X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FImport.pm;h=fc4e72dc93ea0ce978a93f5224c74e5b8178acc3;hb=6e6f7999361925e4c944f308df4bc32a1842cd69;hp=e07eddab4f245a2a47f368d5dd223d26e2a56770;hpb=9aaf9cf73f7b126ad3d2726adf7c764630626d67;p=public-inbox.git diff --git a/lib/PublicInbox/Import.pm b/lib/PublicInbox/Import.pm index e07eddab..fc4e72dc 100644 --- a/lib/PublicInbox/Import.pm +++ b/lib/PublicInbox/Import.pm @@ -1,9 +1,10 @@ -# Copyright (C) 2016-2018 all contributors +# Copyright (C) 2016-2019 all contributors # License: AGPL-3.0+ # # git fast-import-based ssoma-mda MDA replacement -# This is only ever run by public-inbox-mda and public-inbox-learn, -# not the WWW or NNTP code which only requires read-only access. +# This is only ever run by public-inbox-mda, public-inbox-learn +# and public-inbox-watch. Not the WWW or NNTP code which only +# requires read-only access. package PublicInbox::Import; use strict; use warnings; @@ -14,6 +15,7 @@ use PublicInbox::Address; use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp); use PublicInbox::ContentId qw(content_digest); use PublicInbox::MDA; +use POSIX qw(strftime); sub new { my ($class, $git, $name, $email, $ibx) = @_; @@ -49,7 +51,14 @@ sub gfi_start { $self->lock_acquire; local $/ = "\n"; - chomp($self->{tip} = $git->qx(qw(rev-parse --revs-only), $self->{ref})); + my $ref = $self->{ref}; + chomp($self->{tip} = $git->qx(qw(rev-parse --revs-only), $ref)); + if ($self->{path_type} ne '2/38' && $self->{tip}) { + local $/ = "\0"; + my @tree = $git->qx(qw(ls-tree -r -z --name-only), $ref); + chomp @tree; + $self->{-tree} = { map { $_ => 1 } @tree }; + } my $git_dir = $git->{git_dir}; my @cmd = ('git', "--git-dir=$git_dir", qw(fast-import @@ -88,19 +97,13 @@ sub _check_path ($$$$) { $info =~ /\Amissing / ? undef : $info; } -sub check_remove_v1 { - my ($r, $w, $tip, $path, $mime) = @_; - - my $info = _check_path($r, $w, $tip, $path) or return ('MISSING',undef); - $info =~ m!\A100644 blob ([a-f0-9]{40})\t!s or die "not blob: $info"; - my $blob = $1; - - print $w "cat-blob $blob\n" or wfail; +sub _cat_blob ($$$) { + my ($r, $w, $oid) = @_; + print $w "cat-blob $oid\n" or wfail; local $/ = "\n"; - $info = <$r>; + my $info = <$r>; defined $info or die "EOF from fast-import / cat-blob: $!"; - $info =~ /\A[a-f0-9]{40} blob (\d+)\n\z/ or - die "unexpected cat-blob response: $info"; + $info =~ /\A[a-f0-9]{40} blob (\d+)\n\z/ or return; my $left = $1; my $offset = 0; my $buf = ''; @@ -115,7 +118,26 @@ sub check_remove_v1 { $n = read($r, my $lf, 1); defined($n) or die "read final byte of cat-blob failed: $!"; die "bad read on final byte: <$lf>" if $lf ne "\n"; - my $cur = PublicInbox::MIME->new(\$buf); + + # fixup some bugginess in old versions: + $buf =~ s/\A[\r\n]*From [^\r\n]*\r?\n//s; + \$buf; +} + +sub cat_blob { + my ($self, $oid) = @_; + my ($r, $w) = $self->gfi_start; + _cat_blob($r, $w, $oid); +} + +sub check_remove_v1 { + my ($r, $w, $tip, $path, $mime) = @_; + + my $info = _check_path($r, $w, $tip, $path) or return ('MISSING',undef); + $info =~ m!