X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FGit.pm;h=86b80a4edb877152f566c00a7b804952516e907b;hb=d4ba8828ab23f2785be54493495bbf7e1d62c0b0;hp=3d97300c43c3f0fca345d6719d925d3b9ad652eb;hpb=9dfc0b670fc634b54998c3020f173b82de1915ac;p=public-inbox.git diff --git a/lib/PublicInbox/Git.pm b/lib/PublicInbox/Git.pm index 3d97300c..86b80a4e 100644 --- a/lib/PublicInbox/Git.pm +++ b/lib/PublicInbox/Git.pm @@ -1,4 +1,4 @@ -# Copyright (C) 2014-2021 all contributors +# Copyright (C) all contributors # License: GPLv2 or later # # Used to read files from a git repository without excessive forking. @@ -12,25 +12,26 @@ use v5.10.1; use parent qw(Exporter); use POSIX (); use IO::Handle; # ->autoflush -use Errno qw(EINTR EAGAIN); +use Errno qw(EINTR EAGAIN ENOENT); use File::Glob qw(bsd_glob GLOB_NOSORT); use File::Spec (); use Time::HiRes qw(stat); -use PublicInbox::Spawn qw(popen_rd); +use PublicInbox::Spawn qw(popen_rd spawn); use PublicInbox::Tmpfile; use IO::Poll qw(POLLIN); -use Carp qw(croak); +use Carp qw(croak carp); use Digest::SHA (); use PublicInbox::DS qw(dwaitpid); our @EXPORT_OK = qw(git_unquote git_quote); our $PIPE_BUFSIZ = 65536; # Linux default our $in_cleanup; our $RDTIMEO = 60_000; # milliseconds +our $async_warn; # true in read-only daemons -use constant MAX_INFLIGHT => - (($^O eq 'linux' ? 4096 : POSIX::_POSIX_PIPE_BUF()) * 3) - / - 65; # SHA-256 hex size + "\n" in preparation for git using non-SHA1 +# 512: POSIX PIPE_BUF minimum (see pipe(7)) +# 3: @$inflight is flattened [ $OID, $cb, $arg ] +# 65: SHA-256 hex size + "\n" in preparation for git using non-SHA1 +use constant MAX_INFLIGHT => 512 * 3 / 65; my %GIT_ESC = ( a => "\a", @@ -50,14 +51,13 @@ my %ESC_GIT = map { $GIT_ESC{$_} => $_ } keys %GIT_ESC; sub git_unquote ($) { return $_[0] unless ($_[0] =~ /\A"(.*)"\z/); $_[0] = $1; - $_[0] =~ s/\\([\\"abfnrtv])/$GIT_ESC{$1}/g; - $_[0] =~ s/\\([0-7]{1,3})/chr(oct($1))/ge; + $_[0] =~ s!\\([\\"abfnrtv]|[0-3][0-7]{2})!$GIT_ESC{$1}//chr(oct($1))!ge; $_[0]; } sub git_quote ($) { if ($_[0] =~ s/([\\"\a\b\f\n\r\t\013]|[^[:print:]])/ - '\\'.($ESC_GIT{$1}||sprintf("%0o",ord($1)))/egs) { + '\\'.($ESC_GIT{$1}||sprintf("%03o",ord($1)))/egs) { return qq{"$_[0]"}; } $_[0]; @@ -65,13 +65,15 @@ sub git_quote ($) { sub new { my ($class, $git_dir) = @_; + $git_dir =~ tr!/!/!s; + $git_dir =~ s!/*\z!!s; # may contain {-tmp} field for File::Temp::Dir - bless { git_dir => $git_dir, alt_st => '', -git_path => {} }, $class + bless { git_dir => $git_dir }, $class } sub git_path ($$) { my ($self, $path) = @_; - $self->{-git_path}->{$path} ||= do { + $self->{-git_path}->{$path} //= do { local $/ = "\n"; chomp(my $str = $self->qx(qw(rev-parse --git-path), $path)); @@ -90,10 +92,17 @@ sub alternates_changed { # can't rely on 'q' on some 32-bit builds, but `d' works my $st = pack('dd', $st[10], $st[7]); # 10: ctime, 7: size - return 0 if $self->{alt_st} eq $st; + return 0 if ($self->{alt_st} // '') eq $st; $self->{alt_st} = $st; # always a true value } +sub object_format { + $_[0]->{object_format} //= do { + my $fmt = $_[0]->qx(qw(config extensions.objectformat)); + $fmt eq "sha256\n" ? \'sha256' : \undef; + } +} + sub last_check_err { my ($self) = @_; my $fh = $self->{err_c} or return; @@ -113,18 +122,22 @@ sub _bidi_pipe { } return; } - my ($out_r, $out_w); - pipe($out_r, $out_w) or $self->fail("pipe failed: $!"); - my @cmd = (qw(git), "--git-dir=$self->{git_dir}", + pipe(my ($out_r, $out_w)) or $self->fail("pipe failed: $!"); + my $rdr = { 0 => $out_r, pgid => 0 }; + my $gd = $self->{git_dir}; + if ($gd =~ s!