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 {-tmp} field for File::Temp::Dir
54 bless { git_dir => $git_dir, st => \@st, -git_path => {} }, $class
58 my ($self, $path) = @_;
59 $self->{-git_path}->{$path} ||= do {
61 chomp(my $str = $self->qx(qw(rev-parse --git-path), $path));
66 sub alternates_changed {
68 my $alt = git_path($self, 'objects/info/alternates');
69 my @st = stat($alt) or return 0;
70 my $old_st = $self->{st};
71 # 10 - ctime, 7 - size
72 return 0 if ($st[10] == $old_st->[10] && $st[7] == $old_st->[7]);
78 my $fh = $self->{err_c} or return;
79 sysseek($fh, 0, 0) or fail($self, "sysseek failed: $!");
80 defined(sysread($fh, my $buf, -s $fh)) or
81 fail($self, "sysread failed: $!");
86 my ($self, $batch, $in, $out, $pid, $err) = @_;
88 if (defined $err) { # "err_c"
89 my $fh = $self->{$err};
90 sysseek($fh, 0, 0) or fail($self, "sysseek failed: $!");
91 truncate($fh, 0) or fail($self, "truncate failed: $!");
95 my ($in_r, $in_w, $out_r, $out_w);
97 pipe($in_r, $in_w) or fail($self, "pipe failed: $!");
98 pipe($out_r, $out_w) or fail($self, "pipe failed: $!");
99 if ($^O eq 'linux') { # 1031: F_SETPIPE_SZ
100 fcntl($out_w, 1031, 4096);
101 fcntl($in_w, 1031, 4096) if $batch eq '--batch-check';
104 my @cmd = (qw(git), "--git-dir=$self->{git_dir}",
105 qw(-c core.abbrev=40 cat-file), $batch);
106 my $redir = { 0 => fileno($out_r), 1 => fileno($in_w) };
108 open(my $fh, '+>', undef) or fail($self, "open.err failed: $!");
110 $redir->{2} = fileno($fh);
112 my $p = spawn(\@cmd, undef, $redir);
113 defined $p or fail($self, "spawn failed: $!");
115 $out_w->autoflush(1);
116 $self->{$out} = $out_w;
117 $self->{$in} = $in_r;
121 my ($self, $obj, $ref) = @_;
122 my ($retried, $in, $head);
125 batch_prepare($self);
126 $self->{out}->print($obj, "\n") or fail($self, "write error: $!");
130 $head = $in->getline;
131 if ($head =~ / missing$/) {
132 if (!$retried && alternates_changed($self)) {
139 $head =~ /^[0-9a-f]{40} \S+ (\d+)$/ or
140 fail($self, "Unexpected result from git cat-file: $head");
143 my $ref_type = $ref ? ref($ref) : '';
147 $$ref = $size if ($ref_type eq 'SCALAR');
150 if ($ref_type eq 'CODE') {
151 $rv = eval { $ref->($in, \$left) };
156 my $r = read($in, my $x, $left > $max ? $max : $left);
157 defined($r) or fail($self, "read failed: $!");
158 $r == 0 and fail($self, 'exited unexpectedly');
165 my $r = read($in, $buf, $left, $offset);
166 defined($r) or fail($self, "read failed: $!");
167 $r == 0 and fail($self, 'exited unexpectedly');
174 my $r = read($in, my $buf, 1);
175 defined($r) or fail($self, "read failed: $!");
176 fail($self, 'newline missing after blob') if ($r != 1 || $buf ne "\n");
177 die $cb_err if $cb_err;
182 sub batch_prepare ($) { _bidi_pipe($_[0], qw(--batch in out pid)) }
185 my ($self, $obj) = @_;
186 _bidi_pipe($self, qw(--batch-check in_c out_c pid_c err_c));
187 $self->{out_c}->print($obj, "\n") or fail($self, "write error: $!");
189 chomp(my $line = $self->{in_c}->getline);
190 my ($hex, $type, $size) = split(' ', $line);
192 # Future versions of git.git may show 'ambiguous', but for now,
193 # we must handle 'dangling' below (and maybe some other oddball
195 # https://public-inbox.org/git/20190118033845.s2vlrb3wd3m2jfzu@dcvr/T/
196 return if $type eq 'missing' || $type eq 'ambiguous';
