X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FV2Writable.pm;h=b5332da43bfc6e30990c1f227aabc689c5a3c77a;hb=99d6d24847c499387ce65b09305deebfd51f8930;hp=37d27302565e17b66586b00e5742a6b2ef451f41;hpb=2710d3105e70ff467eff9e977325628b4e9dd1c5;p=public-inbox.git diff --git a/lib/PublicInbox/V2Writable.pm b/lib/PublicInbox/V2Writable.pm index 37d27302..b5332da4 100644 --- a/lib/PublicInbox/V2Writable.pm +++ b/lib/PublicInbox/V2Writable.pm @@ -1,4 +1,4 @@ -# Copyright (C) 2018-2019 all contributors +# Copyright (C) 2018-2020 all contributors # License: AGPL-3.0+ # This interface wraps and mimics PublicInbox::Import @@ -7,6 +7,7 @@ package PublicInbox::V2Writable; use strict; use warnings; use base qw(PublicInbox::Lock); +use 5.010_001; use PublicInbox::SearchIdxShard; use PublicInbox::MIME; use PublicInbox::Git; @@ -32,14 +33,29 @@ my $PACKING_FACTOR = 0.4; # to increase Xapian shards our $NPROC_MAX_DEFAULT = 4; +sub detect_nproc () { + for my $nproc (qw(nproc gnproc)) { # GNU coreutils nproc + `$nproc 2>/dev/null` =~ /^(\d+)$/ and return $1; + } + + # getconf(1) is POSIX, but *NPROCESSORS* vars are not + for (qw(_NPROCESSORS_ONLN NPROCESSORS_ONLN)) { + `getconf $_ 2>/dev/null` =~ /^(\d+)$/ and return $1; + } + + # should we bother with `sysctl hw.ncpu`? Those only give + # us total processor count, not online processor count. + undef +} + sub nproc_shards ($) { my ($creat_opt) = @_; my $n = $creat_opt->{nproc} if ref($creat_opt) eq 'HASH'; $n //= $ENV{NPROC}; if (!$n) { - chomp($n = `nproc 2>/dev/null`); - # assume 2 cores if GNU nproc(1) is not available - $n = 2 if !$n; + # assume 2 cores if not detectable or zero + state $NPROC_DETECTED = detect_nproc() || 2; + $n = $NPROC_DETECTED; $n = $NPROC_MAX_DEFAULT if $n > $NPROC_MAX_DEFAULT; } @@ -107,8 +123,11 @@ sub new { # public (for now?) sub init_inbox { - my ($self, $parallel, $skip_epoch) = @_; - $self->{parallel} = $parallel; + my ($self, $shards, $skip_epoch) = @_; + if (defined $shards) { + $self->{parallel} = 0 if $shards == 0; + $self->{shards} = $shards if $shards > 0; + } $self->idx_init; my $epoch_max = -1; git_dir_latest($self, \$epoch_max); @@ -129,12 +148,12 @@ sub add { } # indexes a message, returns true if checkpointing is needed -sub do_idx ($$$$$$$) { - my ($self, $msgref, $mime, $len, $num, $oid, $mid0) = @_; - $self->{over}->add_overview($mime, $len, $num, $oid, $mid0); - my $idx = idx_shard($self, $num % $self->{shards}); - $idx->index_raw($len, $msgref, $num, $oid, $mid0, $mime); - my $n = $self->{transact_bytes} += $len; +sub do_idx ($$$$) { + my ($self, $msgref, $mime, $smsg) = @_; + $self->{over}->add_overview($mime, $smsg, $self); + my $idx = idx_shard($self, $smsg->{num} % $self->{shards}); + $idx->index_raw($msgref, $mime, $smsg, $self); + my $n = $self->{transact_bytes} += $smsg->{bytes}; $n >= (PublicInbox::SearchIdx::BATCH_BYTES * $self->{shards}); } @@ -157,12 +176,14 @@ sub _add { defined $num or return; # duplicate defined $mid0 or die "BUG: $mid0 undefined\n"; my $im = $self->importer; - my $cmt = $im->add($mime); + my $cmt = $im->add($mime, undef, $self); # sets $self->{(au|co)time} $cmt = $im->get_mark($cmt); $self->{last_commit}->[$self->{epoch_max}] = $cmt; - my ($oid, $len, $msgref) = @{$im->{last_object}}; - if (do_idx($self, $msgref, $mime, $len, $num, $oid, $mid0)) { + my $msgref; + my $smsg = bless { mid => $mid0, num => $num }, 'PublicInbox::Smsg'; + ($smsg->{blob}, $smsg->{bytes}, $msgref) = @{$im->{last_object}}; + if (do_idx($self, $msgref, $mime, $smsg)) { $self->checkpoint; } @@ -182,11 +203,10 @@ sub v2_num_for { # crap, Message-ID is already known, hope somebody just resent: foreach my $m (@$mids) { # read-only lookup now safe to do after above barrier - my $existing = lookup_content($self, $mime, $m); # easy, don't store duplicates # note: do not add more diagnostic info here since # it gets noisy on public-inbox-watch restarts - return () if $existing; + return () if content_exists($self, $mime, $m); } # AltId may pre-populate article numbers (e.g. X-Mail-Count @@ -362,7 +382,7 @@ sub rewrite_internal ($$;$$$) { } my $over = $self->{over}; my $cids = content_ids($old_mime); - my $removed; + my @removed; my $mids = mids($old_mime->header_obj); # We avoid introducing new blobs into git since the raw content @@ -372,7 +392,7 @@ sub rewrite_internal ($$;$$$) { my $mark; foreach my $mid (@$mids) { - my %gone; # num => [ smsg, raw ] + my %gone; # num => [ smsg, $mime, raw ] my ($id, $prev); while (my $smsg = $over->next_by_mid($mid, \$id, \$prev)) { my $msg = get_blob($self, $smsg); @@ -383,8 +403,7 @@ sub rewrite_internal ($$;$$$) { my $orig = $$msg; my $cur = PublicInbox::MIME->new($msg); if (content_matches($cids, $cur)) { - $smsg->{mime} = $cur; - $gone{$smsg->{num}} = [ $smsg, \$orig ]; + $gone{$smsg->{num}} = [ $smsg, $cur, \$orig ]; } } my $n = scalar keys %gone; @@ -394,15 +413,16 @@ sub rewrite_internal ($$;$$$) { join(',', sort keys %gone), "\n"; } foreach my $num (keys %gone) { - my ($smsg, $orig) = @{$gone{$num}}; - # $removed should only be set once assuming + my ($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 @@ -422,15 +442,18 @@ sub rewrite_internal ($$;$$$) { my $rewrites = _replace_oids($self, $new_mime, $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) = @_; + my @ret; $self->{-inbox}->with_umask(sub { - rewrite_internal($self, $mime, $cmt_msg); + @ret = rewrite_internal($self, $mime, $cmt_msg); }); + defined($ret[0]) ? @ret : undef; } sub _replace ($$;$$) { @@ -530,17 +553,21 @@ W: $list } # make sure we really got the OID: - my ($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, + bytes => $bytes, + num => $smsg->{num}, + mid => $smsg->{mid}, + }, 'PublicInbox::Smsg'; + do_idx($self, \$raw, $new_mime, $new_smsg); } $rewritten->{rewrites}; } @@ -802,7 +829,7 @@ sub get_blob ($$) { $ibx->msg_by_smsg($smsg); } -sub lookup_content ($$$) { +sub content_exists ($$$) { my ($self, $mime, $mid) = @_; my $over = $self->{over}; my $cids = content_ids($mime); @@ -814,11 +841,7 @@ sub lookup_content ($$$) { next; } my $cur = PublicInbox::MIME->new($msg); - if (content_matches($cids, $cur)) { - $smsg->{mime} = $cur; - return $smsg; - } - + return 1 if content_matches($cids, $cur); # XXX DEBUG_DIFF is experimental and may be removed diff($mid, $cur, $mime) if $ENV{DEBUG_DIFF}; @@ -932,7 +955,13 @@ sub reindex_oid_m ($$$$;$) { } } $sync->{nr}++; - if (do_idx($self, $msgref, $mime, $len, $num, $oid, $mid0)) { + my $smsg = bless { + bytes => $len, + num => $num, + blob => $oid, + mid => $mid0, + }, 'PublicInbox::Smsg'; + if (do_idx($self, $msgref, $mime, $smsg)) { reindex_checkpoint($self, $sync, $git); } } @@ -1028,7 +1057,13 @@ sub reindex_oid ($$$$) { $sync->{mm_tmp}->mid_delete($mid0) or die "failed to delete <$mid0> for article #$num\n"; $sync->{nr}++; - if (do_idx($self, $msgref, $mime, $len, $num, $oid, $mid0)) { + my $smsg = bless { + bytes => $len, + num => $num, + blob => $oid, + mid => $mid0, + }, 'PublicInbox::Smsg'; + if (do_idx($self, $msgref, $mime, $smsg)) { reindex_checkpoint($self, $sync, $git); } } @@ -1249,15 +1284,17 @@ sub index_epoch ($$$) { $pr->("$i.git indexing $range\n"); } - my @cmd = qw(log --raw -r --pretty=tformat:%H + my @cmd = qw(log --raw -r --pretty=tformat:%H.%at.%ct --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 = $_; + if (/\A($x40)\.([0-9]+)\.([0-9]+)$/o) { + $cmt //= $1; + $self->{autime} = $2; + $self->{cotime} = $3; } 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) {