X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FNNTP.pm;h=56d8e010f96b7d2db8061f100213b1e47f6eaf8e;hb=3d41aa23f35501ca92aab8aa42980fa73f7fa74f;hp=e4e3de4adfd3e17dc12bdacb5713c7d33158a527;hpb=682ecd3ad1eec32876f2705e6c60c7dfb30556aa;p=public-inbox.git
diff --git a/lib/PublicInbox/NNTP.pm b/lib/PublicInbox/NNTP.pm
index e4e3de4a..56d8e010 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',
@@ -39,8 +40,6 @@ 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 $nextq = [];
@@ -61,19 +60,8 @@ sub next_tick () {
sub update_idle_time ($) {
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 +82,11 @@ 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 +99,6 @@ sub new ($$$) {
$self->{rbuf} = '';
$self->watch_read(1);
update_idle_time($self);
- $WEAKEN->{"$nntpd"} = $nntpd;
$expt ||= PublicInbox::EvCleanup::later(*expire_old);
$self;
}
@@ -144,6 +127,7 @@ sub process_line ($$) {
my $res = eval { $req->($self, @args) };
my $err = $@;
if ($err && !$self->{closed}) {
+ local $/ = "\n";
chomp($l);
err($self, 'error from: %s (%s)', $l, $err);
$res = '503 program fault - command not performed';
@@ -255,7 +239,6 @@ sub cmd_listgroup ($;$) {
sub parse_time ($$;$) {
my ($date, $time, $gmt) = @_;
- use Time::Local qw();
my ($hh, $mm, $ss) = unpack('A2A2A2', $time);
if (defined $gmt) {
$gmt =~ /\A(?:UTC|GMT)\z/i or die "GM invalid: $gmt";
@@ -267,15 +250,15 @@ sub parse_time ($$;$) {
($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, $YYYY);
}
}
@@ -304,7 +287,6 @@ sub wildmat2re (;$) {
return $_[0] = qr/.*/ if (!defined $_[0] || $_[0] eq '*');
my %keep;
my $salt = rand;
- use Digest::SHA qw(sha1_hex);
my $tmp = $_[0];
$tmp =~ s#(?header_set('Xref', xref($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>");
}
@@ -481,10 +463,9 @@ 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 $s = eval { $ng->msg_by_mid($mid, \$bytes) } or return $err;
+ $s = Email::Simple->new($s);
my $lines;
if ($set_headers) {
set_nntp_headers($s->header_obj, $ng, $n, $mid);
@@ -512,6 +493,12 @@ sub set_art {
$self->{article} = $art if defined $art && $art =~ /\A\d+\z/;
}
+sub _header ($) {
+ my $hdr = $_[0]->header_obj->as_string;
+ utf8::encode($hdr);
+ $hdr
+}
+
sub cmd_article ($;$) {
my ($self, $art) = @_;
my $r = art_lookup($self, $art, 1);
@@ -519,7 +506,7 @@ sub cmd_article ($;$) {
my ($n, $mid, $s) = @$r;
set_art($self, $art);
more($self, "220 $n <$mid> article retrieved - head and body follow");
- do_more($self, $s->header_obj->as_string);
+ do_more($self, _header($s));
do_more($self, "\r\n");
simple_body_write($self, $s);
}
@@ -531,7 +518,7 @@ sub cmd_head ($;$) {
my ($n, $mid, $s) = @$r;
set_art($self, $art);
more($self, "221 $n <$mid> article retrieved - head follows");
- do_more($self, $s->header_obj->as_string);
+ do_more($self, _header($s));
'.'
}
@@ -707,8 +694,7 @@ sub hdr_xref ($$$) { # optimize XHDR Xref [range] for rtin
sub search_header_for {
my ($srch, $mid, $field) = @_;
- my $smsg = $srch->lookup_message($mid) or return;
- $smsg = PublicInbox::SearchMsg->load_doc($smsg->{doc});
+ my $smsg = $srch->lookup_mail($mid) or return;
$smsg->$field;
}
@@ -738,6 +724,7 @@ sub hdr_searchmsg ($$$$) {
foreach my $s (@$msgs) {
$tmp .= $s->num . ' ' . $s->$field . "\r\n";
}
+ utf8::encode($tmp);
do_more($self, $tmp);
# -1 to adjust for implicit increment in long_response
$$i = $nr ? $$i + $nr - 1 : long_response_limit;
@@ -826,7 +813,7 @@ sub over_line ($$) {
my ($num, $smsg) = @_;
# n.b. field access and procedural calls can be
# 10%-15% faster than OO method calls:
- join("\t", $num,
+ my $s = join("\t", $num,
$smsg->{subject},
$smsg->{from},
PublicInbox::SearchMsg::date($smsg),
@@ -834,16 +821,17 @@ sub over_line ($$) {
$smsg->{references},
PublicInbox::SearchMsg::bytes($smsg),
PublicInbox::SearchMsg::lines($smsg));
+ utf8::encode($s);
+ $s
}
sub cmd_over ($;$) {
my ($self, $range) = @_;
if ($range && $range =~ /\A<(.+)>\z/) {
my ($ng, $n) = mid_lookup($self, $1);
- my $smsg = $ng->search->lookup_message($range) or
+ my $smsg = $ng->search->lookup_mail($range) or
return '430 No article with that message-id';
more($self, '224 Overview information follows (multi-line)');
- $smsg = PublicInbox::SearchMsg->load_doc($smsg->{doc});
# Only set article number column if it's the current group
my $self_ng = $self->{ng};
@@ -896,7 +884,6 @@ sub cmd_xpath ($$) {
sub res ($$) {
my ($self, $line) = @_;
- utf8::encode($line);
do_write($self, $line . "\r\n");
}
@@ -908,7 +895,7 @@ sub more ($$) {
sub do_write ($$) {
my ($self, $data) = @_;
my $done = $self->write($data);
- die if $self->{closed};
+ return if $self->{closed};
# Do not watch for readability if we have data in the queue,
# instead re-enable watching for readability when we can
@@ -931,7 +918,6 @@ use constant MSG_MORE => ($^O eq 'linux') ? 0x8000 : 0;
sub do_more ($$) {
my ($self, $data) = @_;
- utf8::encode($data);
if (MSG_MORE && !$self->{write_buf_size}) {
my $n = send($self->{sock}, $data, MSG_MORE);
if (defined $n) {
@@ -996,10 +982,19 @@ 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;