1 # Copyright (C) 2020-2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # backend for a git-cat-file-workalike based on libgit2,
5 # other libgit2 stuff may go here, too.
6 package PublicInbox::Gcf2;
9 use PublicInbox::Spawn qw(which popen_rd); # may set PERL_INLINE_DIRECTORY
10 use Fcntl qw(LOCK_EX SEEK_SET);
11 use Time::HiRes qw(clock_gettime CLOCK_MONOTONIC);
12 use IO::Handle; # autoflush
13 use File::Path qw(make_path);
16 # PublicInbox::Spawn will set PERL_INLINE_DIRECTORY
17 # to ~/.cache/public-inbox/inline-c if it exists
18 my $inline_dir = $ENV{PERL_INLINE_DIRECTORY} //
19 die 'PERL_INLINE_DIRECTORY not defined';
20 make_path($inline_dir);
21 my $f = "$inline_dir/.public-inbox.lock";
22 open my $fh, '+>', $f or die "open($f): $!";
24 # CentOS 7.x ships Inline 0.53, 0.64+ has built-in locking
25 flock($fh, LOCK_EX) or die "LOCK_EX($f): $!\n";
27 my $pc = which($ENV{PKG_CONFIG} // 'pkg-config') //
28 die "pkg-config missing for libgit2";
29 my ($dir) = (__FILE__ =~ m!\A(.+?)/[^/]+\z!);
30 my $ef = "$inline_dir/.public-inbox.pkg-config.err";
31 open my $err, '+>', $ef or die "open($ef): $!";
32 for my $x (qw(libgit2)) {
33 my $rdr = { 2 => $err };
34 my ($l, $pid) = popen_rd([$pc, '--libs', $x], undef, $rdr);
35 $l = do { local $/; <$l> };
38 (my $c, $pid) = popen_rd([$pc, '--cflags', $x], undef, $rdr);
39 $c = do { local $/; <$c> };
43 # note: we name C source files .h to prevent
44 # ExtUtils::MakeMaker from automatically trying to
46 my $f = "$dir/gcf2_$x.h";
47 open(my $src, '<', $f) or die "E: open($f): $!";
50 defined($c_src = <$src>) or die "read $f: $!";
56 seek($err, 0, SEEK_SET);
57 $err = do { local $/; <$err> };
58 die "E: libgit2 not installed: $err\n";
60 open my $oldout, '>&', \*STDOUT or die "dup(1): $!";
61 open my $olderr, '>&', \*STDERR or die "dup(2): $!";
62 open STDOUT, '>&', $fh or die "1>$f: $!";
63 open STDERR, '>&', $fh or die "2>$f: $!";
67 # we use Capitalized and ALLCAPS for compatibility with old Inline::C
69 use Inline C => Config => %CFG, BOOT => q[git_libgit2_init();];
70 use Inline C => $c_src, BUILD_NOISY => 1;
73 open(STDERR, '>&', $olderr) or warn "restore stderr: $!";
74 open(STDOUT, '>&', $oldout) or warn "restore stdout: $!";
76 seek($fh, 0, SEEK_SET);
78 die "Inline::C Gcf2 build failed:\n", $err, "\n", @msg;
83 my ($gcf2, $git_dir) = @_;
84 my $objdir = PublicInbox::Git->new($git_dir)->git_path('objects');
86 # libgit2 (tested 0.27.7+dfsg.1-0.2 and 0.28.3+dfsg.1-1~bpo10+1
87 # in Debian) doesn't handle relative epochs properly when nested
88 # multiple levels. Add all the absolute paths to workaround it,
89 # since $EXTINDEX_DIR/ALL.git/objects/info/alternates uses absolute
90 # paths to reference $V2INBOX_DIR/all.git/objects and
91 # $V2INBOX_DIR/all.git/objects/info/alternates uses relative paths
92 # to refer to $V2INBOX_DIR/git/$EPOCH.git/objects
94 # See https://bugs.debian.org/975607
95 if (open(my $fh, '<', "$objdir/info/alternates")) {
96 chomp(my @abs_alt = grep(m!^/!, <$fh>));
97 $gcf2->add_alternate($_) for @abs_alt;
99 $gcf2->add_alternate($objdir);
103 sub have_unlinked_files () {
104 # FIXME: port gcf2-like over to git.git so we won't need to
106 return 1 if $^O ne 'linux';
107 open my $fh, '<', "/proc/$$/maps" or return;
108 while (<$fh>) { return 1 if /\.(?:idx|pack) \(deleted\)$/ }
112 # Usage: $^X -MPublicInbox::Gcf2 -e PublicInbox::Gcf2::loop [EXPIRE-TIMEOUT]
113 # (see lib/PublicInbox/Gcf2Client.pm)
115 my $exp = $_[0] || $ARGV[0] || 60; # seconds
117 my (%seen, $check_at);
118 STDERR->autoflush(1);
119 STDOUT->autoflush(1);
123 my ($oid, $git_dir) = split(/ /, $_, 2);
124 $seen{$git_dir} //= add_alt($gcf2, $git_dir);
125 if (!$gcf2->cat_oid(1, $oid)) {
126 # retry once if missing. We only get unabbreviated OIDs
127 # from SQLite or Xapian DBs, here, so malicious clients
128 # can't trigger excessive retries:
129 warn "I: $$ $oid missing, retrying in $git_dir\n";
132 %seen = ($git_dir => add_alt($gcf2, $git_dir);
133 $check_at = clock_gettime(CLOCK_MONOTONIC) + $exp;
135 if ($gcf2->cat_oid(1, $oid)) {
136 warn "I: $$ $oid found after retry\n";
138 warn "W: $$ $oid missing after retry\n";
139 print "$oid missing\n"; # mimic git-cat-file
141 } else { # check expiry to deal with deleted pack files
142 my $now = clock_gettime(CLOCK_MONOTONIC);
143 $check_at //= $now + $exp;
144 if ($now > $check_at) {
146 if (have_unlinked_files()) {