X-Git-Url: http://www.git.stargrave.org/?p=public-inbox.git;a=blobdiff_plain;f=lib%2FPublicInbox%2FNNTP.pm;h=e51793505cf515851f7031dc65c982918076f89c;hp=e86832107ca560db8f10dd834e801c4a5c5c42ed;hb=b8c41362f2a5c8fcc6b1846a79c72bfa77565297;hpb=b6974dbc81665427413020414a668ddb742e68f9 diff --git a/lib/PublicInbox/NNTP.pm b/lib/PublicInbox/NNTP.pm index e8683210..e5179350 100644 --- a/lib/PublicInbox/NNTP.pm +++ b/lib/PublicInbox/NNTP.pm @@ -1,5 +1,5 @@ -# Copyright (C) 2015 all contributors -# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt) +# Copyright (C) 2015-2018 all contributors +# License: AGPL-3.0+ # # Each instance of this represents a NNTP client socket package PublicInbox::NNTP; @@ -9,13 +9,14 @@ use base qw(Danga::Socket); use fields qw(nntpd article rbuf ng long_res); use PublicInbox::Search; use PublicInbox::Msgmap; +use PublicInbox::MID qw(mid_escape); use PublicInbox::Git; -use PublicInbox::MID qw(mid2path); require PublicInbox::EvCleanup; use Email::Simple; use POSIX qw(strftime); use Time::HiRes qw(clock_gettime CLOCK_MONOTONIC); -use URI::Escape qw(uri_escape_utf8); +use Digest::SHA qw(sha1_hex); +use Time::Local qw(timegm timelocal); use constant { r501 => '501 command syntax error', r221 => '221 Header follows', @@ -33,14 +34,11 @@ my $LIST_HEADERS = join("\r\n", @OVERVIEW, qw(:bytes :lines Xref To Cc)) . "\r\n"; # disable commands with easy DoS potential: -# LISTGROUP could get pretty bad, too... my %DISABLED; # = map { $_ => 1 } qw(xover list_overview_fmt newnews xhdr); my $EXPMAP; # fd -> [ idle_time, $self ] my $expt; our $EXPTIME = 180; # 3 minutes -my $WEAKEN = {}; # string(nntpd) -> nntpd -my $weakt; my $nextt; my $nextq = []; @@ -61,19 +59,8 @@ sub next_tick () { sub update_idle_time ($) { my ($self) = @_; - my $tmp = $self->{sock} or return; - $tmp = fileno($tmp); - defined $tmp and $EXPMAP->{$tmp} = [ now(), $self ]; -} - -# reduce FD pressure by closing some "git cat-file --batch" processes -# and unused FDs for msgmap and Xapian indices -sub weaken_groups () { - $weakt = undef; - foreach my $nntpd (values %$WEAKEN) { - $_->weaken_all foreach (@{$nntpd->{grouplist}}); - } - $WEAKEN = {}; + my $fd = $self->{fd}; + defined $fd and $EXPMAP->{$fd} = [ now(), $self ]; } sub expire_old () { @@ -94,15 +81,11 @@ sub expire_old () { $EXPMAP = \%new; if ($nr) { $expt = PublicInbox::EvCleanup::later(*expire_old); - weaken_groups(); } else { $expt = undef; # noop to kick outselves out of the loop ASAP so descriptors # really get closed PublicInbox::EvCleanup::asap(sub {}); - - # grace period for reaping resources - $weakt ||= PublicInbox::EvCleanup::later(*weaken_groups); } } @@ -115,7 +98,6 @@ sub new ($$$) { $self->{rbuf} = ''; $self->watch_read(1); update_idle_time($self); - $WEAKEN->{"$nntpd"} = $nntpd; $expt ||= PublicInbox::EvCleanup::later(*expire_old); $self; } @@ -144,6 +126,7 @@ sub process_line ($$) { my $res = eval { $req->($self, @args) }; my $err = $@; if ($err && !