# Used to write to V2 inboxes (see L<public-inbox-v2-format(5)>).
package PublicInbox::V2Writable;
use strict;
-use warnings;
-use base qw(PublicInbox::Lock);
-use 5.010_001;
+use v5.10.1;
+use parent qw(PublicInbox::Lock);
use PublicInbox::SearchIdxShard;
-use PublicInbox::MIME;
+use PublicInbox::IdxStack;
+use PublicInbox::Eml;
use PublicInbox::Git;
use PublicInbox::Import;
use PublicInbox::MID qw(mids references);
-use PublicInbox::ContentId qw(content_id content_digest);
-use PublicInbox::Inbox;
+use PublicInbox::ContentHash qw(content_hash content_digest);
+use PublicInbox::InboxWritable;
use PublicInbox::OverIdx;
use PublicInbox::Msgmap;
use PublicInbox::Spawn qw(spawn popen_rd);
use IO::Handle; # ->autoflush
use File::Temp qw(tempfile);
+my $OID = qr/[a-f0-9]{40,}/;
# an estimate of the post-packed size to the raw uncompressed size
my $PACKING_FACTOR = 0.4;
# Also, shard count may change while -watch is running
# due to "xcpdb --reshard"
if (-d $xpfx) {
- require PublicInbox::Search;
- PublicInbox::Search::load_xapian();
- my $XapianDatabase = $PublicInbox::Search::X{Database};
+ my $XapianDatabase;
foreach my $shard (<$xpfx/*>) {
-d $shard && $shard =~ m!/[0-9]+\z! or next;
+ $XapianDatabase //= do {
+ require PublicInbox::Search;
+ PublicInbox::Search::load_xapian();
+ $PublicInbox::Search::X{Database};
+ };
eval {
$XapianDatabase->new($shard)->close;
$n++;
im => undef, # PublicInbox::Import
parallel => 1,
transact_bytes => 0,
+ total_bytes => 0,
current_info => '',
xpfx => $xpfx,
over => PublicInbox::OverIdx->new("$xpfx/over.sqlite3", 1),
last_commit => [], # git repo -> commit
};
$self->{shards} = count_shards($self) || nproc_shards($creat);
+ $self->{index_max_size} = $v2ibx->{index_max_size};
bless $self, $class;
}
# public (for now?)
sub init_inbox {
- my ($self, $shards, $skip_epoch) = @_;
+ my ($self, $shards, $skip_epoch, $skip_artnum) = @_;
if (defined $shards) {
$self->{parallel} = 0 if $shards == 0;
$self->{shards} = $shards if $shards > 0;
}
$self->idx_init;
+ $self->{mm}->skip_artnum($skip_artnum) if defined $skip_artnum;
my $epoch_max = -1;
git_dir_latest($self, \$epoch_max);
if (defined $skip_epoch && $epoch_max == -1) {
# returns undef on duplicate or spam
# mimics Import::add and wraps it for v2
sub add {
- my ($self, $mime, $check_cb) = @_;
- $self->{-inbox}->with_umask(sub {
- _add($self, $mime, $check_cb)
- });
+ my ($self, $eml, $check_cb) = @_;
+ $self->{-inbox}->with_umask(\&_add, $self, $eml, $check_cb);
}
# indexes a message, returns true if checkpointing is needed
-sub do_idx ($$$$$$$) {
- my ($self, $msgref, $mime, $len, $num, $oid, $mid0) = @_;
- $self->{over}->add_overview($mime, $len, $num, $oid, $mid0);
- my $idx = idx_shard($self, $num % $self->{shards});
- $idx->index_raw($len, $msgref, $num, $oid, $mid0, $mime);
- my $n = $self->{transact_bytes} += $len;
- $n >= (PublicInbox::SearchIdx::BATCH_BYTES * $self->{shards});
+sub do_idx ($$$$) {
+ my ($self, $msgref, $mime, $smsg) = @_;
+ $smsg->{bytes} = $smsg->{raw_bytes} +
+ PublicInbox::SearchIdx::crlf_adjust($$msgref);
+ $self->{over}->add_overview($mime, $smsg);
+ my $idx = idx_shard($self, $smsg->{num} % $self->{shards});
+ $idx->index_raw($msgref, $mime, $smsg);
+ my $n = $self->{transact_bytes} += $smsg->{raw_bytes};
+ $n >= ($PublicInbox::SearchIdx::BATCH_BYTES * $self->{shards});
}
sub _add {
# spam check:
if ($check_cb) {
- $mime = $check_cb->($mime) or return;
+ $mime = $check_cb->($mime, $self->{-inbox}) or return;
}
# All pipes (> $^F) known to Perl 5.6+ have FD_CLOEXEC set,
my ($num, $mid0) = v2_num_for($self, $mime);
defined $num or return; # duplicate
- defined $mid0 or die "BUG: $mid0 undefined\n";
+ defined $mid0 or die "BUG: \$mid0 undefined\n";
my $im = $self->importer;
- my $cmt = $im->add($mime);
+ my $smsg = bless { mid => $mid0, num => $num }, 'PublicInbox::Smsg';
+ my $cmt = $im->add($mime, undef, $smsg); # sets $smsg->{ds|ts|blob}
$cmt = $im->get_mark($cmt);