198 if ($hex eq 'dangling' || $hex eq 'notdir' || $hex eq 'loop') {
199 $size = $type + length("\n");
200 my $r = read($self->{in_c}, my $buf, $size);
201 defined($r) or fail($self, "read failed: $!");
205 ($hex, $type, $size);
209 my ($self, $in, $out, $pid, $expire) = @_;
210 my $rfh = $self->{$in} or return;
211 if (defined $expire) {
212 # at least FreeBSD 11.2 and Linux 4.20 update mtime of the
213 # read end of a pipe when the pipe is written to; dunno
215 my $mtime = (stat($rfh))[9];
216 return if $mtime > $expire;
218 my $p = delete $self->{$pid} or return;
219 foreach my $f ($in, $out) {
226 my ($self, $msg) = @_;
232 my ($self, @cmd) = @_;
233 @cmd = ('git', "--git-dir=$self->{git_dir}", @cmd);
238 my ($self, @cmd) = @_;
239 my $fh = $self->popen(@cmd);
240 defined $fh or return;
242 return <$fh> if wantarray;
247 # returns true if there are pending "git cat-file" processes
249 my ($self, $expire) = @_;
250 _destroy($self, qw(in out pid), $expire);
251 _destroy($self, qw(in_c out_c pid_c), $expire);
252 !!($self->{pid} || $self->{pid_c});
255 # assuming a well-maintained repo, this should be a somewhat
256 # accurate estimation of its size
257 # TODO: show this in the WWW UI as a hint to potential cloners
261 my $pack_dir = git_path($self, 'objects/pack');
262 foreach my $p (glob("$pack_dir/*.pack")) {
268 sub DESTROY { cleanup(@_) }
273 # don't show full FS path, basename should be OK:
274 if ($self->{git_dir} =~ m!/([^/]+)(?:/\.git)?\z!) {
275 $ret = "/path/to/$1";
277 wantarray ? ($ret) : $ret;
280 # show the blob URL for cgit/gitweb/whatever
282 my ($self, $oid) = @_;
283 # blob_url_format = "https://example.com/foo.git/blob/%s"
284 if (my $bfu = $self->{blob_url_format}) {
285 return map { sprintf($_, $oid) } @$bfu if wantarray;
286 return sprintf($bfu->[0], $oid);
291 sub host_prefix_url ($$) {
292 my ($env, $url) = @_;
293 return $url if index($url, '//') >= 0;
294 my $scheme = $env->{'psgi.url_scheme'};
295 my $host_port = $env->{HTTP_HOST} ||
296 "$env->{SERVER_NAME}:$env->{SERVER_PORT}";
297 "$scheme://$host_port". ($env->{SCRIPT_NAME} || '/') . $url;
301 my ($self, $env) = @_;
302 if (my $urls = $self->{cgit_url}) {
303 return map { host_prefix_url($env, $_) } @$urls;
308 sub commit_title ($$) {
309 my ($self, $oid) = @_; # PublicInbox::Git, $sha1hex
310 my $buf = cat_file($self, $oid) or return;
312 ($$buf =~ /\r?\n\r?\n([^\r\n]+)\r?\n?/)[0]
321 PublicInbox::Git - git wrapper
329 use PublicInbox::Git;
330 chomp(my $git_dir = `git rev-parse --git-dir`);
331 $git_dir or die "GIT_DIR= must be specified\n";
332 my $git = PublicInbox::Git->new($git_dir);
336 Unstable API outside of the L</new> method.
337 It requires L<git(1)> to be installed.
345 my $git = PublicInbox::Git->new($git_dir);
347 Initialize a new PublicInbox::Git object for use with L<PublicInbox::Import>
348 This is the only public API method we support. Everything else
349 in this module is subject to change.
353 L<Git>, L<PublicInbox::Import>
357 All feedback welcome via plain-text mail to L<mailto:meta@public-inbox.org>
359 The mail archives are hosted at L<https://public-inbox.org/meta/>
363 Copyright (C) 2016 all contributors L<mailto:meta@public-inbox.org>
365 License: AGPL-3.0+ L<http://www.gnu.org/licenses/agpl-3.0.txt>