/([^/]+/[^/]+)\z!/!) { + $rdr->{-C} = $gd; + $gd = $1; + } + my @cmd = (qw(git), "--git-dir=$gd", qw(-c core.abbrev=40 cat-file), $batch); - my $redir = { 0 => $out_r }; if ($err) { my $id = "git.$self->{git_dir}$batch.err"; my $fh = tmpfile($id) or $self->fail("tmpfile($id): $!"); $self->{$err} = $fh; - $redir->{2} = $fh; + $rdr->{2} = $fh; } - my ($in_r, $p) = popen_rd(\@cmd, undef, $redir); + my ($in_r, $p) = popen_rd(\@cmd, undef, $rdr); $self->{$pid} = $p; $self->{"$pid.owner"} = $$; $out_w->autoflush(1); @@ -132,6 +145,7 @@ sub _bidi_pipe { fcntl($out_w, 1031, 4096); fcntl($in_r, 1031, 4096) if $batch eq '--batch-check'; } + $out_w->blocking(0); $self->{$out} = $out_w; $self->{$in} = $in_r; } @@ -154,7 +168,8 @@ sub my_read ($$$) { return; # unrecoverable error } } - \substr($$rbuf, 0, $len, ''); + my $no_pad = substr($$rbuf, 0, $len, ''); + \$no_pad; } sub my_readline ($$) { @@ -175,8 +190,8 @@ sub my_readline ($$) { } } -sub cat_async_retry ($$$$$) { - my ($self, $inflight, $req, $cb, $arg) = @_; +sub cat_async_retry ($$) { + my ($self, $inflight) = @_; # {inflight} may be non-existent, but if it isn't we delete it # here to prevent cleanup() from waiting: @@ -185,12 +200,15 @@ sub cat_async_retry ($$$$$) { $self->{inflight} = $inflight; batch_prepare($self); - my $buf = "$req\n"; + my $buf = ''; for (my $i = 0; $i < @$inflight; $i += 3) { $buf .= "$inflight->[$i]\n"; } + $self->{out}->blocking(1); # brand new pipe, should never block print { $self->{out} } $buf or $self->fail("write error: $!"); - unshift(@$inflight, \$req, $cb, $arg); # \$ref to indicate retried + $self->{out}->blocking(0); + my $req = shift @$inflight; + unshift(@$inflight, \$req); # \$ref to indicate retried cat_async_step($self, $inflight); # take one step } @@ -198,8 +216,8 @@ sub cat_async_retry ($$$$$) { sub cat_async_step ($$) { my ($self, $inflight) = @_; die 'BUG: inflight empty or odd' if scalar(@$inflight) < 3; - my ($req, $cb, $arg) = splice(@$inflight, 0, 3); - my $rbuf = delete($self->{cat_rbuf}) // \(my $new = ''); + my ($req, $cb, $arg) = @$inflight[0, 1, 2]; + my $rbuf = delete($self->{rbuf}) // \(my $new = ''); my ($bref, $oid, $type, $size); my $head = my_readline($self->{in}, $rbuf); # ->fail may be called via Gcf2Client.pm @@ -213,8 +231,7 @@ sub cat_async_step ($$) { # ref($req) indicates it's already been retried # -gcf2 retries internally, so it never hits this path: if (!ref($req) && !$in_cleanup && $self->alternates_changed) { - return cat_async_retry($self, $inflight, - $req, $cb, $arg); + return cat_async_retry($self, $inflight); } $type = 'missing'; $oid = ref($req) ? $$req : $req if $oid eq ''; @@ -222,9 +239,10 @@ sub cat_async_step ($$) { my $err = $! ? " ($!)" : ''; $self->fail("bad result from async cat-file: $head$err"); } - $self->{cat_rbuf} = $rbuf if $$rbuf ne ''; + $self->{rbuf} = $rbuf if $$rbuf ne ''; + splice(@$inflight, 0, 3); # don't retry $cb on ->fail eval { $cb->($bref, $oid, $type, $size, $arg) }; - warn "E: $oid: $@\n" if $@; + async_err($self, $req, $oid, $@, 'cat') if $@; } sub cat_async_wait ($) { @@ -240,24 +258,23 @@ sub batch_prepare ($) { } sub _cat_file_cb { - my ($bref, undef, undef, $size, $result) = @_; - @$result = ($bref, $size); + my ($bref, $oid, $type, $size, $result) = @_; + @$result = ($bref, $oid, $type, $size); } sub cat_file { - my ($self, $oid, $sizeref) = @_; + my ($self, $oid) = @_; my $result = []; cat_async($self, $oid, \&_cat_file_cb, $result); cat_async_wait($self); - $$sizeref = $result->[1] if $sizeref; - $result->[0]; + wantarray ? @$result : $result->[0]; } sub check_async_step ($$) { my ($self, $inflight_c) = @_; die 'BUG: inflight empty or odd' if scalar(@$inflight_c) < 3; - my ($req, $cb, $arg) = splice(@$inflight_c, 0, 3); - my $rbuf = delete($self->{chk_rbuf}) // \(my $new = ''); + my ($req, $cb, $arg) = @$inflight_c[0, 1, 2]; + my $rbuf = delete($self->{rbuf_c}) // \(my $new = ''); chomp(my $line = my_readline($self->{in_c}, $rbuf)); my ($hex, $type, $size) = split(/ /, $line); @@ -269,9 +286,10 @@ sub check_async_step ($$) { my $ret = my_read($self->{in_c}, $rbuf, $type + 1); $self->fail(defined($ret) ? 'read EOF' : "read: $!") if !$ret; } - $self->{chk_rbuf} = $rbuf if $$rbuf ne ''; + $self->{rbuf_c} = $rbuf if $$rbuf ne ''; + splice(@$inflight_c, 0, 3); # don't retry $cb on ->fail eval { $cb->($hex, $type, $size, $arg, $self) }; - warn "E: check($req) $@\n" if $@; + async_err($self, $req, $hex, $@, 'check') if $@; } sub check_async_wait ($) { @@ -290,13 +308,27 @@ sub check_async_begin ($) { $self->{inflight_c} = []; } +sub write_all { + my ($self, $out, $buf, $read_step, $inflight) = @_; + $read_step->($self, $inflight) while @$inflight >= MAX_INFLIGHT; + do { + my $w = syswrite($out, $buf); + if (defined $w) { + return if $w == length($buf); + warn "chop: $w"; + substr($buf, 0, $w, ''); # sv_chop + } elsif ($! != EAGAIN) { + $self->fail("write: $!"); + } else { warn "E: $!" } + $read_step->($self, $inflight); + } while (1); +} + sub check_async ($$$$) { my ($self, $oid, $cb, $arg) = @_; my $inflight_c = $self->{inflight_c} // check_async_begin($self); - while (scalar(@$inflight_c) >= MAX_INFLIGHT) { - check_async_step($self, $inflight_c); - } - print { $self->{out_c} } $oid, "\n" or $self->fail("write error: $!"); + write_all($self, $self->{out_c}, $oid."\n", + \&check_async_step, $inflight_c); push(@$inflight_c, $oid, $cb, $arg); } @@ -331,27 +363,39 @@ sub _destroy { dwaitpid($p) if $$ == $self->{"$pid.owner"}; } -sub cat_async_abort ($) { +sub async_abort ($) { my ($self) = @_; - if (my $inflight = $self->{inflight}) { - while (@$inflight) { - my ($req, $cb, $arg) = splice(@$inflight, 0, 3); - $req =~ s/ .*//; # drop git_dir for Gcf2Client - eval { $cb->(undef, $req, undef, undef, $arg) }; - warn "E: $req: $@ (in abort)\n" if $@; + while (scalar(@{$self->{inflight_c} // []}) || + scalar(@{$self->{inflight} // []})) { + for my $c ('', '_c') { + my $q = $self->{"inflight$c"} or next; + while (@$q) { + my ($req, $cb, $arg) = splice(@$q, 0, 3); + $req = $$req if ref($req); + $req =~ s/ .