about summary refs log tree commit
diff options
context:
space:
mode:
authorEric Wong <e@80x24.org>2019-10-22 07:45:38 +0000
committerEric Wong <e@80x24.org>2019-10-23 18:13:30 +0000
commitf9fc5cf2e8cefda4a56d937ed217e47689fd7e49 (patch)
tree669cde20959431be6cb54996c3b43ebe38cc7e23
parenta4b2bb62a0dde5054c5d2c0a7ea3fb3ff1806beb (diff)
parent1d6e0dbc3bad19acbfd90a9841a334dcaa0e4641 (diff)
downloadpublic-inbox-f9fc5cf2e8cefda4a56d937ed217e47689fd7e49.tar.gz
* regen:
  v2writable: use msgmap as multi_mid queue
  v2writable: move git->cleanup to the correct place
  v2writable: reindex handles 3-headered monsters
  v2writable: improve "num_for" API and disambiguate
  v2writable: set unindexed article number
-rw-r--r--TODO3
-rw-r--r--lib/PublicInbox/OverIdx.pm14
-rw-r--r--lib/PublicInbox/V2Writable.pm307
-rw-r--r--t/v2reindex.t66
4 files changed, 300 insertions, 90 deletions
diff --git a/TODO b/TODO
index 7eecde08..f9122a5d 100644
--- a/TODO
+++ b/TODO
@@ -136,3 +136,6 @@ all need to be considered for everything we introduce)
   for coderepos
 
 * configurable diff output for solver-generated blobs
+
+* fix search for messages with multiple Subject:/To:/From:/Date:
+  headers (some wacky examples out there...)
diff --git a/lib/PublicInbox/OverIdx.pm b/lib/PublicInbox/OverIdx.pm
index e8df01c4..01ca6f11 100644
--- a/lib/PublicInbox/OverIdx.pm
+++ b/lib/PublicInbox/OverIdx.pm
@@ -337,6 +337,20 @@ sub remove_oid {
         $nr;
 }
 
+sub num_mid0_for_oid {
+        my ($self, $oid, $mid) = @_;
+        my ($num, $mid0);
+        $self->begin_lazy;
+        each_by_mid($self, $mid, ['ddd'], sub {
+                my ($smsg) = @_;
+                my $blob = $smsg->{blob};
+                return 1 if (!defined($blob) || $blob ne $oid); # continue;
+                ($num, $mid0) = ($smsg->{num}, $smsg->{mid});
+                0; # done
+        });
+        ($num, $mid0);
+}
+
 sub create_tables {
         my ($dbh) = @_;
 
diff --git a/lib/PublicInbox/V2Writable.pm b/lib/PublicInbox/V2Writable.pm
index 6a88f62a..ad2e8e62 100644
--- a/lib/PublicInbox/V2Writable.pm
+++ b/lib/PublicInbox/V2Writable.pm
@@ -19,6 +19,7 @@ use PublicInbox::Msgmap;
 use PublicInbox::Spawn qw(spawn);
 use PublicInbox::SearchIdx;
 use IO::Handle;
+use File::Temp qw(tempfile);
 
 # an estimate of the post-packed size to the raw uncompressed size
 my $PACKING_FACTOR = 0.4;
@@ -155,8 +156,7 @@ sub _add {
         # leaking FDs to it...
         $self->idx_init;
 
-        my $mid0;
-        my $num = num_for($self, $mime, \$mid0);
+        my ($num, $mid0) = v2_num_for($self, $mime);
         defined $num or return; # duplicate
         defined $mid0 or die "BUG: $mid0 undefined\n";
         my $im = $self->importer;
@@ -172,16 +172,15 @@ sub _add {
         $cmt;
 }
 
-sub num_for {
-        my ($self, $mime, $mid0) = @_;
+sub v2_num_for {
+        my ($self, $mime) = @_;
         my $mids = mids($mime->header_obj);
         if (@$mids) {
                 my $mid = $mids->[0];
                 my $num = $self->{mm}->mid_insert($mid);
                 if (defined $num) { # common case
-                        $$mid0 = $mid;
-                        return $num;
-                };
+                        return ($num, $mid);
+                }
 
                 # crap, Message-ID is already known, hope somebody just resent:
                 foreach my $m (@$mids) {
@@ -190,7 +189,7 @@ sub num_for {
                         # 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 $existing;
                 }
 
                 # AltId may pre-populate article numbers (e.g. X-Mail-Count
@@ -201,8 +200,7 @@ sub num_for {
                         my $num = $self->{mm}->num_for($mid);
 