$self->{last_commit}->[$self->{epoch_max}] = $cmt;
- my ($oid, $len, $msgref) = @{$im->{last_object}};
- if (do_idx($self, $msgref, $mime, $len, $num, $oid, $mid0)) {
+ my $msgref = delete $smsg->{-raw_email};
+ if (do_idx($self, $msgref, $mime, $smsg)) {
$self->checkpoint;
}
# crap, Message-ID is already known, hope somebody just resent:
foreach my $m (@$mids) {
# read-only lookup now safe to do after above barrier
- my $existing = lookup_content($self, $mime, $m);
# easy, don't store duplicates
# note: do not add more diagnostic info here since
# it gets noisy on public-inbox-watch restarts
- return () if $existing;
+ return () if content_exists($self, $mime, $m);
}
# AltId may pre-populate article numbers (e.g. X-Mail-Count
$self->{idx_shards}->[$shard_i];
}
+sub _idx_init { # with_umask callback
+ my ($self, $opt) = @_;
+ $self->lock_acquire unless $opt && $opt->{-skip_lock};
+ $self->{over}->create;
+
+ # xcpdb can change shard count while -watch is idle
+ my $nshards = count_shards($self);
+ $self->{shards} = $nshards if $nshards && $nshards != $self->{shards};
+
+ # need to create all shards before initializing msgmap FD
+ # idx_shards must be visible to all forked processes
+ my $max = $self->{shards} - 1;
+ my $idx = $self->{idx_shards} = [];
+ push @$idx, PublicInbox::SearchIdxShard->new($self, $_) for (0..$max);
+
+ # Now that all subprocesses are up, we can open the FDs
+ # for SQLite:
+ my $mm = $self->{mm} = PublicInbox::Msgmap->new_file(
+ "$self->{-inbox}->{inboxdir}/msgmap.sqlite3", 1);
+ $mm->{dbh}->begin_work;
+}
+
# idempotent
sub idx_init {
my ($self, $opt) = @_;
# do not leak read-only FDs to child processes, we only have these
# FDs for duplicate detection so they should not be
# frequently activated.
+ # delete @$ibx{qw(git mm search)};
delete $ibx->{$_} foreach (qw(git mm search));
- my $indexlevel = $ibx->{indexlevel};
- if ($indexlevel && $indexlevel eq 'basic') {
- $self->{parallel} = 0;
- }
-
+ $self->{parallel} = 0 if ($ibx->{indexlevel}//'') eq 'basic';
if ($self->{parallel}) {
pipe(my ($r, $w)) or die "pipe failed: $!";
# pipe for barrier notifications doesn't need to be big,
$w->autoflush(1);
}
- my $over = $self->{over};
$ibx->umask_prepare;
- $ibx->with_umask(sub {
- $self->lock_acquire unless ($opt && $opt->{-skip_lock});
- $over->create;
-
- # xcpdb can change shard count while -watch is idle
- my $nshards = count_shards($self);
- if ($nshards && $nshards != $self->{shards}) {
- $self->{shards} = $nshards;
- }
-
- # need to create all shards before initializing msgmap FD
- my $max = $self->{shards} - 1;
-
- # idx_shards must be visible to all forked processes
- my $idx = $self->{idx_shards} = [];
- for my $i (0..$max) {
- push @$idx, PublicInbox::SearchIdxShard->new($self, $i);
- }
-
- # Now that all subprocesses are up, we can open the FDs
- # for SQLite:
- my $mm = $self->{mm} = PublicInbox::Msgmap->new_file(
- "$self->{-inbox}->{inboxdir}/msgmap.sqlite3", 1);
- $mm->{dbh}->begin_work;
- });
+ $ibx->with_umask(\&_idx_init, $self, $opt);
}
# returns an array mapping [ epoch => latest_commit ]
$rewrites;
}
-sub content_ids ($) {
+sub content_hashes ($) {
my ($mime) = @_;
- my @cids = ( content_id($mime) );
+ my @chashes = ( content_hash($mime) );
+ # We still support Email::MIME, here, and
# Email::MIME->as_string doesn't always round-trip, so we may
- # use a second content_id
- my $rt = content_id(PublicInbox::MIME->new(\($mime->as_string)));
- push @cids, $rt if $cids[0] ne $rt;
- \@cids;
+ # use a second content_hash
+ my $rt = content_hash(PublicInbox::Eml->new(\($mime->as_string)));
+ push @chashes, $rt if $chashes[0] ne $rt;
+ \@chashes;
}
sub content_matches ($$) {
- my ($cids, $existing) = @_;
- my $cid = content_id($existing);
- foreach (@$cids) {
- return 1 if $_ eq $cid
+ my ($chashes, $existing) = @_;
+ my $chash = content_hash($existing);
+ foreach (@$chashes) {
+ return 1 if $_ eq $chash
}
0
}
# used for removing or replacing (purging)
sub rewrite_internal ($$;$$$) {
