1 # Copyright (C) 2014-2018 all contributors <meta@public-inbox.org>
2 # License: GPLv2 or later <https://www.gnu.org/licenses/gpl-2.0.txt>
4 # Used to read files from a git repository without excessive forking.
5 # Used in our web interfaces as well as our -nntpd server.
6 # This is based on code in Git.pm which is GPLv2+, but modified to avoid
7 # dependence on environment variables for compatibility with mod_perl.
8 # There are also API changes to simplify our usage and data set.
9 package PublicInbox::Git;
14 use PublicInbox::Spawn qw(spawn popen_rd);
15 use base qw(Exporter);
16 our @EXPORT_OK = qw(git_unquote git_quote);
29 my %ESC_GIT = map { $GIT_ESC{$_} => $_ } keys %GIT_ESC;
32 # unquote pathnames used by git, see quote.c::unquote_c_style.c in git.git
34 return $_[0] unless ($_[0] =~ /\A"(.*)"\z/);
36 $_[0] =~ s/\\([\\"abfnrtv])/$GIT_ESC{$1}/g;
37 $_[0] =~ s/\\([0-7]{1,3})/chr(oct($1))/ge;
42 if ($_[0] =~ s/([\\"\a\b\f\n\r\t\013]|[^[:print:]])/
43 '\\'.($ESC_GIT{$1}||sprintf("%0o",ord($1)))/egs) {
50 my ($class, $git_dir) = @_;
53 # may contain {-wt} field (working-tree (File::Temp::Dir))
54 bless { git_dir => $git_dir, st => \@st }, $class
57 sub alternates_changed {
59 my $alt = "$self->{git_dir}/objects/info/alternates";
60 my @st = stat($alt) or return 0;
61 my $old_st = $self->{st};
62 # 10 - ctime, 7 - size
63 return 0 if ($st[10] == $old_st->[10] && $st[7] == $old_st->[7]);
69 my $fh = $self->{err_c} or return;
70 sysseek($fh, 0, 0) or fail($self, "sysseek failed: $!");
71 defined(sysread($fh, my $buf, -s $fh)) or
72 fail($self, "sysread failed: $!");
77 my ($self, $batch, $in, $out, $pid, $err) = @_;
79 if (defined $err) { # "err_c"
80 my $fh = $self->{$err};
81 sysseek($fh, 0, 0) or fail($self, "sysseek failed: $!");
82 truncate($fh, 0) or fail($self, "truncate failed: $!");
86 my ($in_r, $in_w, $out_r, $out_w);
88 pipe($in_r, $in_w) or fail($self, "pipe failed: $!");
89 pipe($out_r, $out_w) or fail($self, "pipe failed: $!");
90 if ($^O eq 'linux') { # 1031: F_SETPIPE_SZ
91 fcntl($out_w, 1031, 4096);
92 fcntl($in_w, 1031, 4096) if $batch eq '--batch-check';
95 my @cmd = (qw(git), "--git-dir=$self->{git_dir}",
96 qw(-c core.abbrev=40 cat-file), $batch);
97 my $redir = { 0 => fileno($out_r), 1 => fileno($in_w) };
99 open(my $fh, '+>', undef) or fail($self, "open.err failed: $!");
101 $redir->{2} = fileno($fh);
103 my $p = spawn(\@cmd, undef, $redir);
104 defined $p or fail($self, "spawn failed: $!");
106 $out_w->autoflush(1);
107 $self->{$out} = $out_w;
108 $self->{$in} = $in_r;
112 my ($self, $obj, $ref) = @_;
113 my ($retried, $in, $head);
116 batch_prepare($self);
117 $self->{out}->print($obj, "\n") or fail($self, "write error: $!");
121 $head = $in->getline;
122 if ($head =~ / missing$/) {
123 if (!$retried && alternates_changed($self)) {
130 $head =~ /^[0-9a-f]{40} \S+ (\d+)$/ or
131 fail($self, "Unexpected result from git cat-file: $head");
134 my $ref_type = $ref ? ref($ref) : '';
138 $$ref = $size if ($ref_type eq 'SCALAR');
141 if ($ref_type eq 'CODE') {
142 $rv = eval { $ref->($in, \$left) };
147 my $r = read($in, my $x, $left > $max ? $max : $left);
148 defined($r) or fail($self, "read failed: $!");
