#!/usr/bin/perl -w
-# Copyright (C) 2015, all contributors <meta@public-inbox.org>
+# Copyright (C) 2015 all contributors <meta@public-inbox.org>
# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
#
# Incremental (or one-shot) importer of a slrnpull news spool
use Email::Filter;
use Email::LocalDelivery;
sub usage { "Usage:\n".join('',grep(/\t/, `head -n 10 $0`)) }
+my $exit = 0;
+my $sighandler = sub { $exit = 1 };
+$SIG{INT} = $sighandler;
+$SIG{TERM} = $sighandler;
my $spool = shift @ARGV or die usage();
my $recipient = $ENV{ORIGINAL_RECIPIENT};
defined $recipient or die usage();
-my @mda = qw(public-inbox-mda);
my $config = PublicInbox::Config->new;
my $cfg = $config->lookup($recipient);
defined $cfg or exit(1);
-use Data::Dumper; print STDERR Dumper($cfg);
+my @mda;
+if ($ENV{'FILTER'}) {
+ @mda = qw(public-inbox-mda);
+} else {
+ @mda = (qw(ssoma-mda -1), $cfg->{mainrepo});
+}
-sub get_min {
+sub key {
my ($cfg) = @_;
- $cfg->{importslrnspoolstate} || 0;
+ "publicinbox.$cfg->{listname}.importslrnspoolstate";
+}
+
+sub get_min {
+ my $f = PublicInbox::Config->default_file;
+ my @cmd = (qw/git config/, "--file=$f", key($cfg));
+ use IPC::Run qw/run/;
+
+ my $in = '';
+ my $out = '';
+ unless (run(\@cmd, \$in, \$out)) {
+ $out = 0;
+ }
+ int($out);
}
sub set_min {
my ($cfg, $num) = @_;
my $f = PublicInbox::Config->default_file;
- my @cmd = (qw/git config/, "--file=$f",
- "publicinbox.$cfg->{listname}.importslrnspoolstate", $num);
+ my @cmd = (qw/git config/, "--file=$f", key($cfg), $num);
system(@cmd) == 0 or die join(' ', @cmd). " failed: $?\n";
}
my $max_gap = 10000;
my $max = $n + $max_gap;
-for (; $n < $max; $n++) {
+for (; $exit == 0 && $n < $max; $n++) {
my $fn = "$spool/$n";
print STDERR $fn, "\n";
open(my $fh, '<', $fn) or next;
$max = $n + $max_gap;
- my $f = Email::Filter->new(data => eval { local $/; <$fh> });
- my $s = $f->simple;
- # gmane rewrites Received headers, which increases spamminess
- my @h = $s->header('Original-Received');
- if (@h) {
- $s->header_set('Received', @h);
- $s->header_set('Original-Received');
- }
+ # prevent process growth by forking a new process for each message
+ my $pid = fork;
+ die "failed to fork: $!\n" unless defined $pid;
- # triggers for the SA HEADER_SPAM rule
- foreach my $drop (qw(Approved)) { $s->header_set($drop) }
+ if ($pid == 0) {
+ my $f = Email::Filter->new(data => eval { local $/; <$fh> });
+ close $fh;
+ $fh = undef;
+ my $s = $f->simple;
- # appears to be an old gmane bug:
- $s->header_set('connect()');
+ # gmane rewrites Received headers, which increases spamminess
+ # Some older archives set Original-To
+ foreach my $x (qw(Received To)) {
+ my @h = $s->header("Original-$x");
+ if (@h) {
+ $s->header_set($x, @h);
+ $s->header_set("Original-$x");
+ }
+ }
- $f->exit(0);
- $f->pipe(@mda);
+ # triggers for the SA HEADER_SPAM rule
+ foreach my $drop (qw(Approved)) { $s->header_set($drop) }
+
+ # appears to be an old gmane bug:
+ $s->header_set('connect()');
+
+ $f->exit(0);
+ $f->pipe(@mda);
+ exit 0;
+ } else {
+ close $fh;
+ waitpid($pid, 0);
+ die "error: $?\n" if $?;
+ }
$ok = $n + 1;
set_min($cfg, $ok);
}