- my ($self, $old_mime, $cmt_msg, $new_mime, $sref) = @_;
+ my ($self, $old_eml, $cmt_msg, $new_eml, $sref) = @_;
$self->idx_init;
my ($im, $need_reindex, $replace_map);
if ($sref) {
$replace_map = {}; # oid => sref
- $need_reindex = [] if $new_mime;
+ $need_reindex = [] if $new_eml;
} else {
$im = $self->importer;
}
my $over = $self->{over};
- my $cids = content_ids($old_mime);
- my $removed;
- my $mids = mids($old_mime->header_obj);
+ my $chashes = content_hashes($old_eml);
+ my $removed = [];
+ my $mids = mids($old_eml->header_obj);
# We avoid introducing new blobs into git since the raw content
# can be slightly different, so we do not need the user-supplied
- # message now that we have the mids and content_id
- $old_mime = undef;
+ # message now that we have the mids and content_hash
+ $old_eml = undef;
my $mark;
foreach my $mid (@$mids) {
- my %gone; # num => [ smsg, raw ]
+ my %gone; # num => [ smsg, $mime, raw ]
my ($id, $prev);
while (my $smsg = $over->next_by_mid($mid, \$id, \$prev)) {
my $msg = get_blob($self, $smsg);
next; # continue
}
my $orig = $$msg;
- my $cur = PublicInbox::MIME->new($msg);
- if (content_matches($cids, $cur)) {
- $smsg->{mime} = $cur;
- $gone{$smsg->{num}} = [ $smsg, \$orig ];
+ my $cur = PublicInbox::Eml->new($msg);
+ if (content_matches($chashes, $cur)) {
+ $gone{$smsg->{num}} = [ $smsg, $cur, \$orig ];
}
}
my $n = scalar keys %gone;
join(',', sort keys %gone), "\n";
}
foreach my $num (keys %gone) {
- my ($smsg, $orig) = @{$gone{$num}};
+ my ($smsg, $mime, $orig) = @{$gone{$num}};
# $removed should only be set once assuming
# no bugs in our deduplication code:
- $removed = $smsg;
+ $removed = [ undef, $mime, $smsg ];
my $oid = $smsg->{blob};
if ($replace_map) {
$replace_map->{$oid} = $sref;
} else {
($mark, undef) = $im->remove($orig, $cmt_msg);
+ $removed->[0] = $mark;
}
$orig = undef;
if ($need_reindex) { # ->replace
$self->{last_commit}->[$self->{epoch_max}] = $cmt;
}
if ($replace_map && scalar keys %$replace_map) {
- my $rewrites = _replace_oids($self, $new_mime, $replace_map);
+ my $rewrites = _replace_oids($self, $new_eml, $replace_map);
return { rewrites => $rewrites, need_reindex => $need_reindex };
}
- $removed;
+ defined($mark) ? $removed : undef;
}
-# public
+# public (see PublicInbox::Import->remove), but note the 3rd element
+# (retval[2]) is not part of the stable API shared with Import->remove
sub remove {
- my ($self, $mime, $cmt_msg) = @_;
- $self->{-inbox}->with_umask(sub {
- rewrite_internal($self, $mime, $cmt_msg);
- });
+ my ($self, $eml, $cmt_msg) = @_;
+ my $r = $self->{-inbox}->with_umask(\&rewrite_internal,
+ $self, $eml, $cmt_msg);
+ defined($r) && defined($r->[0]) ? @$r: undef;
}
sub _replace ($$;$$) {
- my ($self, $old_mime, $new_mime, $sref) = @_;
- my $rewritten = $self->{-inbox}->with_umask(sub {
- rewrite_internal($self, $old_mime, undef, $new_mime, $sref);
- }) or return;
+ my ($self, $old_eml, $new_eml, $sref) = @_;
+ my $arg = [ $self, $old_eml, undef, $new_eml, $sref ];
+ my $rewritten = $self->{-inbox}->with_umask(\&rewrite_internal,
+ $self, $old_eml, undef, $new_eml, $sref) or return;
my $rewrites = $rewritten->{rewrites};
# ->done is called if there are rewrites since we gc+prune from git
sub git_hash_raw ($$) {
my ($self, $raw) = @_;
# grab the expected OID we have to reindex:
- open my $tmp_fh, '+>', undef or die "failed to open tmp: $!";
- $tmp_fh->autoflush(1);
- print $tmp_fh $$raw or die "print \$tmp_fh: $!";
- sysseek($tmp_fh, 0, 0) or die "seek failed: $!";
-
+ pipe(my($in, $w)) or die "pipe: $!";
my $git_dir = $self->{-inbox}->git->{git_dir};
my $cmd = ['git', "--git-dir=$git_dir", qw(hash-object --stdin)];
- my $r = popen_rd($cmd, undef, { 0 => $tmp_fh });
+ my $r = popen_rd($cmd, undef, { 0 => $in });
+ print $w $$raw or die "print \$w: $!";
+ close $w or die "close \$w: $!";
local $/ = "\n";
chomp(my $oid = <$r>);
close $r or die "git hash-object failed: $?";
