X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FLeiToMail.pm;h=90db30e53089942be3c512033e23f17147089e39;hb=97a9139a75729c41bb91acd711e8b088e6379600;hp=997ce5993c897c653b76426c366bee4b5d67b7f6;hpb=528d69bbbda3cd26d9cc317eff92d21d980b2056;p=public-inbox.git diff --git a/lib/PublicInbox/LeiToMail.pm b/lib/PublicInbox/LeiToMail.pm index 997ce599..90db30e5 100644 --- a/lib/PublicInbox/LeiToMail.pm +++ b/lib/PublicInbox/LeiToMail.pm @@ -9,10 +9,10 @@ use parent qw(PublicInbox::IPC); use PublicInbox::Eml; use PublicInbox::ProcessPipe; use PublicInbox::Spawn qw(spawn); -use PublicInbox::PktOp qw(pkt_do); use Symbol qw(gensym); use IO::Handle; # ->autoflush use Fcntl qw(SEEK_SET SEEK_END O_CREAT O_EXCL O_WRONLY); +use PublicInbox::Syscall qw(rename_noreplace); my %kw2char = ( # Maildir characters draft => 'D', @@ -110,39 +110,33 @@ sub _mboxcl_common ($$$) { $$buf .= 'Content-Length: '.length($$bdy).$crlf. 'Lines: '.$lines.$crlf.$crlf; substr($$bdy, 0, 0, $$buf); # prepend header - $_[0] = $bdy; + $$bdy .= $crlf; + $bdy; } # mboxcl still escapes "From " lines sub eml2mboxcl { my ($eml, $smsg) = @_; my $buf = _mbox_hdr_buf($eml, 'mboxcl', $smsg); - my $crlf = $eml->{crlf}; - if (my $bdy = delete $eml->{bdy}) { - $$bdy =~ s/^From />From /gm; - _mboxcl_common($buf, $bdy, $crlf); - } - $$buf .= $crlf; - $buf; + my $bdy = delete($eml->{bdy}) // \(my $empty = ''); + $$bdy =~ s/^From />From /gm; + _mboxcl_common($buf, $bdy, $eml->{crlf}); } # mboxcl2 has no "From " escaping sub eml2mboxcl2 { my ($eml, $smsg) = @_; my $buf = _mbox_hdr_buf($eml, 'mboxcl2', $smsg); - my $crlf = $eml->{crlf}; - if (my $bdy = delete $eml->{bdy}) { - _mboxcl_common($buf, $bdy, $crlf); - } - $$buf .= $crlf; - $buf; + my $bdy = delete($eml->{bdy}) // \(my $empty = ''); + _mboxcl_common($buf, $bdy, $eml->{crlf}); } sub git_to_mail { # git->cat_async callback my ($bref, $oid, $type, $size, $arg) = @_; + $type // return; # called by git->async_abort my ($write_cb, $smsg) = @$arg; - if ($type eq 'missing' && $smsg->{-lms_ro}) { - if ($bref = $smsg->{-lms_ro}->local_blob($oid, 1)) { + if ($type eq 'missing' && $smsg->{-lms_rw}) { + if ($bref = $smsg->{-lms_rw}->local_blob($oid, 1)) { $type = 'blob'; $size = length($$bref); } @@ -165,7 +159,7 @@ sub _post_augment_mbox { # open a compressor process from top-level process my $zsfx = $self->{zsfx} or return; my $cmd = PublicInbox::MboxReader::zsfx2cmd($zsfx, undef, $lei); my ($r, $w) = @{delete $lei->{zpipe}}; - my $rdr = { 0 => $r, 1 => $lei->{1}, 2 => $lei->{2} }; + my $rdr = { 0 => $r, 1 => $lei->{1}, 2 => $lei->{2}, pgid => 0 }; my $pid = spawn($cmd, undef, $rdr); my $pp = gensym; my $dup = bless { "pid.$pid" => $cmd }, ref($lei); @@ -199,7 +193,7 @@ sub _mbox_write_cb ($$) { my $dedupe = $lei->{dedupe}; $dedupe->prepare_dedupe; my $lse = $lei->{lse}; # may be undef - my $set_recent = $dedupe->dedupe_nr; + my $set_recent = $dedupe->has_entries; sub { # for git_to_mail my ($buf, $smsg, $eml) = @_; $eml //= PublicInbox::Eml->new($buf); @@ -223,14 +217,14 @@ sub update_kw_maybe ($$$$) { my $c = $lse->kw_changed($eml, $kw, my $docids = []); my $vmd = { kw => $kw }; if (scalar @$docids) { # already in lei/store - $lei->{sto}->ipc_do('set_eml_vmd', undef, $vmd, $docids) if $c; + $lei->{sto}->wq_do('set_eml_vmd', undef, $vmd, $docids) if $c; } elsif (my $xoids = $lei->{ale}->xoids_for($eml)) { # it's in an external, only set kw, here - $lei->{sto}->ipc_do('set_xvmd', $xoids, $eml, $vmd); + $lei->{sto}->wq_do('set_xvmd', $xoids, $eml, $vmd); } else { # never-before-seen, import the whole thing # XXX this is critical in protecting against accidental # data loss without --augment - $lei->{sto}->ipc_do('set_eml', $eml, $vmd); + $lei->{sto}->wq_do('set_eml', $eml, $vmd); } } @@ -269,10 +263,8 @@ sub _buf2maildir ($$$$) { $rand = ''; do { $base = $rand.$common.':2,'.kw2suffix($kw); - } while (!($ok = link($tmp, $dst.$base)) && $!{EEXIST} && - ($rand = _rand.',')); - die "link($tmp, $dst$base): $!" unless $ok; - unlink($tmp) or warn "W: failed to unlink $tmp: $!\n"; + } while (!($ok = rename_noreplace($tmp, $dst.$base)) && + $!{EEXIST} && ($rand = _rand.',')); \$base; } else { my $err = "Error writing $smsg->{blob} to $dst: $!\n"; @@ -288,13 +280,14 @@ sub _maildir_write_cb ($$) { $dedupe->prepare_dedupe if $dedupe; my $dst = $lei->{ovv}->{dst}; my $lse = $lei->{lse}; # may be undef - my $sto = $lei->{opt}->{'mail-sync'} ? $lei->{sto} : undef; - my $out = $sto ? 'maildir:'.$lei->abs_path($dst) : undef; + my $lms = $self->{-lms_rw}; + my $out = $lms ? 'maildir:'.$lei->abs_path($dst) : undef; + $lms->lms_write_prepare if $lms; # Favor cur/ and only write to new/ when augmenting. This # saves MUAs from having to do a mass rename when the initial # search result set is huge. - my $dir = $dedupe && $dedupe->dedupe_nr ? 'new/' : 'cur/'; + my $dir = $dedupe && $dedupe->has_entries ? 'new/' : 'cur/'; sub { # for git_to_mail my ($bref, $smsg, $eml) = @_; $dst // return $lei->fail; # dst may be undef-ed in last run @@ -304,7 +297,7 @@ sub _maildir_write_cb ($$) { $lse->xsmsg_vmd($smsg) if $lse; my $n = _buf2maildir($dst, $bref // \($eml->as_string), $smsg, $dir); - $sto->ipc_do('set_sync_info', $smsg->{blob}, $out, $n) if $sto; + $lms->set_src($smsg->oidbin, $out, $n) if $lms; ++$lei->{-nr_write}; } } @@ -319,7 +312,8 @@ sub _imap_write_cb ($$) { my $folder = $uri->mailbox; $uri->uidvalidity($mic->uidvalidity($folder)); my $lse = $lei->{lse}; # may be undef - my $sto = $lei->{opt}->{'mail-sync'} ? $lei->{sto} : undef; + my $lms = $self->{-lms_rw}; + $lms->lms_write_prepare if $lms; sub { # for git_to_mail my ($bref, $smsg, $eml) = @_; $mic // return $lei->fail; # mic may be undef-ed in last run @@ -333,9 +327,8 @@ sub _imap_write_cb ($$) { die $err; } # imap_append returns UID if IMAP server has UIDPLUS extension - ($sto && $uid =~ /\A[0-9]+\z/) and - $sto->ipc_do('set_sync_info', - $smsg->{blob}, $$uri, $uid + 0); + ($lms && $uid =~ /\A[0-9]+\z/) and + $lms->set_src($smsg->oidbin, $$uri, $uid + 0); ++$lei->{-nr_write}; } } @@ -368,7 +361,8 @@ sub _v2_write_cb ($$) { my ($bref, $smsg, $eml) = @_; $eml //= PublicInbox::Eml->new($bref); return if $dedupe && $dedupe->is_dup($eml, $smsg); - $lei->{v2w}->ipc_do('add', $eml); # V2Writable->add + $lei->{v2w}->wq_do('add', $eml); # V2Writable->add + ++$lei->{-nr_write}; } } @@ -384,23 +378,31 @@ sub new { my $fmt = $lei->{ovv}->{fmt}; my $dst = $lei->{ovv}->{dst}; my $self = bless {}, $cls; + my @conflict; if ($fmt eq 'maildir') { require PublicInbox::MdirReader; $self->{base_type} = 'maildir'; -e $dst && !