                         if (defined $num && !$self->{over}->get_art($num)) {
-                                $$mid0 = $mid;
-                                return $num;
+                                return ($num, $mid);
                         }
                 }
 
@@ -215,39 +213,38 @@ sub num_for {
                         $num = $self->{mm}->mid_insert($m);
                         if (defined $num) {
                                 warn "alternative <$m> for <$mid> found\n";
-                                $$mid0 = $m;
-                                return $num;
+                                return ($num, $m);
                         }
                 }
         }
         # none of the existing Message-IDs are good, generate a new one:
-        num_for_harder($self, $mime, $mid0);
+        v2_num_for_harder($self, $mime);
 }
 
-sub num_for_harder {
-        my ($self, $mime, $mid0) = @_;
+sub v2_num_for_harder {
+        my ($self, $mime) = @_;
 
         my $hdr = $mime->header_obj;
         my $dig = content_digest($mime);
-        $$mid0 = PublicInbox::Import::digest2mid($dig, $hdr);
-        my $num = $self->{mm}->mid_insert($$mid0);
+        my $mid0 = PublicInbox::Import::digest2mid($dig, $hdr);
+        my $num = $self->{mm}->mid_insert($mid0);
         unless (defined $num) {
                 # it's hard to spoof the last Received: header
                 my @recvd = $hdr->header_raw('Received');
                 $dig->add("Received: $_") foreach (@recvd);
-                $$mid0 = PublicInbox::Import::digest2mid($dig, $hdr);
-                $num = $self->{mm}->mid_insert($$mid0);
+                $mid0 = PublicInbox::Import::digest2mid($dig, $hdr);
+                $num = $self->{mm}->mid_insert($mid0);
 
                 # fall back to a random Message-ID and give up determinism:
                 until (defined($num)) {
                         $dig->add(rand);
-                        $$mid0 = PublicInbox::Import::digest2mid($dig, $hdr);
-                        warn "using random Message-ID <$$mid0> as fallback\n";
-                        $num = $self->{mm}->mid_insert($$mid0);
+                        $mid0 = PublicInbox::Import::digest2mid($dig, $hdr);
+                        warn "using random Message-ID <$mid0> as fallback\n";
+                        $num = $self->{mm}->mid_insert($mid0);
                 }
         }
-        PublicInbox::Import::append_mid($hdr, $$mid0);
-        $num;
+        PublicInbox::Import::append_mid($hdr, $mid0);
+        ($num, $mid0);
 }
 
 sub idx_shard {
@@ -767,7 +764,6 @@ sub import_init {
 # XXX experimental
 sub diff ($$$) {
         my ($mid, $cur, $new) = @_;
-        use File::Temp qw(tempfile);
 
         my ($ah, $an) = tempfile('email-cur-XXXXXXXX', TMPDIR => 1);
         print $ah $cur->as_string or die "print: $!";
@@ -845,81 +841,184 @@ sub mark_deleted ($$$$) {
         }
 }
 
-sub reindex_oid ($$$$) {
-        my ($self, $sync, $git, $oid) = @_;
-        my $len;
+sub reindex_checkpoint ($$$) {
+        my ($self, $sync, $git) = @_;
+
+        $git->cleanup;
+        $sync->{mm_tmp}->atfork_prepare;
+        $self->done; # release lock
+
+        if (my $pr = $sync->{-opt}->{-progress}) {
+                my ($bn) = (split('/', $git->{git_dir}))[-1];
+                $pr->("$bn ".sprintf($sync->{-regen_fmt}, $sync->{nr}));
+        }
+
+        # allow -watch or -mda to write...
+        $self->idx_init; # reacquire lock
+        $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;
 
-        # get the NNTP article number we used before, highest number wins
-        # and gets deleted from sync->{mm_tmp};
-        my $mid0;
-        my $num = -1;
-        my $del = 0;
-        foreach my $mid (@$mids) {
-                $del += delete($sync->{D}->{"$mid\0$cid"}) ? 1 : 0;
-                my $n = $sync->{mm_tmp}->num_for($mid);
-                if (defined $n && $n > $num) {
-                        $mid0 = $mid;
-                        $num = $n;
-                        $self->{mm}->mid_set($num, $mid0);
+        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(#$num)";
                 }
         }
-        if (!defined($mid0) && !$del) {
-                $num = $sync->{regen}--;
-                die "BUG: ran out of article numbers\n" if $num <= 0;
-                my $mm = $self->{mm};
-                foreach my $mid (reverse @$mids) {
-                        if ($mm->mid_set($num, $mid) == 1) {
-                                $mid0 = $mid;
-                                last;
-                        }
+        unless (defined($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, $mid);
+                        last;
                 }
-                if (!defined($mid0)) {
-                        my $id = '<' . join('> <', @$mids) . '>';
-                        warn "Message-ID $id unusable for $num\n";
-                        foreach my $mid (@$mids) {
-                                defined(my $n = $mm->num_for($mid)) or next;
-                                warn "#$n previously mapped for <$mid>\n";
+        }
+        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);
+        }
+}
 