- $oid =~ /\A[a-f0-9]{40}\z/ or die "OID not expected: $oid";
+ $oid =~ /\A$OID\z/ or die "OID not expected: $oid";
$oid;
}
}
# make sure we really got the OID:
- my ($oid, $type, $len) = $self->{-inbox}->git->check($expect_oid);
- $oid eq $expect_oid or die "BUG: $expect_oid not found after replace";
+ my ($blob, $type, $bytes) = $self->{-inbox}->git->check($expect_oid);
+ $blob eq $expect_oid or die "BUG: $expect_oid not found after replace";
# don't leak FDs to Xapian:
$self->{-inbox}->git->cleanup;
# reindex modified messages:
for my $smsg (@$need_reindex) {
- my $num = $smsg->{num};
- my $mid0 = $smsg->{mid};
- do_idx($self, \$raw, $new_mime, $len, $num, $oid, $mid0);
+ my $new_smsg = bless {
+ blob => $blob,
+ raw_bytes => $bytes,
+ num => $smsg->{num},
+ mid => $smsg->{mid},
+ }, 'PublicInbox::Smsg';
+ my $sync = { autime => $smsg->{ds}, cotime => $smsg->{ts} };
+ $new_smsg->populate($new_mime, $sync);
+ do_idx($self, \$raw, $new_mime, $new_smsg);
}
$rewritten->{rewrites};
}
my $bnote = $self->{bnote} or return;
my $r = $bnote->[0];
while (scalar keys %$barrier) {
- defined(my $l = $r->getline) or die "EOF on barrier_wait: $!";
+ defined(my $l = readline($r)) or die "EOF on barrier_wait: $!";
$l =~ /\Abarrier (\d+)/ or die "bad line on barrier_wait: $l";
delete $barrier->{$1} or die "bad shard[$1] on barrier wait";
}
$dbh->begin_work;
}
+ $self->{total_bytes} += $self->{transact_bytes};
$self->{transact_bytes} = 0;
}
}
$self->{over}->disconnect;
delete $self->{bnote};
- $self->{transact_bytes} = 0;
- $self->lock_release if $shards;
+ my $nbytes = $self->{total_bytes};
+ $self->{total_bytes} = 0;
+ $self->lock_release(!!$nbytes) if $shards;
$self->{-inbox}->git->cleanup;
}
my $pfx = "$self->{-inbox}->{inboxdir}/git";
my $all = "$self->{-inbox}->{inboxdir}/all.git";
-
- unless (-d $all) {
- PublicInbox::Import::init_bare($all);
- }
+ PublicInbox::Import::init_bare($all) unless -d $all;
my $info_dir = "$all/objects/info";
my $alt = "$info_dir/alternates";
my (%alt, $new);
$mode = (stat($fh))[2] & 07777;
# we assign a sort score to every alternate and favor
- # the newest (highest numbered) one when we
+ # the newest (highest numbered) one because loose objects
+ # require scanning epochs and only the latest epoch is
+ # expected to see loose objects
my $score;
my $other = 0; # in case admin adds non-epoch repos
%alt = map {;
sub git_init {
my ($self, $epoch) = @_;
my $git_dir = "$self->{-inbox}->{inboxdir}/git/$epoch.git";
- my @cmd = (qw(git init --bare -q), $git_dir);
- PublicInbox::Import::run_die(\@cmd);
- @cmd = (qw/git config/, "--file=$git_dir/config",
+ PublicInbox::Import::init_bare($git_dir);
+ my @cmd = (qw/git config/, "--file=$git_dir/config",
'include.path', '../../all.git/config');
PublicInbox::Import::run_die(\@cmd);
fill_alternates($self, $epoch);
my ($self, $git, $packed_bytes, $tmp) = @_;
my $im = PublicInbox::Import->new($git, undef, undef, $self->{-inbox});
$im->{bytes_added} = int($packed_bytes / $PACKING_FACTOR);
- $im->{want_object_info} = 1;
$im->{lock_path} = undef;
$im->{path_type} = 'v2';
$self->{im} = $im unless $tmp;
$ibx->msg_by_smsg($smsg);
}
-sub lookup_content ($$$) {
+sub content_exists ($$$) {
my ($self, $mime, $mid) = @_;
my $over = $self->{over};
- my $cids = content_ids($mime);
+ my $chashes = content_hashes($mime);
my ($id, $prev);
while (my $smsg = $over->next_by_mid($mid, \$id, \$prev)) {
my $msg = get_blob($self, $smsg);
warn "broken smsg for $mid\n";
next;
}
- my $cur = PublicInbox::MIME->new($msg);
- if (content_matches($cids, $cur)) {
- $smsg->{mime} = $cur;
- return $smsg;
- }
-
+ my $cur = PublicInbox::Eml->new($msg);
+ return 1 if content_matches($chashes, $cur);
# XXX DEBUG_DIFF is experimental and may be removed
diff($mid, $cur, $mime) if $ENV{DEBUG_DIFF};
$self->{bnote}->[1];
}
-sub mark_deleted ($$$$) {
- my ($self, $sync, $git, $oid) = @_;
- my $msgref = $git->cat_file($oid);
- my $mime = PublicInbox::MIME->new($$msgref);