\A100644 blob ([a-f0-9]{40})\t!s or die "not blob: $info"; + my $oid = $1; + my $msg = _cat_blob($r, $w, $oid) or die "BUG: cat-blob $1 failed"; + my $cur = PublicInbox::MIME->new($msg); my $cur_s = $cur->header('Subject'); $cur_s = '' unless defined $cur_s; my $cur_m = $mime->header('Subject'); @@ -156,8 +178,8 @@ sub _update_git_info ($$) { run_die([@cmd, 'update-server-info'], undef); ($self->{path_type} eq '2/38') and eval { require PublicInbox::SearchIdx; - my $inbox = $self->{inbox} || $git_dir; - my $s = PublicInbox::SearchIdx->new($inbox); + my $ibx = $self->{inbox} || $git_dir; + my $s = PublicInbox::SearchIdx->new($ibx); $s->index_sync({ ref => $self->{ref} }); }; eval { run_die([@cmd, qw(gc --auto)], undef) } if $do_gc; @@ -185,6 +207,7 @@ sub get_mark { my ($r, $w) = $self->gfi_start; print $w "get-mark $mark\n" or wfail; defined(my $oid = <$r>) or die "get-mark failed, need git 2.6.0+\n"; + chomp($oid); $oid; } @@ -238,7 +261,8 @@ sub remove { if (defined $path) { print $w "D $path\n\n" or wfail; } else { - print $w "M 100644 :$blob _/D\n\n" or wfail; + clean_tree_v2($self, $w, 'd'); + print $w "M 100644 :$blob d\n\n" or wfail; } $self->{nchg}++; (($self->{tip} = ":$commit"), $cur); @@ -255,10 +279,12 @@ sub extract_author_info ($) { my $sender = ''; my $from = $mime->header('From'); + $from ||= ''; my ($email) = PublicInbox::Address::emails($from); my ($name) = PublicInbox::Address::names($from); if (!defined($name) || !defined($email)) { $sender = $mime->header('Sender'); + $sender ||= ''; if (!defined($name)) { ($name) = PublicInbox::Address::names($sender); } @@ -311,12 +337,21 @@ sub v1_mid0 ($) { my $mids = mids($hdr); if (!scalar(@$mids)) { # spam often has no Message-Id - my $mid0 = digest2mid(content_digest($mime)); + my $mid0 = digest2mid(content_digest($mime), $hdr); append_mid($hdr, $mid0); return $mid0; } $mids->[0]; } +sub clean_tree_v2 ($$$) { + my ($self, $w, $keep) = @_; + my $tree = $self->{-tree} or return; #v2 only + delete $tree->{$keep}; + foreach (keys %$tree) { + print $w "D $_\n" or wfail; + } + %$tree = ($keep => 1); +} # returns undef on duplicate # returns the :MARK of the most recent commit @@ -362,7 +397,7 @@ sub add { # v2: we need this for Xapian if ($self->{want_object_info}) { - chomp(my $oid = $self->get_mark(":$blob")); + my $oid = $self->get_mark(":$blob"); $self->{last_object} = [ $oid, $n, \$str ]; } my $ref = $self->{ref}; @@ -373,6 +408,8 @@ sub add { print $w "reset $ref\n" or wfail; } + # Mime decoding can create nulls replace them with spaces to protect git + $subject =~ tr/\0/ /; utf8::encode($subject); print $w "commit $ref\nmark :$commit\n", "author $name <$email> $author_time_raw\n", @@ -382,14 +419,15 @@ sub add { if ($tip ne '') { print $w 'from ', ($parent ? $parent : $tip), "\n" or wfail; } + clean_tree_v2($self, $w, $path); print $w "M 100644 :$blob $path\n\n" or wfail; $self->{nchg}++; $self->{tip} = ":$commit"; } -sub run_die ($;$) { - my ($cmd, $env) = @_; - my $pid = spawn($cmd, $env, undef); +sub run_die ($;$$) { + my ($cmd, $env, $rdr) = @_; + my $pid = spawn($cmd, $env, $rdr); defined $pid or die "spawning ".join(' ', @$cmd)." failed: $!"; waitpid($pid, 0) == $pid or die join(' ', @$cmd) .' did not finish'; $? == 0 or die join(' ', @$cmd) . " failed: $?\n"; @@ -407,27 +445,139 @@ sub done { _update_git_info($self, 1) if delete $self->{nchg}; $self->lock_release; + + $self->{git}->cleanup; } sub atfork_child { my ($self) = @_; foreach my $f (qw(in out)) { + next unless defined($self->{$f}); close $self->{$f} or die "failed to close import[$f]: $!\n"; } } -sub digest2mid ($) { - my ($dig) = @_; +sub digest2mid ($$) { + my ($dig, $hdr) = @_; my $b64 = $dig->clone->b64digest; # Make our own URLs nicer: # See "Base 64 Encoding with URL and Filename Safe Alphabet" in RFC4648 $b64 =~ tr!+/=!