$self->{closed}) { + local $/ = "\n"; chomp($l); err($self, 'error from: %s (%s)', $l, $err); $res = '503 program fault - command not performed'; @@ -241,21 +224,17 @@ sub cmd_listgroup ($;$) { } $self->{ng} or return '412 no newsgroup selected'; - long_response($self, 0, long_response_limit, sub { - my ($i) = @_; - my $nr = $self->{ng}->mm->id_batch($$i, sub { - my ($ary) = @_; - more($self, join("\r\n", @$ary)); - }); - - # -1 to adjust for implicit increment in long_response - $$i = $nr ? $$i + $nr - 1 : long_response_limit; + my $n = 0; + long_response($self, sub { + my $ary = $self->{ng}->mm->ids_after(\$n); + scalar @$ary or return; + more($self, join("\r\n", @$ary)); + 1; }); } sub parse_time ($$;$) { my ($date, $time, $gmt) = @_; - use Time::Local qw(); my ($hh, $mm, $ss) = unpack('A2A2A2', $time); if (defined $gmt) { $gmt =~ /\A(?:UTC|GMT)\z/i or die "GM invalid: $gmt"; @@ -267,15 +246,15 @@ sub parse_time ($$;$) { ($YYYY, $MM, $DD) = unpack('A4A2A2', $date); } else { # legacy clients send YYMMDD ($YYYY, $MM, $DD) = unpack('A2A2A2', $date); - if ($YYYY > strftime('%y', @now)) { - my $cur_year = $now[5] + 1900; + my $cur_year = $now[5] + 1900; + if ($YYYY > $cur_year) { $YYYY += int($cur_year / 1000) * 1000 - 100; } } if ($gmt) { - Time::Local::timegm($ss, $mm, $hh, $DD, $MM - 1, $YYYY); + timegm($ss, $mm, $hh, $DD, $MM - 1, $YYYY); } else { - Time::Local::timelocal($ss, $mm, $hh, $DD, $MM - 1, $YYYY); + timelocal($ss, $mm, $hh, $DD, $MM - 1, $YYYY); } } @@ -304,7 +283,6 @@ sub wildmat2re (;$) { return $_[0] = qr/.*/ if (!defined $_[0] || $_[0] eq '*'); my %keep; my $salt = rand; - use Digest::SHA qw(sha1_hex); my $tmp = $_[0]; $tmp =~ s#(? 1, limit => 1000, offset => 0 }; - long_response($self, 0, long_response_limit, sub { - my ($i) = @_; + my $prev = 0; + long_response($self, sub { my $srch = $srch[0]; - my $res = $srch->query($ts, $opts); - my $msgs = $res->{msgs}; - if (my $nr = scalar @$msgs) { + my $msgs = $srch->query_ts($ts, $prev); + if (scalar @$msgs) { more($self, '<' . join(">\r\n<", map { $_->mid } @$msgs ). '>'); - $opts->{offset} += $nr; + $prev = $msgs->[-1]->{num}; } else { shift @srch; if (@srch) { # continue onto next newsgroup - $opts->{offset} = 0; + $prev = 0; + return 1; } else { # break out of the long response. - $$i = long_response_limit; + return; } } }); @@ -439,7 +415,7 @@ sub set_nntp_headers { $hdr->header_set('Xref', xref($ng, $n)); header_append($hdr, 'List-Post', "{-primary_address}>"); if (my $url = $ng->base_url) { - $mid = uri_escape_utf8($mid); + $mid = mid_escape($mid); header_append($hdr, 'Archived-At', "<$url$mid/>"); header_append($hdr, 'List-Archive', "<$url>"); } @@ -481,19 +457,16 @@ find_mid: defined $mid or return $err; } found: - my $o = 'HEAD:' . mid2path($mid); - my $bytes; - my $s = eval { Email::Simple->new($ng->git->cat_file($o, \$bytes)) }; - return $err unless $s; - my $lines; + my $smsg = $ng->search->{over_ro}->get_art($n) or return $err; + my $msg = $ng->msg_by_smsg($smsg) or return $err; + my $s = Email::Simple->new($msg); if ($set_headers) { set_nntp_headers($s->header_obj, $ng, $n, $mid); - $lines = $s->body =~ tr!\n!\n!; # must be last $s->body_set('') if ($set_headers == 2); } - [ $n, $mid, $s, $bytes, $lines, $ng ]; + [ $n, $mid, $s, $smsg->bytes, $smsg->lines, $ng ]; } sub simple_body_write ($$) { @@ -591,8 +564,8 @@ sub get_range ($$) { [ $beg, $end ]; } -sub long_response ($$$$) { - my ($self, $beg, $end, $cb) = @_; +sub long_response ($$) { + my ($self, $cb) = @_; die "BUG: nested long response" if $self->{long_res}; my $fd = $self->{fd}; @@ -603,23 +576,14 @@ sub long_response ($$$$) { $self->watch_read(0); my $t0 = now(); $self->{long_res} = sub { - # limit our own running time for fairness with other - # clients and to avoid buffering too much: - my $lim = 100; - - my $err; - do { - eval { $cb->(\$beg) }; - } until (($err = $@) || $self->{closed} || - ++$beg > $end || !