- my $mids = mids($mime->header_obj);
- my $cid = content_id($mime);
- foreach my $mid (@$mids) {
- $sync->{D}->{"$mid\0$cid"} = $oid;
- }
-}
-
sub reindex_checkpoint ($$$) {
my ($self, $sync, $git) = @_;
$sync->{mm_tmp}->atfork_parent;
}
-# only for a few odd messages with multiple Message-IDs
-sub reindex_oid_m ($$$$;$) {
- my ($self, $sync, $git, $oid, $regen_num) = @_;
- $self->{current_info} = "multi_mid $oid";
- my ($num, $mid0, $len);
- my $msgref = $git->cat_file($oid, \$len);
- my $mime = PublicInbox::MIME->new($$msgref);
- my $mids = mids($mime->header_obj);
- my $cid = content_id($mime);
- die "BUG: reindex_oid_m called for <=1 mids" if scalar(@$mids) <= 1;
-
- for my $mid (reverse @$mids) {
- delete($sync->{D}->{"$mid\0$cid"}) and
- die "BUG: reindex_oid should handle <$mid> delete";
- }
- my $over = $self->{over};
- for my $mid (reverse @$mids) {
- ($num, $mid0) = $over->num_mid0_for_oid($oid, $mid);
- next unless defined $num;
- if (defined($regen_num) && $regen_num != $num) {
- die "BUG: regen(#$regen_num) != over(#$num)";
- }
- }
- unless (defined($num)) {
- for my $mid (reverse @$mids) {
- # is this a number we got before?
- my $n = $sync->{mm_tmp}->num_for($mid);
- next unless defined $n;
- next if defined($regen_num) && $regen_num != $n;
- ($num, $mid0) = ($n, $mid);
- last;
- }
- }
- if (defined($num)) {
- $sync->{mm_tmp}->num_delete($num);
- } elsif (defined $regen_num) {
- $num = $regen_num;
- for my $mid (reverse @$mids) {
- $self->{mm}->mid_set($num, $mid) == 1 or next;
- $mid0 = $mid;
- last;
- }
- unless (defined $mid0) {
- warn "E: cannot regen #$num\n";
- return;
- }
- } else { # fixup bugs in old mirrors on reindex
- for my $mid (reverse @$mids) {
- $num = $self->{mm}->mid_insert($mid);
- next unless defined $num;
- $mid0 = $mid;
- last;
- }
- if (defined $mid0) {
- if ($sync->{reindex}) {
- warn "reindex added #$num <$mid0>\n";
- }
- } else {
- warn "E: cannot find article #\n";
- return;
- }
- }
- $sync->{nr}++;
- if (do_idx($self, $msgref, $mime, $len, $num, $oid, $mid0)) {
- reindex_checkpoint($self, $sync, $git);
- }
-}
-
-sub check_unindexed ($$$) {
- my ($self, $num, $mid0) = @_;
- my $unindexed = $self->{unindexed} // {};
- my $n = delete($unindexed->{$mid0});
- defined $n or return;
- if ($n != $num) {
- die "BUG: unindexed $n != $num <$mid0>\n";
- } else {
- $self->{mm}->mid_set($num, $mid0);
- }
-}
-
-# reuse Msgmap to store num => oid mapping (rather than num => mid)
-sub multi_mid_q_new () {
- my ($fh, $fn) = tempfile('multi_mid-XXXXXXX', EXLOCK => 0, TMPDIR => 1);
- my $multi_mid = PublicInbox::Msgmap->new_file($fn, 1);
- $multi_mid->{dbh}->do('PRAGMA synchronous = OFF');
- # for Msgmap->DESTROY:
- $multi_mid->{tmp_name} = $fn;
- $multi_mid->{pid} = $$;
- close $fh or die "failed to close $fn: $!";
- $multi_mid
-}
-
-sub multi_mid_q_push ($$) {
- my ($sync, $oid) = @_;
- my $multi_mid = $sync->{multi_mid} //= multi_mid_q_new();
- if ($sync->{reindex}) { # no regen on reindex
- $multi_mid->mid_insert($oid);
- } else {
- my $num = $sync->{regen}--;
- die "BUG: ran out of article numbers" if $num <= 0;
- $multi_mid->mid_set($num, $oid);
- }
-}
-
sub reindex_oid ($$$$) {
my ($self, $sync, $git, $oid) = @_;
+ return if PublicInbox::SearchIdx::too_big($self, $git, $oid);
my ($num, $mid0, $len);
my $msgref = $git->cat_file($oid, \$len);
return if $len == 0; # purged
- my $mime = PublicInbox::MIME->new($$msgref);
+ my $mime = PublicInbox::Eml->new($$msgref);
my $mids = mids($mime->header_obj);
- my $cid = content_id($mime);
+ my $chash = content_hash($mime);
if (scalar(@$mids) == 0) {
warn "E: $oid has no Message-ID, skipping\n";
return;
- } elsif (scalar(@$mids) == 1) {
- my $mid = $mids->[0];
-
- # was the file previously marked as deleted?, skip if so
- if (delete($sync->{D}->{"$mid\0$cid"})) {
- if (!$sync->{reindex}) {
- $num = $sync->{regen}--;
- $self->{mm}->num_highwater($num);
- }
- return;
- }
+ }
- # is this a number we got before?