-d _ and die "$dst exists and is not a directory\n"; $lei->{ovv}->{dst} = $dst .= '/' if substr($dst, -1) ne '/'; + $lei->{opt}->{save} //= \1 if $lei->{cmd} eq 'q'; } elsif (substr($fmt, 0, 4) eq 'mbox') { require PublicInbox::MboxReader; $self->can("eml2$fmt") or die "bad mbox format: $fmt\n"; $self->{base_type} = 'mbox'; + if ($lei->{cmd} eq 'q' && + (($lei->path_to_fd($dst) // -1) < 0) && + (-f $dst || !-e _)) { + $lei->{opt}->{save} //= \1; + } } elsif ($fmt =~ /\Aimaps?\z/) { require PublicInbox::NetWriter; require PublicInbox::URIimap; - my $net = PublicInbox::NetWriter->new; + # {net} may exist from "lei up" for auth + my $net = $lei->{net} // PublicInbox::NetWriter->new; $net->{quiet} = $lei->{opt}->{quiet}; my $uri = PublicInbox::URIimap->new($dst)->canonical; - $net->add_url($uri); + $net->add_url($$uri); my $err = $net->errors($lei); return $lei->fail($err) if $err; $uri->mailbox or return $lei->fail("No mailbox: $dst"); @@ -408,17 +410,21 @@ sub new { $dst = $lei->{ovv}->{dst} = $$uri; # canonicalized $lei->{net} = $net; $self->{base_type} = 'imap'; - } elsif ($fmt eq 'text') { + $lei->{opt}->{save} //= \1 if $lei->{cmd} eq 'q'; + } elsif ($fmt eq 'text' || $fmt eq 'reply') { require PublicInbox::LeiViewText; - $lei->{lvt} = PublicInbox::LeiViewText->new($lei); + $lei->{lvt} = PublicInbox::LeiViewText->new($lei, $fmt); $self->{base_type} = 'text'; + $self->{-wq_nr_workers} = 1; # for pager + @conflict = qw(mua save); } elsif ($fmt eq 'v2') { die "--dedupe=oid and v2 are incompatible\n" if ($lei->{opt}->{dedupe}//'') eq 'oid'; $self->{base_type} = 'v2'; + $self->{-wq_nr_workers} = 1; # v2 has shards $lei->{opt}->{save} = \1; - die "--mua incompatible with v2\n" if $lei->{opt}->{mua}; $dst = $lei->{ovv}->{dst} = $lei->abs_path($dst); + @conflict = qw(mua sort); } else { die "bad mail --format=$fmt\n"; } @@ -426,19 +432,17 @@ sub new { (-d $dst || (-e _ && !-w _)) and die "$dst exists and is not a writable file\n"; } - if ($self->{base_type} eq 'text') { - my @err = map { - defined($lei->{opt}->{$_}) ? "--$_" : (); - } (qw(mua save)); - die "@err incompatible with $fmt\n" if @err; - } + my @err = map { defined($lei->{opt}->{$_}) ? "--$_" : () } @conflict; + die "@err incompatible with $fmt\n" if @err; $self->{dst} = $dst; $lei->{dedupe} = $lei->{lss} // do { my $dd_cls = 'PublicInbox::'. ($lei->{opt}->{save} ? 'LeiSavedSearch' : 'LeiDedupe'); eval "require $dd_cls"; die "$dd_cls: $@" if $@; - $dd_cls->new($lei); + my $dd = $dd_cls->new($lei); + $lei->{lss} //= $dd if $dd && $dd->can('cfg_set'); + $dd; }; $self; } @@ -457,9 +461,19 @@ sub _pre_augment_maildir { open $self->{poke_dh}, '<', "${dst}cur" or die "open ${dst}cur: $!"; } +sub clobber_dst_prepare ($;$) { + my ($lei, $f) = @_; + if (my $lms = defined($f) ? $lei->lms : undef) { + $lms->lms_write_prepare; + $lms->forget_folders($f); + } + my $dedupe = $lei->{dedupe} or return; + $dedupe->reset_dedupe if $dedupe->can('reset_dedupe'); +} + sub _do_augment_maildir { my ($self, $lei) = @_; - return if ($lei->{opt}->{save} // 0) < 0; + return if $lei->{cmd} eq 'up'; my $dst = $lei->{ovv}->{dst}; my $lse = $lei->{opt}->{'import-before'} ? $lei->{lse} : undef; my $mdr = PublicInbox::MdirReader->new; @@ -471,9 +485,11 @@ sub _do_augment_maildir { $dedupe->pause_dedupe; } } elsif ($lse) { + clobber_dst_prepare($lei, "maildir:$dst"); $mdr->{shard_info} = $self->{shard_info}; $mdr->maildir_each_eml($dst, \&_md_update, $lei, $lse, 1); } else {# clobber existing Maildir + clobber_dst_prepare($lei, "maildir:$dst"); $mdr->maildir_each_file($dst, \&_unlink); } } @@ -490,7 +506,7 @@ sub _imap_augment_or_delete { # PublicInbox::NetReader::imap_each cb sub _do_augment_imap { my ($self, $lei) = @_; - return if ($lei->{opt}->{save} // 0) < 0; + return if $lei->{cmd} eq 'up'; my $net = $lei->{net}; my $lse = $lei->{opt}->{'import-before'} ? $lei->{lse} : undef; if ($lei->{opt}->{augment}) { @@ -502,11 +518,13 @@ sub _do_augment_imap { } } elsif ($lse) { my $delete_mic; + clobber_dst_prepare($lei, "$self->{uri}"); $net->imap_each($self->{uri}, \&_imap_augment_or_delete, $lei, $lse, \$delete_mic); $delete_mic->expunge if $delete_mic; } elsif (!$self->{-wq_worker_nr}) { # undef or 0 # clobber existing IMAP folder + clobber_dst_prepare($lei, "$self->{uri}"); $net->imap_delete_all($self->{uri}); } } @@ -566,6 +584,8 @@ sub _pre_augment_mbox { if $imp_before && !ref($imp_before); die "--augment specified but $dst is not seekable\n" if $lei->{opt}->{augment}; + die "cannot --save with unseekable $dst\n" if + $lei->{dedupe} && $lei->{dedupe}->can('reset_dedupe'); } if ($self->{zsfx} = PublicInbox::MboxReader::zsfx($dst)) { pipe(my ($r, $w)) or die "pipe: $!"; @@ -584,10 +604,10 @@ sub _do_augment_mbox { my ($self, $lei) = @_; return unless $self->{seekable}; my $opt = $lei->{opt}; - return if ($opt->{save} // 0) < 0; + return if $lei->{cmd} eq 'up'; my $out = $lei->{1}; my ($fmt, $dst) = @{$lei->{ovv}}{qw(fmt dst)}; - return unless -s $out; + return clobber_dst_prepare($lei) unless -s $out; unless ($opt->{augment} || $opt->{'import-before'}) { truncate($out, 0) or die "truncate($dst): $!"; return; @@ -602,6 +622,8 @@ sub _do_augment_mbox { if ($opt->{augment}) { $dedupe = $lei->{dedupe}; $dedupe->prepare_dedupe if $dedupe; + } else { + clobber_dst_prepare($lei); } if ($opt->{'import-before'}) { # the default my $lse = $lei->{lse}; @@ -618,6 +640,12 @@ sub _do_augment_mbox { $dedupe->pause_dedupe if $dedupe; } +sub v2w_done_wait { # dwaitpid callback + my ($arg, $pid) = @_; + my ($v2w, $lei) = @$arg; + $lei->child_error($?, "error for $v2w->{ibx}->{inboxdir}") if $?; +} + sub _pre_augment_v2 { my ($self, $lei) = @_; my $dir = $self->{dst}; @@ -639,9 +667,10 @@ sub _pre_augment_v2 { } PublicInbox::InboxWritable->new($ibx, @creat); $ibx->init_inbox if @creat; - my $v2w = $lei->{v2w} = $ibx->importer; - $v2w->ipc_lock_init("$dir/ipc.lock"); - $v2w->ipc_worker_spawn("lei/v2w $dir", $lei->oldset, { lei => $lei }); + my $v2w = $ibx->importer; + $v2w->wq_workers_start("lei/v2w $dir", 1, $lei->oldset, {lei => $lei}); + $v2w->wq_wait_async(\&v2w_done_wait, $lei); + $lei->{v2w} = $v2w; return if !