, first:
981 while (scalar keys %$todo) {
982 for my $x (keys %$todo) {
984 # resolve multi-level references
985 while ($m && defined($nxt = $m->{$x}->{reference})) {
986 exists($todo->{$nxt}) or last;
988 $m->{$x}->{reference} = undef;
989 $m->{$nxt}->{reference} = undef;
991 E: dependency loop detected (`$x' => `$nxt'), breaking
996 my $y = delete $todo->{$x} // next; # already done
999 return if !keep_going($self);
1001 last; # restart %$todo iteration
1005 # $end->DESTROY will call fgrp_fetch_all once all references
1006 # in $LIVE are gone, and do_reap will eventually drain $LIVE
1011 sub dump_manifest ($$) {
1013 # write the smaller manifest if epochs were skipped so
1014 # users won't have to delete manifest if they +w an
1015 # epoch they no longer want to skip
1016 my $json = PublicInbox::Config->json->encode($m);
1017 my $mtime = (stat($ft))[9];
1018 seek($ft, SEEK_SET, 0) or die "seek($ft): $!";
1019 truncate($ft, 0) or die "truncate($ft): $!";
1020 gzip(\$json => $ft) or die "gzip($ft): $GzipError";
1021 $ft->flush or die "flush($ft): $!";
1022 utime($mtime, $mtime, "$ft") or die "utime(..., $ft): $!";
1025 sub dump_project_list ($$) {
1026 my ($self, $m) = @_;
1027 my $f = $self->{'-project-list'} // return;
1028 my $old = PublicInbox::Git::try_cat($f);
1031 open my $dh, '<', '.' or die "open(.): $!";
1032 chdir($self->{dst}) or die "chdir($self->{dst}): $!";
1033 my @local = grep { -e $_ ? ($new{$_} = undef) : 1 } split(/\n/s, $old);
1034 chdir($dh) or die "chdir(restore): $!";
1036 $new{substr($_, 1)} = 1 for keys %$m; # drop leading '/'
1037 my @list = sort keys %new;
1038 my @remote = grep { !defined($new{$_}) } @list;
1039 my %lnk = map { substr($_, 1) => undef } @{$self->{-new_symlinks}};
1040 @remote = grep { !exists($lnk{$_}) } @remote;
1042 warn <<EOM if @remote;
1043 The following local repositories are ignored/gone from $self->{src}:
1045 warn "\t", $_, "\n" for @remote;
1046 warn <<EOM if @local;
1047 The following repos in $f no longer exist on the filesystem:
1049 warn "\t", $_, "\n" for @local;
1051 my (undef, $dn, $bn) = File::Spec->splitpath($f);
1052 atomic_write($dn, $bn, join("\n", @list, ''));
1055 # FIXME: this gets confused by single inbox instance w/ global manifest.js.gz
1058 my $uri = URI->new($self->{src});
1059 my $lei = $self->{lei};
1060 my $curl = $self->{curl} //= PublicInbox::LeiCurl->new($lei) or return;
1061 $self->{-torsocks} //= $curl->torsocks($lei, $uri) or return;
1062 my $path = $uri->path;
1063 chop($path) eq '/' or die "BUG: $uri not canonicalized";
1064 $uri->path($path . '/manifest.js.gz');
1065 my $manifest = $self->{-manifest} // "$self->{dst}/manifest.js.gz";
1066 my %opt = (UNLINK => 1, SUFFIX => '.tmp', TMPDIR => 1);
1067 if (!$self->{dry_run} && $manifest =~ m!\A(.+?)/[^/]+\z! and -d $1) {
1068 $opt{DIR} = $1; # allows fast rename(2) w/o EXDEV
1069 delete $opt{TMPDIR};
1071 my $ft = File::Temp->new(TEMPLATE => '.manifest-XXXX', %opt);
1072 my $cmd = $curl->for_uri($lei, $uri, qw(-R -o), $ft->filename);
1073 push(@$cmd, '-z', $manifest) if -f $manifest;
1074 my $mf_url = "$uri";
1075 %opt = map { $_ => $lei->{$_} } (0..2);
1076 my $cerr = run_reap($lei, $cmd, \%opt);
1078 return try_scrape($self) if ($cerr >> 8) == 22; # 404 missing
1079 return $lei->child_error($cerr, "@$cmd failed");
1082 # bail out if curl -z/--timecond hit 304 Not Modified, $ft will be empty
1083 return $lei->qerr("# $manifest unchanged") if -f $manifest && !-s $ft;
1085 my $m = eval { decode_manifest($ft, $ft, $uri) };