--$lim || $self->{write_buf_size}); - - if ($err || $self->{closed}) { + my $more = eval { $cb->() }; + if ($@ || $self->{closed}) { $self->{long_res} = undef; - if ($err) { + if ($@) { err($self, "%s during long response[$fd] - %0.6f", - $err, now() - $t0); + $@, now() - $t0); } if ($self->{closed}) { out($self, " deferred[$fd] aborted - %0.6f", @@ -628,7 +592,7 @@ sub long_response ($$$$) { update_idle_time($self); $self->watch_read(1); } - } elsif (!$lim || $self->{write_buf_size}) { + } elsif ($more) { # $self->{write_buf_size}: # no recursion, schedule another call ASAP # but only after all pending writes are done update_idle_time($self); @@ -660,10 +624,13 @@ sub hdr_message_id ($$$) { # optimize XHDR Message-ID [range] for slrnpull. my $mm = $self->{ng}->mm; my ($beg, $end) = @$r; more($self, $xhdr ? r221 : r225); - long_response($self, $beg, $end, sub { - my ($i) = @_; - my $mid = $mm->mid_for($$i); - more($self, "$$i <$mid>") if defined $mid; + long_response($self, sub { + my $r = $mm->msg_range(\$beg, $end); + @$r or return; + more($self, join("\r\n", map { + "$_->[0] <$_->[1]>" + } @$r)); + 1; }); } } @@ -703,27 +670,33 @@ sub hdr_xref ($$$) { # optimize XHDR Xref [range] for rtin my $mm = $ng->mm; my ($beg, $end) = @$r; more($self, $xhdr ? r221 : r225); - long_response($self, $beg, $end, sub { - my ($i) = @_; - my $mid = $mm->mid_for($$i); - more($self, "$$i ".xref($ng, $$i)) if defined $mid; + long_response($self, sub { + my $r = $mm->msg_range(\$beg, $end); + @$r or return; + more($self, join("\r\n", map { + # TODO: use $_->[1] (mid) to fill + # Xref: from other inboxes + my $num = $_->[0]; + "$num ".xref($ng, $num); + } @$r)); + 1; }); } } sub search_header_for { - my ($srch, $mid, $field) = @_; - my $smsg = $srch->lookup_message($mid) or return; - $smsg = PublicInbox::SearchMsg->load_doc($smsg->{doc}); - $smsg->$field; + my ($srch, $num, $field) = @_; + my $smsg = $srch->{over_ro}->get_art($num) or return; + return PublicInbox::SearchMsg::date($smsg) if $field eq 'date'; + $smsg->{$field}; } sub hdr_searchmsg ($$$$) { my ($self, $xhdr, $field, $range) = @_; if (defined $range && $range =~ /\A<(.+)>\z/) { # Message-ID my ($ng, $n) = mid_lookup($self, $1); - return r430 unless $n; - my $v = search_header_for($ng->search, $range, $field); + return r430 unless defined $n; + my $v = search_header_for($ng->search, $n, $field); hdr_mid_response($self, $xhdr, $ng, $n, $range, $v); } else { # numeric range $range = $self->{article} unless defined $range; @@ -733,21 +706,17 @@ sub hdr_searchmsg ($$$$) { return $r unless ref $r; my ($beg, $end) = @$r; more($self, $xhdr ? r221 : r225); - my $off = 0; - long_response($self, $beg, $end, sub { - my ($i) = @_; - my $res = $srch->query_xover($beg, $end, $off); - my $msgs = $res->{msgs}; + my $cur = $beg; + long_response($self, sub { + my $msgs = $srch->query_xover($cur, $end); my $nr = scalar @$msgs or return; - $off += $nr; my $tmp = ''; foreach my $s (@$msgs) { $tmp .