*//; # drop git_dir for Gcf2Client + eval { $cb->(undef, $req, undef, undef, $arg) }; + warn "E: (in abort) $req: $@" if $@; + } + delete $self->{"inflight$c"}; + delete $self->{"rbuf$c"}; } - delete $self->{cat_rbuf}; - delete $self->{inflight}; } cleanup($self); } sub fail { # may be augmented in subclasses my ($self, $msg) = @_; - cat_async_abort($self); + async_abort($self); croak(ref($self) . ' ' . ($self->{git_dir} // '') . ": $msg"); } +sub async_err ($$$$$) { + my ($self, $req, $oid, $err, $action) = @_; + $req = $$req if ref($req); # retried + my $msg = "E: $action $req ($oid): $err"; + $async_warn ? carp($msg) : $self->fail($msg); +} + # $git->popen(qw(show f00)); # or # $git->popen(qw(show f00), { GIT_CONFIG => ... }, { 2 => ... }); sub popen { @@ -363,10 +407,8 @@ sub popen { # same args as popen above sub qx { - my $self = shift; - my $fh = $self->popen(@_); + my $fh = popen(@_); if (wantarray) { - local $/ = "\n"; my @ret = <$fh>; close $fh; # caller should check $? @ret; @@ -378,31 +420,40 @@ sub qx { } } +sub date_parse { + my $self = shift; + map { + substr($_, length('--max-age='), -1) + } $self->qx('rev-parse', map { "--since=$_" } @_); +} + # check_async and cat_async may trigger the other, so ensure they're # both completely done by using this: sub async_wait_all ($) { my ($self) = @_; while (scalar(@{$self->{inflight_c} // []}) || scalar(@{$self->{inflight} // []})) { - $self->check_async_wait; - $self->cat_async_wait; + check_async_wait($self); + cat_async_wait($self); } } # returns true if there are pending "git cat-file" processes sub cleanup { - my ($self) = @_; + my ($self, $lazy) = @_; + return 1 if $lazy && (scalar(@{$self->{inflight_c} // []}) || + scalar(@{$self->{inflight} // []})); local $in_cleanup = 1; delete $self->{async_cat}; + delete $self->{async_chk}; async_wait_all($self); delete $self->{inflight}; delete $self->{inflight_c}; - _destroy($self, qw(cat_rbuf in out pid)); - _destroy($self, qw(chk_rbuf in_c out_c pid_c err_c)); - !!($self->{pid} || $self->{pid_c}); + _destroy($self, qw(rbuf in out pid)); + _destroy($self, qw(rbuf_c in_c out_c pid_c err_c)); + undef; } - # assuming a well-maintained repo, this should be a somewhat # accurate estimation of its size # TODO: show this in the WWW UI as a hint to potential cloners @@ -419,13 +470,9 @@ sub packed_bytes { sub DESTROY { cleanup(@_) } sub local_nick ($) { - my ($self) = @_; - my $ret = '???'; # don't show full FS path, basename should be OK: - if ($self->{git_dir} =~ m!/([^/]+)(?:/\.git)?\z!) { - $ret = "/path/to/$1"; - } - wantarray ? ($ret) : $ret; + $_[0]->{nick} // ($_[0]->{git_dir} =~ m!/([^/]+?)(?:/*\.git/*)?\z! ? + "$1.git" : undef); } sub host_prefix_url ($$) { @@ -437,12 +484,22 @@ sub host_prefix_url ($$) { "$scheme://$host_port". ($env->{SCRIPT_NAME} || '/') . $url; } +sub base_url { # for coderepos, PSGI-only + my ($self, $env) = @_; # env - PSGI env + my $url = host_prefix_url($env, ''); + # for mount in Plack::Builder + $url .= '/' if substr($url, -1, 1) ne '/'; + $url . $self->{nick} . '/'; +} + +sub isrch {} # TODO + sub pub_urls { my ($self, $env) = @_; if (my $urls = $self->{cgit_url}) { return map { host_prefix_url($env, $_) } @$urls; } - local_nick($self); + (local_nick($self) // '???'); } sub cat_async_begin { @@ -456,49 +513,44 @@ sub cat_async_begin { sub cat_async ($$$;$) { my ($self, $oid, $cb, $arg) = @_; my $inflight = $self->{inflight} // cat_async_begin($self); - while (scalar(@$inflight) >= MAX_INFLIGHT) { - cat_async_step($self, $inflight); - } - print { $self->{out} } $oid, "\n" or $self->fail("write error: $!"); + write_all($self, $self->{out}, $oid."