-        if (!defined($mid0) || $del) {
-                if (!defined($mid0) && $del) { # expected for deletes
-                        $num = $sync->{regen}--;
-                        $self->{mm}->num_highwater($num) if !$sync->{reindex};
-                        return
+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 ($$) {
+        my ($sync, $oid) = @_;
+        my $multi_mid = $sync->{multi_mid} //= multi_mid_q_new();
+        if ($sync->{reindex}) { # no regen on reindex
+                $multi_mid->mid_insert($oid);
+        } else {
+                my $num = $sync->{regen}--;
+                die "BUG: ran out of article numbers" if $num <= 0;
+                $multi_mid->mid_set($num, $oid);
+        }
+}
+
+sub reindex_oid ($$$$) {
+        my ($self, $sync, $git, $oid) = @_;
+        my ($num, $mid0, $len);
+        my $msgref = $git->cat_file($oid, \$len);
+        return if $len == 0; # purged
+        my $mime = PublicInbox::MIME->new($$msgref);
+        my $mids = mids($mime->header_obj);
+        my $cid = content_id($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;
                 }
 
-                my $id = '<' . join('> <', @$mids) . '>';
-                defined($mid0) or
-                        warn "Skipping $id, no article number found\n";
-                if ($del && defined($mid0)) {
-                        warn "$id was deleted $del " .
-                                "time(s) but mapped to article #$num\n";
+                # is this a number we got before?
+                $num = $sync->{mm_tmp}->num_for($mid);
+                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;
+                }
+        } else { # multiple MIDs are a weird case:
+                my $del = 0;
+                for (@$mids) {
+                        $del += delete($sync->{D}->{"$_\0$cid"}) // 0;
+                }
+                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);
                 }
                 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)) {
-                $git->cleanup;
-                $sync->{mm_tmp}->atfork_prepare;
-                $self->done; # release lock
-
-                if (my $pr = $sync->{-opt}->{-progress}) {
-                        my ($bn) = (split('/', $git->{git_dir}))[-1];
-                        $pr->("$bn ".sprintf($sync->{-regen_fmt}, $sync->{nr}));
-                }
-
-                # allow -watch or -mda to write...
-                $self->idx_init; # reacquire lock
-                $sync->{mm_tmp}->atfork_parent;
+                reindex_checkpoint($self, $sync, $git);
         }
 }
 
@@ -1051,8 +1150,9 @@ sub unindex_oid_remote ($$$) {
         $self->{over}->remove_oid($oid, $mid);
 }
 
-sub unindex_oid ($$$) {
-        my ($self, $git, $oid) = @_;
+sub unindex_oid ($$$;$) {
+        my ($self, $git, $oid, $unindexed) = @_;
+        my $mm = $self->{mm};
         my $msgref = $git->cat_file($oid);
         my $mime = PublicInbox::MIME->new($msgref);
         my $mids = mids($mime->header_obj);
@@ -1072,8 +1172,11 @@ sub unindex_oid ($$$) {
                                 join(',',sort keys %gone), "\n";
                 }
                 foreach my $num (keys %gone) {
-                        $self->{unindexed}->{$_}++;
-                        $self->{mm}->num_delete($num);
+                        if ($unindexed) {
+                                my $mid0 = $mm->mid_for($num);
+                                $unindexed->{$mid0} = $num;
+                        }
+                        $mm->num_delete($num);
                 }
                 unindex_oid_remote($self, $oid, $mid);
         }
@@ -1082,20 +1185,21 @@ sub unindex_oid ($$$) {
 my $x40 = qr/[a-f0-9]{40}/;
 sub unindex ($$$$) {
         my ($self, $sync, $git, $unindex_range) = @_;
-        my $un = $self->{unindexed} ||= {}; # num => removal count
-        my $before = scalar keys %$un;
+        my $unindexed = $self->{unindexed} ||= {}; # $mid0 => $num
+        my $before = scalar keys %$unindexed;
+        # order does not matter, here:
         my @cmd = qw(log --raw -r
                         --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$/o or next;
-                unindex_oid($self, $git, $1);
+                unindex_oid($self, $git, $1, $unindexed);
         }
         delete $self->{reindex_pipe};
         $fh = undef;
 
         return unless $sync->{-opt}->{prune};
-        my $after = scalar keys %$un;
+        my $after = scalar keys %$unindexed;
         return if $before == $after;
 