= $s->num . ' ' . $s->$field . "\r\n"; } utf8::encode($tmp); do_more($self, $tmp); - # -1 to adjust for implicit increment in long_response - $$i = $nr ? $$i + $nr - 1 : long_response_limit; + $cur = $msgs->[-1]->{num} + 1; }); } } @@ -821,11 +790,11 @@ sub cmd_xrover ($;$) { my $mm = $ng->mm; my $srch = $ng->search; more($self, '224 Overview information follows'); - long_response($self, $beg, $end, sub { - my ($i) = @_; - my $mid = $mm->mid_for($$i) or return; - my $h = search_header_for($srch, $mid, 'references'); - more($self, "$$i $h"); + + long_response($self, sub { + my $h = search_header_for($srch, $beg, 'references'); + more($self, "$beg $h") if defined($h); + $beg++ < $end; }); } @@ -837,10 +806,10 @@ sub over_line ($$) { $smsg->{subject}, $smsg->{from}, PublicInbox::SearchMsg::date($smsg), - '<'.PublicInbox::SearchMsg::mid($smsg).'>', + "<$smsg->{mid}>", $smsg->{references}, - PublicInbox::SearchMsg::bytes($smsg), - PublicInbox::SearchMsg::lines($smsg)); + $smsg->{bytes}, + $smsg->{lines}); utf8::encode($s); $s } @@ -849,10 +818,9 @@ sub cmd_over ($;$) { my ($self, $range) = @_; if ($range && $range =~ /\A<(.+)>\z/) { my ($ng, $n) = mid_lookup($self, $1); - my $smsg = $ng->search->lookup_message($range) or - return '430 No article with that message-id'; + defined $n or return r430; + my $smsg = $ng->search->{over_ro}->get_art($n) or return r430; more($self, '224 Overview information follows (multi-line)'); - $smsg = PublicInbox::SearchMsg->load_doc($smsg->{doc}); # Only set article number column if it's the current group my $self_ng = $self->{ng}; @@ -872,21 +840,16 @@ sub cmd_xover ($;$) { my ($beg, $end) = @$r; more($self, "224 Overview information follows for $beg to $end"); my $srch = $self->{ng}->search; - my $off = 0; - long_response($self, $beg, $end, sub { - my ($i) = @_; - my $res = $srch->query_xover($beg, $end, $off); - my $msgs = $res->{msgs}; + my $cur = $beg; + long_response($self, sub { + my $msgs = $srch->query_xover($cur, $end); my $nr = scalar @$msgs or return; - $off += $nr; # OVERVIEW.FMT more($self, join("\r\n", map { - over_line(PublicInbox::SearchMsg::num($_), $_); + over_line($_->{num}, $_); } @$msgs)); - - # -1 to adjust for implicit increment in long_response - $$i = $nr ? $$i + $nr - 1 : long_response_limit; + $cur = $msgs->[-1]->{num} + 1; }); } @@ -916,7 +879,7 @@ sub more ($$) { sub do_write ($$) { my ($self, $data) = @_; my $done = $self->write($data); - die if $self->{closed}; + return if $self->{closed}; # Do not watch for readability if we have data in the queue, # instead re-enable watching for readability when we can @@ -1003,10 +966,19 @@ sub watch_read { $rv; } +sub not_idle_long ($$) { + my ($self, $now) = @_; + defined(my $fd = $self->{fd}) or return; + my $ary = $EXPMAP->{$fd} or return; + my $exp_at = $ary->[0] + $EXPTIME; + $exp_at > $now; +} + # for graceful shutdown in PublicInbox::Daemon: -sub busy () { - my ($self) = @_; - ($self->{rbuf} ne '' || $self->{long_res} || $self->{write_buf_size}); +sub busy { + my ($self, $now) = @_; + ($self->{rbuf} ne '' || $self->{long_res} || $self->{write_buf_size} || + not_idle_long($self, $now)); } 1;