1088 return try_scrape($self);
1090 local $self->{chg} = {};
1091 local $self->{-local_manifest} = load_current_manifest($self);
1092 local $self->{-new_symlinks} = [];
1093 my ($path_pfx, $n, $multi) = multi_inbox($self, \$path, $m);
1094 return $lei->child_error(1, $multi) if !ref($multi);
1095 my $v2 = delete $multi->{v2};
1097 for my $name (sort keys %$v2) {
1098 my $epochs = delete $v2->{$name};
1099 my %v2_epochs = map {
1100 $uri->path($n > 1 ? $path_pfx.$path.$_
1102 my ($e) = ("$uri" =~ m!/([0-9]+)\.git\z!);
1103 $e // die "no [0-9]+\.git in `$uri'";
1104 $e => [ $uri->clone, $_ ];
1106 ("$uri" =~ m!\A(.+/)git/[0-9]+\.git\z!) or
1107 die "BUG: `$uri' !~ m!/git/[0-9]+.git!";
1108 local $self->{cur_src} = $1;
1109 local $self->{cur_dst} = $self->{dst};
1110 if ($n > 1 && $uri->path =~ m!\A\Q$path_pfx$path\E/(.+)/
1111 git/[0-9]+\.git\z!x) {
1112 $self->{cur_dst} .= "/$1";
1114 index($self->{cur_dst}, "\n") >= 0 and die <<EOM;
1115 E: `$self->{cur_dst}' must not contain newline
1117 clone_v2_prep($self, \%v2_epochs, $m);
1118 return if !keep_going($self);
1121 if (my $v1 = delete $multi->{v1}) {
1122 my $p = $path_pfx.$path;
1123 chop($p) if substr($p, -1, 1) eq '/';
1125 for my $name (@$v1) {
1126 my $task = bless { %$self }, __PACKAGE__;
1127 $task->{-ent} = $m->{$name} //
1128 die("BUG: no `$name' in manifest");
1129 $task->{cur_src} = "$uri";
1130 $task->{cur_dst} = $task->{dst};
1131 $task->{-key} = $name;
1133 $task->{cur_dst} .= $name;
1134 $task->{cur_src} .= $name;
1136 index($task->{cur_dst}, "\n") >= 0 and die <<EOM;
1137 E: `$task->{cur_dst}' must not contain newline
1139 $task->{cur_src} .= '/';
1140 my $dep = $task->{-ent}->{reference} // '';
1141 push @{$TODO->{$dep}}, $task; # for clone_all
1142 $self->{any_want}->{$name} = 1;
1145 delete local $lei->{opt}->{epoch} if defined($v2);
1146 clone_all($self, $m);
1147 return if $self->{dry_run} || !keep_going($self);
1149 # set by clone_v2_prep/-I/--exclude
1150 my $mis = delete $self->{chg}->{fp_mismatch};
1152 my $t = (stat($ft))[9];
1153 $t = strftime('%F %k:%M:%S %z', localtime($t));
1155 W: Fingerprints for the following repositories do not match
1157 W: These repositories may have updated since $t:
1159 warn "\t", $_, "\n" for @$mis;
1160 warn <<EOM if !$self->{lei}->{opt}->{prune};
1161 W: The above fingerprints may never match without --prune
1164 dump_manifest($m => $ft) if delete($self->{chg}->{manifest}) || $mis;
1165 my $bad = delete $self->{chg}->{badlink};
1166 warn(<<EOM, map { ("\t", $_, "\n") } @$bad) if $bad;
1167 W: The following exist and have not been converted to symlinks
1169 dump_project_list($self, $m);
1170 ft_rename($ft, $manifest, 0666);
1173 sub start_clone_url {
1175 return try_manifest($self) if $self->{src} =~ m!\Ahttps?://!;
1176 die "TODO: non-HTTP/HTTPS clone of $self->{src} not supported, yet";
1179 sub do_mirror { # via wq_io_do or public-inbox-clone
1181 my $lei = $self->{lei};
1182 $self->{dry_run} = 1 if $lei->{opt}->{'dry-run'};
1183 umask($lei->{client_umask}) if defined $lei->{client_umask};
1184 $self->{-initial_clone} = 1 if !-d $self->{dst};
1185 if (defined(my $puh = $lei->{opt}->{'post-update-hook'})) {
1186 require Text::ParseWords;
1188 my $pfx = [ Text::ParseWords::shellwords($_) ];
1189 push @{$self->{-puh}}, $pfx;
1193 my $ic = $lei->{opt}->{'inbox-config'} //= 'always';
1194 $ic =~ /\A(?:v1|v2|always|never)\z/s or die <<"";
1195 --inbox-config must be one of `always', `v2', `v1', or `never'
1197 # we support these switches with '' (empty string).