149 $r == 0 and fail($self, 'exited unexpectedly');
156 my $r = read($in, $buf, $left, $offset);
157 defined($r) or fail($self, "read failed: $!");
158 $r == 0 and fail($self, 'exited unexpectedly');
165 my $r = read($in, my $buf, 1);
166 defined($r) or fail($self, "read failed: $!");
167 fail($self, 'newline missing after blob') if ($r != 1 || $buf ne "\n");
168 die $cb_err if $cb_err;
173 sub batch_prepare ($) { _bidi_pipe($_[0], qw(--batch in out pid)) }
176 my ($self, $obj) = @_;
177 _bidi_pipe($self, qw(--batch-check in_c out_c pid_c err_c));
178 $self->{out_c}->print($obj, "\n") or fail($self, "write error: $!");
180 chomp(my $line = $self->{in_c}->getline);
181 my ($hex, $type, $size) = split(' ', $line);
182 return if $type eq 'missing';
184 # "dead" in git.git shows "dangling 4\ndead\n", not sure why
185 # https://public-inbox.org/git/20190118033845.s2vlrb3wd3m2jfzu@dcvr/
186 # so handle the oddball stuff just in case
187 if ($hex eq 'dangling' || $hex eq 'notdir' || $hex eq 'loop') {
188 $size = $type + length("\n");
189 my $r = read($self->{in_c}, my $buf, $size);
190 defined($r) or fail($self, "read failed: $!");
194 ($hex, $type, $size);
198 my ($self, $in, $out, $pid) = @_;
199 my $p = delete $self->{$pid} or return;
200 foreach my $f ($in, $out) {
207 my ($self, $msg) = @_;
213 my ($self, @cmd) = @_;
214 @cmd = ('git', "--git-dir=$self->{git_dir}", @cmd);
219 my ($self, @cmd) = @_;
220 my $fh = $self->popen(@cmd);
221 defined $fh or return;
223 return <$fh> if wantarray;
230 _destroy($self, qw(in out pid));
231 _destroy($self, qw(in_c out_c pid_c));
234 # assuming a well-maintained repo, this should be a somewhat
235 # accurate estimation of its size
236 # TODO: show this in the WWW UI as a hint to potential cloners
240 foreach my $p (glob("$self->{git_dir}/objects/pack/*.pack")) {
246 sub DESTROY { cleanup(@_) }
251 # don't show full FS path, basename should be OK:
252 if ($self->{git_dir} =~ m!/([^/]+)(?:/\.git)?\z!) {
253 $ret = "/path/to/$1";
255 wantarray ? ($ret) : $ret;
258 # show the blob URL for cgit/gitweb/whatever
260 my ($self, $oid) = @_;
261 # blob_url_format = "https://example.com/foo.git/blob/%s"
262 if (my $bfu = $self->{blob_url_format}) {
263 return map { sprintf($_, $oid) } @$bfu if wantarray;
264 return sprintf($bfu->[0], $oid);
271 if (my $urls = $self->{cgit_url}) {
283 PublicInbox::Git - git wrapper
291 use PublicInbox::Git;
292 chomp(my $git_dir = `git rev-parse --git-dir`);
293 $git_dir or die "GIT_DIR= must be specified\n";
294 my $git = PublicInbox::Git->new($git_dir);
298 Unstable API outside of the L</new> method.
299 It requires L<git(1)> to be installed.
307 my $git = PublicInbox::Git->new($git_dir);
309 Initialize a new PublicInbox::Git object for use with L<PublicInbox::Import>
310 This is the only public API method we support. Everything else
311 in this module is subject to change.
315 L<Git>, L<PublicInbox::Import>
319 All feedback welcome via plain-text mail to L<mailto:meta@public-inbox.org>
321 The mail archives are hosted at L<https://public-inbox.org/meta/>
325 Copyright (C) 2016 all contributors L<mailto:meta@public-inbox.org>
327 License: AGPL-3.0+ L<http://www.gnu.org/licenses/agpl-3.0.txt>