package PublicInbox::V2Writable;
use strict;
use warnings;
-use Fcntl qw(:flock :DEFAULT);
+use base qw(PublicInbox::Lock);
use PublicInbox::SearchIdxPart;
use PublicInbox::SearchIdxSkeleton;
use PublicInbox::MIME;
xap_rw => undef, # PublicInbox::V2SearchIdx
xap_ro => undef,
partitions => $nparts,
+ parallel => 1,
transact_bytes => 0,
+ lock_path => "$dir/inbox.lock",
# limit each repo to 1GB or so
rotate_bytes => int((1024 * 1024 * 1024) / $PACKING_FACTOR),
};
- bless $self, $class
+ bless $self, $class;
}
# returns undef on duplicate or spam
my $nparts = $self->{partitions};
my $part = $num % $nparts;
my $idx = $self->idx_part($part);
- $idx->index_raw($len, $msgref, $num, $oid, $mid0);
+ $idx->index_raw($len, $msgref, $num, $oid, $mid0, $mime);
my $n = $self->{transact_bytes} += $len;
if ($n > (PublicInbox::SearchIdx::BATCH_BYTES * $nparts)) {
$self->checkpoint;
foreach my $m (@$mids) {
# read-only lookup now safe to do after above barrier
my $existing = $self->lookup_content($mime, $m);
- if ($existing) {
- warn "<$m> resent\n";
- return; # easy, don't store duplicates
- }
+ # 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;
}
# very unlikely:
my $hdr = $mime->header_obj;
my $dig = content_digest($mime);
- $$mid0 = $dig->clone->hexdigest . '@localhost';
+ $$mid0 = PublicInbox::Import::digest2mid($dig);
my $num = $self->{skel}->{mm}->mid_insert($$mid0);
unless (defined $num) {
# it's hard to spoof the last Received: header
my @recvd = $hdr->header_raw('Received');
$dig->add("Received: $_") foreach (@recvd);
- $$mid0 = $dig->clone->hexdigest . '@localhost';
+ $$mid0 = PublicInbox::Import::digest2mid($dig);
$num = $self->{skel}->{mm}->mid_insert($$mid0);
# fall back to a random Message-ID and give up determinism:
until (defined($num)) {
$dig->add(rand);
- $$mid0 = $dig->clone->hexdigest . '@localhost';
+ $$mid0 = PublicInbox::Import::digest2mid($dig);
warn "using random Message-ID <$$mid0> as fallback\n";
$num = $self->{skel}->{mm}->mid_insert($$mid0);
}
# frequently activated.
delete $ibx->{$_} foreach (qw(git mm search));
+ $self->lock_acquire;
+
# first time initialization, first we create the skeleton pipe:
my $skel = $self->{skel} = PublicInbox::SearchIdxSkeleton->new($self);
sub done {
my ($self) = @_;
+ my $locked = defined $self->{idx_parts};
my $im = delete $self->{im};
$im->done if $im; # PublicInbox::Import::done
$self->searchidx_checkpoint(0);
+ $self->lock_release if $locked;
}
sub checkpoint {
sub barrier {
my ($self) = @_;
- # For safety, we ensure git checkpoint is complete before because
- # the data in git is still more important than what is in Xapian.
- # Performance may be gained by delaying ->progress call but we
- # lose safety
if (my $im = $self->{im}) {
- $im->checkpoint;
- $im->progress('checkpoint');
+ $im->barrier;
}
my $skel = $self->{skel};
my $parts = $self->{idx_parts};
# Now deal with Xapian
$skel->barrier_init(scalar(@$parts));
# each partition needs to issue a barrier command to skel:
- $_->barrier foreach @$parts;
+ $_->remote_barrier foreach @$parts;
$skel->barrier_wait; # wait for each Xapian partition
my $im = PublicInbox::Import->new($git, undef, undef, $self->{-inbox});
$im->{bytes_added} = int($packed_bytes / $PACKING_FACTOR);
$im->{want_object_info} = 1;
- $im->{ssoma_lock} = 0;
+ $im->{lock_path} = undef;
$im->{path_type} = 'v2';
$self->{im} = $im;
}
+# XXX experimental
+sub diff ($$$) {
+ my ($mid, $cur, $new) = @_;
+ use File::Temp qw(tempfile);
+ use PublicInbox::Spawn qw(spawn);
+
+ my ($ah, $an) = tempfile('email-cur-XXXXXXXX');
+ print $ah $cur->as_string or die "print: $!";
+ close $ah or die "close: $!";
+ my ($bh, $bn) = tempfile('email-new-XXXXXXXX');
+ print $bh $new->as_string or die "print: $!";
+ close $bh or die "close: $!";
+ my $cmd = [ qw(diff -u), $an, $bn ];
+ print STDERR "# MID conflict <$mid>\n";
+ my $pid = spawn($cmd, undef, { 1 => 2 });
+ defined $pid or die "diff failed to spawn $!";
+ waitpid($pid, 0) == $pid or die "diff did not finish";
+ unlink($an, $bn);
+}
+
sub lookup_content {
my ($self, $mime, $mid) = @_;
my $ibx = $self->{-inbox};
$found = $smsg;
return 0; # break out of loop
}
+
+ # XXX DEBUG_DIFF is experimental and may be removed
+ diff($mid, $cur, $mime) if $ENV{DEBUG_DIFF};
+
1; # continue
});
$found;