1198 # defaults match example conf distributed with grokmirror
1199 my @pairs = qw(objstore objstore manifest manifest.js.gz
1200 project-list projects.list);
1202 my ($k, $default) = splice(@pairs, 0, 2);
1203 my $v = $lei->{opt}->{$k} // next;
1204 $v = $default if $v eq '';
1205 $v = "$self->{dst}/$v" if $v !~ m!\A\.{0,2}/!;
1206 $self->{"-$k"} = $v;
1211 local $FGRP_TODO = {};
1212 my $iv = $lei->{opt}->{'inbox-version'} //
1213 return start_clone_url($self);
1214 return clone_v1($self) if $iv == 1;
1215 die "bad --inbox-version=$iv\n" if $iv != 2;
1216 die <<EOM if $self->{src} !~ m!://!;
1217 cloning local v2 inboxes not supported
1219 try_scrape($self, 1);
1221 $lei->fail($@) if $@;
1225 my ($cls, $lei, $src, $dst) = @_;
1226 my $self = bless { src => $src, dst => $dst }, $cls;
1227 $lei->request_umask;
1228 my ($op_c, $ops) = $lei->workers_start($self, 1);
1229 $lei->{wq1} = $self;
1230 $self->wq_io_do('do_mirror', []);
1232 $lei->wait_wq_events($op_c, $ops);
1235 sub ipc_atfork_child {
1237 $self->{lei}->_lei_atfork_child;
1238 $self->SUPER::ipc_atfork_child;
1241 sub write_makefile {
1242 my ($dir, $ibx_ver) = @_;
1243 my $f = "$dir/Makefile";
1244 if (sysopen my $fh, $f, O_CREAT|O_EXCL|O_WRONLY) {
1245 print $fh <<EOM or die "print($f) $!";
1246 # This is a v$ibx_ver public-inbox, see the public-inbox-v$ibx_ver-format(5)
1247 # manpage for more information on the format. This Makefile is
1248 # intended as a familiar wrapper for users unfamiliar with
1249 # public-inbox-* commands.
1251 # See the respective manpages for public-inbox-fetch(1),
1252 # public-inbox-index(1), etc for more information on
1253 # some of the commands used by this Makefile.
1255 # This Makefile will not be modified nor read by public-inbox,
1256 # so you may edit it freely with your own convenience targets
1257 # and notes. public-inbox-fetch will recreate it if removed.
1259 print $fh <<'EOM' or die "print($f): $!";
1260 # the default target:
1262 @echo Common targets:
1263 @echo ' make fetch - fetch from remote git repostorie(s)'
1264 @echo ' make update - fetch and update index '
1266 @echo Rarely needed targets:
1267 @echo ' make reindex - may be needed for new features/bugfixes'
1268 @echo ' make compact - rewrite Xapian storage to save space'
1269 @echo ' make index - initial index after clone
1274 @if ! public-inbox-fetch --exit-code; \
1277 test $$c -eq 127 && exit 0; \
1279 elif test -f msgmap.sqlite3 || test -f public-inbox/msgmap.sqlite3; \
1281 public-inbox-index; \
1283 echo 'public-inbox index not initialized'; \
1284 echo 'see public-inbox-index(1) man page'; \
1289 public-inbox-index --reindex
1291 public-inbox-compact
1293 .PHONY : help fetch update index reindex compact
1295 close $fh or die "close($f): $!";
1297 die "open($f): $!" unless $!{EEXIST};