$lei->{opt}->{shared}; my $d = "$lei->{ale}->{git}->{git_dir}/objects"; my $al = "$dir/git/0.git/objects/info/alternates"; @@ -669,8 +698,10 @@ sub do_augment { # slow, runs in wq worker # fast (spawn compressor or mkdir), runs in same process as pre_augment sub post_augment { my ($self, $lei, @args) = @_; + $self->{-au_noted}++ and $lei->qerr("# writing to $self->{dst} ..."); + my $wait = $lei->{opt}->{'import-before'} ? - $lei->{sto}->ipc_do('checkpoint', 1) : 0; + $lei->{sto}->wq_do('checkpoint', 1) : 0; # _post_augment_mbox my $m = $self->can("_post_augment_$self->{base_type}") or return; $m->($self, $lei, @args); @@ -681,7 +712,7 @@ sub do_post_auth { my ($self) = @_; my $lei = $self->{lei}; # lei_xsearch can start as soon as all l2m workers get here - pkt_do($lei->{pkt_op_p}, 'incr_start_query') or + $lei->{pkt_op_p}->pkt_do('incr_start_query') or die "incr_start_query: $!"; my $aug; if (lock_free($self)) { # all workers do_augment @@ -692,16 +723,15 @@ sub do_post_auth { } else { # Maildir $self->{shard_info} = [ $mod, $shard ]; } - $aug = '+'; # incr_post_augment + $aug = 'incr_post_augment'; } elsif ($self->{-wq_worker_nr} == 0) { # 1st worker do_augment - $aug = '.'; # do_post_augment + $aug = 'do_post_augment'; } if ($aug) { local $0 = 'do_augment'; eval { do_augment($self, $lei) }; $lei->fail($@) if $@; - pkt_do($lei->{pkt_op_p}, $aug) == 1 or - die "do_post_augment trigger: $!"; + $lei->{pkt_op_p}->pkt_do($aug) or die "pkt_do($aug): $!"; } # done augmenting, connect the compressor pipe for each worker if (my $zpipe = delete $lei->{zpipe}) { @@ -724,9 +754,6 @@ sub ipc_atfork_child { my ($self) = @_; my $lei = $self->{lei}; $lei->_lei_atfork_child; - if (my $lse = $lei->{lse}) { - $self->{-lms_ro} = $lse->{-lms_ro} //= $lse->lms; - } $lei->{auth}->do_auth_atfork($self) if $lei->{auth}; $SIG{__WARN__} = PublicInbox::Eml::warn_ignore_cb(); $self->SUPER::ipc_atfork_child; @@ -748,29 +775,49 @@ sub poke_dst { sub write_mail { # via ->wq_io_do my ($self, $smsg, $eml) = @_; return $self->{wcb}->(undef, $smsg, $eml) if $eml; - $smsg->{-lms_ro} = $self->{-lms_ro}; + $smsg->{-lms_rw} = $self->{-lms_rw}; $self->{lei}->{ale}->git->cat_async($smsg->{blob}, \&git_to_mail, [$self->{wcb}, $smsg]); } sub wq_atexit_child { my ($self) = @_; + local $PublicInbox::DS::in_loop = 0; # waitpid synchronously my $lei = $self->{lei}; delete $self->{wcb}; $lei->{ale}->git->async_wait_all; my $nr = delete($lei->{-nr_write}) or return; return if $lei->{early_mua} || !$lei->{-progress} || !$lei->{pkt_op_p}; - require PublicInbox::PktOp; - PublicInbox::PktOp::pkt_do($lei->{pkt_op_p}, 'l2m_progress', $nr); + $lei->{pkt_op_p}->pkt_do('l2m_progress', $nr); +} + +# runs on a 1s timer in lei-daemon +sub augment_inprogress { + my ($err, $opt, $dst, $au_noted) = @_; + eval { + return if $$au_noted++ || !$err || !defined(fileno($err)); + print $err '# '.($opt->{'import-before'} ? + "importing non-external contents of $dst" : ( + ($opt->{dedupe} // 'content') ne 'none') ? + "scanning old contents of $dst for dedupe" : + "removing old contents of $dst")." ...\n"; + }; + warn "E: $@ ($dst)" if $@; } # called in top-level lei-daemon when LeiAuth is done sub net_merge_all_done { - my ($self) = @_; + my ($self, $lei) = @_; + if ($PublicInbox::DS::in_loop && + $self->can("_do_augment_$self->{base_type}") && + !$lei->{opt}->{quiet}) { + $self->{-au_noted} = 0; + PublicInbox::DS::add_timer(1, \&augment_inprogress, + $lei->{2}, $lei->{opt}, + $self->{dst}, \$self->{-au_noted}); + } $self->wq_broadcast('do_post_auth'); - $self->wq_close(1); + $self->wq_close; } -no warnings 'once'; # the following works even when LeiAuth is lazy-loaded -*net_merge_all = \&PublicInbox::LeiAuth::net_merge_all; 1;