\n", \&cat_async_step, $inflight); push(@$inflight, $oid, $cb, $arg); } -sub async_prefetch { - my ($self, $oid, $cb, $arg) = @_; - if (my $inflight = $self->{inflight}) { - # we could use MAX_INFLIGHT here w/o the halving, - # but lets not allow one client to monopolize a git process - if (scalar(@$inflight) < int(MAX_INFLIGHT/2)) { - print { $self->{out} } $oid, "\n" or - $self->fail("write error: $!"); - return push(@$inflight, $oid, $cb, $arg); - } - } - undef; +# returns the modified time of a git repo, same as the "modified" field +# of a grokmirror manifest +sub modified ($) { + # committerdate:unix is git 2.9.4+ (2017-05-05), so using raw instead + my $fh = popen($_[0], qw[for-each-ref --sort=-committerdate + --format=%(committerdate:raw) --count=1]); + (split(/ /, <$fh> // time))[0] + 0; # integerize for JSON } -sub extract_cmt_time { - my ($bref, undef, undef, undef, $modified) = @_; +sub try_cat { + my ($path) = @_; + open(my $fh, '<', $path) or return ''; + local $/; + <$fh> // ''; +} - if ($$bref =~ /^committer .*?> ([0-9]+) [\+\-]?[0-9]+/sm) { - my $cmt_time = $1 + 0; - $$modified = $cmt_time if $cmt_time > $$modified; - } +sub cat_desc ($) { + my $desc = try_cat($_[0]); + chomp $desc; + utf8::decode($desc); + $desc =~ s/\s+/ /smg; + $desc eq '' ? undef : $desc; } -# returns the modified time of a git repo, same as the "modified" field -# of a grokmirror manifest -sub modified ($) { - my ($self) = @_; - my $modified = 0; - my $fh = popen($self, qw(rev-parse --branches)); - local $/ = "\n"; - while (my $oid = <$fh>) { - chomp $oid; - cat_async($self, $oid, \&extract_cmt_time, \$modified); - } - cat_async_wait($self); - $modified || time; +sub description { + cat_desc("$_[0]->{git_dir}/description") // 'Unnamed repository'; +} + +sub cloneurl { + my ($self, $env) = @_; + $self->{cloneurl} // do { + my @urls = split(/\s+/s, try_cat("$self->{git_dir}/cloneurl")); + scalar(@urls) ? ($self->{cloneurl} = \@urls) : undef; + } // [ substr(base_url($self, $env), 0, -1) ]; } # for grokmirror, which doesn't read gitweb.description @@ -523,14 +575,8 @@ sub manifest_entry { chomp(my $owner = $self->qx('config', 'gitweb.owner')); utf8::decode($owner); $ent->{owner} = $owner eq '' ? undef : $owner; - my $desc = ''; - if (open($fh, '<', "$git_dir/description")) { - local $/ = "\n"; - chomp($desc = <$fh>); - utf8::decode($desc); - } - $desc = 'Unnamed repository' if $desc eq ''; - if (defined $epoch && $desc =~ /\AUnnamed repository/) { + my $desc = description($self); + if (defined $epoch && index($desc, 'Unnamed repository') == 0) { $desc = "$default_desc [epoch $epoch]"; } $ent->{description} = $desc; @@ -551,6 +597,28 @@ sub manifest_entry { $ent; } +# returns true if there are pending cat-file processes +sub cleanup_if_unlinked { + my ($self) = @_; + return cleanup($self, 1) if $^O ne 'linux'; + # Linux-specific /proc/$PID/maps access + # TODO: support this inside git.git + my $ret = 0; + for my $fld (qw(pid pid_c)) { + my $pid = $self->{$fld} // next; + open my $fh, '<', "/proc/$pid/maps" or return cleanup($self, 1); + while (<$fh>) { + # n.b. we do not restart for unlinked multi-pack-index + # since it's not too huge, and the startup cost may + # be higher. + /\.(?:idx|pack) \(deleted\)$/ and + return cleanup($self, 1); + } + ++$ret; + } + $ret; +} + 1; __END__ =pod