- $num = $sync->{mm_tmp}->num_for($mid);
+ # {unindexed} is unlikely
+ if ((my $unindexed = $self->{unindexed}) && scalar(@$mids) == 1) {
+ $num = delete($unindexed->{$mids->[0]});
if (defined $num) {
- $mid0 = $mid;
- check_unindexed($self, $num, $mid0);
- } else {
- $num = $sync->{regen}--;
- die "BUG: ran out of article numbers" if $num <= 0;
- if ($self->{mm}->mid_set($num, $mid) != 1) {
- warn "E: unable to assign $num => <$mid>\n";
- return;
- }
- $mid0 = $mid;
+ $mid0 = $mids->[0];
+ $self->{mm}->mid_set($num, $mid0);
+ delete($self->{unindexed}) if !keys(%$unindexed);
+ }
+ }
+ if (!defined($num)) { # reuse if reindexing (or duplicates)
+ my $over = $self->{over};
+ for my $mid (@$mids) {
+ ($num, $mid0) = $over->num_mid0_for_oid($oid, $mid);
+ last if defined $num;
}
- } else { # multiple MIDs are a weird case:
- my $del = 0;
- for (@$mids) {
- $del += delete($sync->{D}->{"$_\0$cid"}) // 0;
+ }
+ $mid0 //= do { # is this a number we got before?
+ $num = $sync->{mm_tmp}->num_for($mids->[0]);
+ defined($num) ? $mids->[0] : undef;
+ };
+ if (!defined($num)) {
+ for (my $i = $#$mids; $i >= 1; $i--) {
+ $num = $sync->{mm_tmp}->num_for($mids->[$i]);
+ if (defined($num)) {
+ $mid0 = $mids->[$i];
+ last;
+ }
}
- if ($del) {
- unindex_oid_remote($self, $oid, $_) for @$mids;
- # do not delete from {mm_tmp}, since another
- # single-MID message may use it.
- } else { # handle them at the end:
- multi_mid_q_push($sync, $oid);
+ }
+ if (defined($num)) {
+ $sync->{mm_tmp}->num_delete($num);
+ } else { # never seen
+ $num = $self->{mm}->mid_insert($mids->[0]);
+ if (defined($num)) {
+ $mid0 = $mids->[0];
+ } else { # rare, try the rest of them, backwards
+ for (my $i = $#$mids; $i >= 1; $i--) {
+ $num = $self->{mm}->mid_insert($mids->[$i]);
+ if (defined($num)) {
+ $mid0 = $mids->[$i];
+ last;
+ }
+ }
}
+ }
+ if (!defined($num)) {
+ warn "E: $oid <", join('> <', @$mids), "> is a duplicate\n";
return;
}
- $sync->{mm_tmp}->mid_delete($mid0) or
- die "failed to delete <$mid0> for article #$num\n";
$sync->{nr}++;
- if (do_idx($self, $msgref, $mime, $len, $num, $oid, $mid0)) {
+ my $smsg = bless {
+ raw_bytes => $len,
+ num => $num,
+ blob => $oid,
+ mid => $mid0,
+ }, 'PublicInbox::Smsg';
+ $smsg->populate($mime, $sync);
+ if (do_idx($self, $msgref, $mime, $smsg)) {
reindex_checkpoint($self, $sync, $git);
}
}
$range;
}
+sub prepare_range_stack {
+ my ($git, $sync, $range) = @_;
+ # Don't bump num_highwater on --reindex by using {D}.
+ # We intentionally do NOT use {D} in the non-reindex case because
+ # we want NNTP article number gaps from unindexed messages to
+ # show up in mirrors, too.