-_!d; - # We can make this more meaningful with a date prefix or other things, - # but this is only needed for crap that fails to generate a Message-ID - # or reuses one. In other words, it's usually spammers who hit this - # so they don't deserve nice Message-IDs :P - $b64 . '@localhost'; + # Add a date prefix to prevent a leading '-' in case that trips + # up some tools (e.g. if a Message-ID were a expected as a + # command-line arg) + my $dt = msg_datestamp($hdr); + $dt = POSIX::strftime('%Y%m%d%H%M%S', gmtime($dt)); + "$dt.$b64" . '@z'; +} + +sub clean_purge_buffer { + my ($oids, $buf) = @_; + my $cmt_msg = 'purged '.join(' ',@$oids)."\n"; + @$oids = (); + + foreach my $i (0..$#$buf) { + my $l = $buf->[$i]; + if ($l =~ /^author .* (\d+ [\+-]?\d+)$/) { + $buf->[$i] = "author <> $1\n"; + } elsif ($l =~ /^data (\d+)/) { + $buf->[$i++] = "data " . length($cmt_msg) . "\n"; + $buf->[$i] = $cmt_msg; + last; + } + } +} + +sub purge_oids { + my ($self, $purge) = @_; + my $tmp = "refs/heads/purge-".((keys %$purge)[0]); + my $old = $self->{'ref'}; + my $git = $self->{git}; + my @export = (qw(fast-export --no-data --use-done-feature), $old); + my $rd = $git->popen(@export); + my ($r, $w) = $self->gfi_start; + my @buf; + my $npurge = 0; + my @oids; + my ($done, $mark); + my $tree = $self->{-tree}; + while (<$rd>) { + if (/^reset (?:.+)/) { + push @buf, "reset $tmp\n"; + } elsif (/^commit (?:.+)/) { + if (@buf) { + $w->print(@buf) or wfail; + @buf = (); + } + push @buf, "commit $tmp\n"; + } elsif (/^data (\d+)/) { + # only commit message, so $len is small: + my $len = $1; # + 1 for trailing "\n" + push @buf, $_; + my $n = read($rd, my $buf, $len) or die "read: $!"; + $len == $n or die "short read ($n < $len)"; + push @buf, $buf; + } elsif (/^M 100644 ([a-f0-9]+) (\w+)/) { + my ($oid, $path) = ($1, $2); + $tree->{$path} = 1; + if ($purge->{$oid}) { + push @oids, $oid; + my $cmd = "M 100644 inline $path\ndata 0\n\n"; + push @buf, $cmd; + } else { + push @buf, $_; + } + } elsif (/^D (\w+)/) { + my $path = $1; + push @buf, $_ if $tree->{$path}; + } elsif ($_ eq "\n") { + if (@oids) { + my $out = join('', @buf); + $out =~ s/^/# /sgm; + warn "purge rewriting\n", $out, "\n"; + clean_purge_buffer(\@oids, \@buf); + $npurge++; + } + $w->print(@buf, "\n") or wfail; + @buf = (); + } elsif ($_ eq "done\n") { + $done = 1; + } elsif (/^mark :(\d+)$/) { + push @buf, $_; + $mark = $1; + } else { + push @buf, $_; + } + } + close $rd or die "close fast-export failed: $?"; + if (@buf) { + $w->print(@buf) or wfail; + } + die 'done\n not seen from fast-export' unless $done; + chomp(my $cmt = $self->get_mark(":$mark")) if $npurge; + $self->{nchg} = 0; # prevent _update_git_info until update-ref: + $self->done; + my @git = ('git', "--git-dir=$git->{git_dir}"); + + run_die([@git, qw(update-ref), $old, $tmp]) if $npurge; + + run_die([@git, qw(update-ref -d), $tmp]); + + return if $npurge == 0; + + run_die([@git, qw(-c gc.reflogExpire=now gc --prune=all)]); + my $err = 0; + foreach my $oid (keys %$purge) { + my @info = $git->check($oid); + if (@info) { + warn "$oid not purged\n"; + $err++; + } + } + _update_git_info($self, 0); + die "Failed to purge $err object(s)\n" if $err; + $cmt; } 1; @@ -487,8 +637,8 @@ version 1.0 =head1 DESCRIPTION An importer and remover for public-inboxes which takes L -messages as input and stores them in a ssoma repository as -documented in L, +messages as input and stores them in a git repository as +documented in L, except it does not allow duplicate Message-IDs. It requires L and L to be installed.