         # ensure any blob can not longer be accessed via dumb HTTP
@@ -1188,11 +1292,34 @@ sub index_sync {
 
         # unindex is required for leftovers if "deletes" affect messages
         # in a previous fetch+index window:
+        my $git;
         if (my @leftovers = values %{delete $sync->{D}}) {
-                my $git = $self->{-inbox}->git;
-                unindex_oid($self, $git, $_) for @leftovers;
-                $git->cleanup;
+                $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;
         $self->done;
 
         if (my $nr = $sync->{nr}) {
diff --git a/t/v2reindex.t b/t/v2reindex.t
index 7c5a6b07..52711f8f 100644
--- a/t/v2reindex.t
+++ b/t/v2reindex.t
@@ -431,4 +431,70 @@ ok(!-d $xap, 'Xapian directories removed again');
                   ], 'msgmap as expected' );
 }
 
+# A real example from linux-renesas-soc on lore where a 3-headed monster
+# of a message has 3 sets of common headers.  Another normal message
+# previously existed with a single Message-ID that conflicts with one
+# of the Message-IDs in the 3-headed monster.
+{
+        my @warn;
+        local $SIG{__WARN__} = sub { push @warn, @_ };
+        my %config = %$ibx_config;
+        $config{indexlevel} = 'basic';
+        my $ibx = PublicInbox::Inbox->new(\%config);
+        my $im = PublicInbox::V2Writable->new($ibx);
+        my $m3 = PublicInbox::MIME->new(<<'EOF');
+Date: Tue, 24 May 2016 14:34:22 -0700 (PDT)
+Message-Id: <20160524.143422.552507610109476444.d@example.com>
+To: t@example.com
+Cc: c@example.com
+Subject: Re: [PATCH v2 2/2]
+From: <f@example.com>
+In-Reply-To: <1463825855-7363-2-git-send-email-y@example.com>
+References: <1463825855-7363-1-git-send-email-y@example.com>
+        <1463825855-7363-2-git-send-email-y@example.com>
+Date: Wed, 25 May 2016 10:01:51 +0900
+From: h@example.com
+To: g@example.com
+Cc: m@example.com
+Subject: Re: [PATCH]
+Message-ID: <20160525010150.GD7292@example.com>
+References: <1463498133-23918-1-git-send-email-g+r@example.com>
+In-Reply-To: <1463498133-23918-1-git-send-email-g+r@example.com>
+From: s@example.com
+To: h@example.com
+Cc: m@example.com
+Subject: [PATCH 12/13]
+Date: Wed, 01 Jun 2016 01:32:35 +0300
+Message-ID: <1923946.Jvi0TDUXFC@wasted.example.com>
+In-Reply-To: <13205049.n7pM8utpHF@wasted.example.com>
+References: <13205049.n7pM8utpHF@wasted.example.com>
+
+Somehow we got a message with 3 sets of headers into one
+message, could've been something broken on the archiver side.
+EOF
+
+        my $m1 = PublicInbox::MIME->new(<<'EOF');
+From: a@example.com
+To: t@example.com
+Subject: [PATCH 12/13]
+Date: Wed, 01 Jun 2016 01:32:35 +0300
+Message-ID: <1923946.Jvi0TDUXFC@wasted.example.com>
+In-Reply-To: <13205049.n7pM8utpHF@wasted.example.com>
+References: <13205049.n7pM8utpHF@wasted.example.com>
+
+This is probably one of the original messages
+
+EOF
+        $im->add($m1);
+        $im->add($m3);
+        $im->done;
+        remove_tree($xap);
+        eval { $im->index_sync() };
+        is($@, '', 'no error from initial indexing');
+        is_deeply(\@warn, [], 'no warnings from initial index');
+        eval { $im->index_sync({reindex=>1}) };
+        is($@, '', 'no error from reindexing after reused Message-ID (x3)');
+        is_deeply(\@warn, [], 'no warnings on reindex');
+}
+
 done_testing();