+ my $D = $sync->{D} //= $sync->{reindex} ? {} : undef; # OID_BIN => NR
+
+ my $fh = $git->popen(qw(log --raw -r --pretty=tformat:%at-%ct-%H
+ --no-notes --no-color --no-renames --no-abbrev),
+ $range);
+ my ($at, $ct, $stk);
+ while (<$fh>) {
+ if (/\A([0-9]+)-([0-9]+)-($OID)$/o) {
+ ($at, $ct) = ($1 + 0, $2 + 0);
+ $stk //= PublicInbox::IdxStack->new($3);
+ } elsif (/\A:\d{6} 100644 $OID ($OID) [AM]\td$/o) {
+ my $oid = $1;
+ if ($D) { # reindex case
+ $D->{pack('H*', $oid)}++;
+ } else { # non-reindex case:
+ $stk->push_rec('d', $at, $ct, $oid);
+ }
+ } elsif (/\A:\d{6} 100644 $OID ($OID) [AM]\tm$/o) {
+ my $oid = $1;
+ if ($D) {
+ my $oid_bin = pack('H*', $oid);
+ my $nr = --$D->{$oid_bin};
+ delete($D->{$oid_bin}) if $nr <= 0;
+
+ # nr < 0 (-1) means it never existed
+ $stk->push_rec('m', $at, $ct, $oid) if $nr < 0;
+ } else {
+ $stk->push_rec('m', $at, $ct, $oid);
+ }
+ }
+ }
+ close $fh or die "git log failed: \$?=$?";
+ $stk ? $stk->read_prepare : undef;
+}
+
sub sync_prepare ($$$) {
my ($self, $sync, $epoch_max) = @_;
my $pr = $sync->{-opt}->{-progress};
next if $?; # new repo
my $range = log_range($self, $sync, $git, $i, $tip) or next;
- $sync->{ranges}->[$i] = $range;
-
# can't use 'rev-list --count' if we use --diff-filter
$pr->("$i.git counting $range ... ") if $pr;
- my $n = 0;
- my $fh = $git->popen(qw(log --pretty=tformat:%H
- --no-notes --no-color --no-renames
- --diff-filter=AM), $range, '--', 'm');
- ++$n while <$fh>;
- close $fh or die "git log failed: \$?=$?";
- $pr->("$n\n") if $pr;
- $regen_max += $n;
+ my $stk = prepare_range_stack($git, $sync, $range);
+ my $nr = $stk ? $stk->num_records : 0;
+ $pr->("$nr\n") if $pr;
+ $sync->{stacks}->[$i] = $stk if $stk;
+ $regen_max += $nr;
+ }
+
+ # XXX this should not happen unless somebody bypasses checks in
+ # our code and blindly injects "d" file history into git repos
+ if (my @leftovers = keys %{delete($sync->{D}) // {}}) {
+ warn('W: unindexing '.scalar(@leftovers)." leftovers\n");
+ my $git = $self->{-inbox}->git;
+ for my $oid (@leftovers) {
+ $oid = unpack('H*', $oid);
+ $self->{current_info} = "leftover $oid";
+ unindex_oid($self, $git, $oid);
+ }
+ $git->cleanup;
}
-
return 0 if (!$regen_max && !keys(%{$self->{unindex_range}}));
# reindex should NOT see new commits anymore, if we do,
sub unindex_oid_remote ($$$) {
my ($self, $oid, $mid) = @_;
- $_->remote_remove($oid, $mid) foreach @{$self->{idx_shards}};
- $self->{over}->remove_oid($oid, $mid);
+ my @removed = $self->{over}->remove_oid($oid, $mid);
+ for my $num (@removed) {
+ my $idx = idx_shard($self, $num % $self->{shards});
+ $idx->remote_remove($oid, $num);
+ }
}
sub unindex_oid ($$$;$) {
my ($self, $git, $oid, $unindexed) = @_;
my $mm = $self->{mm};
my $msgref = $git->cat_file($oid);
- my $mime = PublicInbox::MIME->new($msgref);
+ my $mime = PublicInbox::Eml->new($msgref);
my $mids = mids($mime->header_obj);
$mime = $msgref = undef;
my $over = $self->{over};
my ($id, $prev);
while (my $smsg = $over->next_by_mid($mid, \$id, \$prev)) {
$gone{$smsg->{num}} = 1 if $oid eq $smsg->{blob};
- 1; # continue
}
- my $n = scalar keys %gone;
- next unless $n;
+ my $n = scalar(keys(%gone)) or next;
if ($n > 1) {
warn "BUG: multiple articles linked to $oid\n",
join(',',sort keys %gone), "\n";
}
}
-my $x40 = qr/[a-f0-9]{40}/;
sub unindex ($$$$) {
my ($self, $sync, $git, $unindex_range) = @_;
my $unindexed = $self->{unindexed} ||= {}; # $mid0 => $num
--no-notes --no-color --no-abbrev --no-renames);
my $fh = $self->{reindex_pipe} = $git->popen(@cmd, $unindex_range);
while (<$fh>) {
- /\A:\d{6} 100644 $x40 ($x40) [AM]\tm$/o or next;
+ /\A:\d{6} 100644 $OID ($OID) [AM]\tm$/o or next;
unindex_oid($self, $git, $1, $unindexed);
}
delete $self->{reindex_pipe};
my $git_dir = git_dir_n($self, $i);
die 'BUG: already reindexing!' if $self->{reindex_pipe};
-d $git_dir or return; # missing epochs are fine
- fill_alternates($self, $i);
my $git = PublicInbox::Git->new($git_dir);
if (my $unindex_range = delete $sync->{unindex_range}->{$i}) {
unindex($self, $sync, $git, $unindex_range);
}
- defined(my $range = $sync->{ranges}->[$i]) or return;
- if (my $pr = $sync->{-opt}->{-progress}) {
- $pr->("$i.git indexing $range\n");
- }
-
- my @cmd = qw(log --raw -r --pretty=tformat:%H
- --no-notes --no-color --no-abbrev --no-renames);
- my $fh = $self->{reindex_pipe} = $git->popen(@cmd, $range);
- my $cmt;
- while (<$fh>) {
- chomp;
- $self->{current_info} = "$i.git $_";
- if (/\A$x40$/o && !defined($cmt)) {
- $cmt = $_;
- } elsif (/\A:\d{6} 100644 $x40 ($x40) [AM]\tm$/o) {
- reindex_oid($self, $sync, $git, $1);
- } elsif (/\A:\d{6} 100644 $x40 ($x40) [AM]\td$/o) {
- mark_deleted($self, $sync, $git, $1);
+ defined(my $stk = $sync->{stacks}->[$i]) or return;
+ $sync->{stacks}->[$i] = undef;
+ while (my ($f, $at, $ct, $oid) = $stk->pop_rec) {
+ $self->{current_info} = "$i.git $oid";
+ if ($f eq 'm') {
+ $sync->{autime} = $at;
+ $sync->{cotime} = $ct;
+ reindex_oid($self, $sync, $git, $oid);
+ } elsif ($f eq 'd') {
+ unindex_oid($self, $git, $oid);
}
}
- close $fh or die "git log failed: \$?=$?";
- delete $self->{reindex_pipe};
- update_last_commit($self, $git, $i, $cmt) if defined $cmt;
+ delete @$sync{qw(autime cotime)};
+ update_last_commit($self, $git, $i, $stk->{latest_cmt});
}
# public, called by public-inbox-index
my $latest = git_dir_latest($self, \$epoch_max);
return unless defined $latest;
$self->idx_init($opt); # acquire lock
+ fill_alternates($self, $epoch_max);
+ $self->{over}->rethread_prepare($opt);
my $sync = {
- D => {}, # "$mid\0$cid" => $oid
unindex_range => {}, # EPOCH => oid_old..oid_new
reindex => $opt->{reindex},
-opt => $opt
};
$sync->{ranges} = sync_ranges($self, $sync, $epoch_max);
- $sync->{regen} = sync_prepare($self, $sync, $epoch_max);
-
- if ($sync->{regen}) {
+ if (sync_prepare($self, $sync, $epoch_max)) {
# tmp_clone seems to fail if inside a transaction, so
# we rollback here (because we opened {mm} for reading)
# Note: we do NOT rely on DBI transactions for atomicity;
$sync->{mm_tmp} = $self->{mm}->tmp_clone;
}
- # work backwards through history
- for (my $i = $epoch_max; $i >= 0; $i--) {
- index_epoch($self, $sync, $i);
- }
-
- # unindex is required for leftovers if "deletes" affect messages
- # in a previous fetch+index window:
- my $git;
- if (my @leftovers = values %{delete $sync->{D}}) {
- $git = $self->{-inbox}->git;
- for my $oid (@leftovers) {
- $self->{current_info} = "leftover $oid";
- unindex_oid($self, $git, $oid);
- }
- }
- if (my $multi_mid = delete $sync->{multi_mid}) {
- $git //= $self->{-inbox}->git;
- my ($min, $max) = $multi_mid->minmax;
- if ($sync->{reindex}) {
- # we may need to create new Message-IDs if mirrors
- # were initially indexed with old versions
- for (my $i = $max; $i >= $min; $i--) {
- my $oid = $multi_mid->mid_for($i);
- next unless defined $oid;
- reindex_oid_m($self, $sync, $git, $oid);
- }
- } else { # regen on initial index
- for my $num ($min..$max) {
- my $oid = $multi_mid->mid_for($num);
- next unless defined $oid;
- reindex_oid_m($self, $sync, $git, $oid, $num);
- }
- }
- }
- $git->cleanup if $git;
+ # work forwards through history
+ index_epoch($self, $sync, $_) for (0..$epoch_max);
$self->done;
if (my $nr = $sync->{nr}) {
my $pr = $sync->{-opt}->{-progress};
$pr->('all.git '.sprintf($sync->{-regen_fmt}, $nr)) if $pr;
}
+ $self->{over}->rethread_done($opt);
# reindex does not pick up new changes, so we rerun w/o it:
if ($opt->{reindex}) {
my %again = %$opt;
$sync = undef;
- delete @again{qw(reindex -skip_lock)};
+ delete @again{qw(rethread reindex -skip_lock)};
index_sync($self, \%again);
}
}