scripts/import_gmane_spool: misc updates

We may promote this to be a real script, since public-inbox-mda
is idempotent.

diff --git a/scripts/import_gmane_spool b/scripts/import_gmane_spool
index b102a21d5bc81868590b46f8d524df4f0060f755..e153ba4c4d6bec729dd38fa9ee5ccb1a7936eaed 100755 (executable)
--- a/scripts/import_gmane_spool
+++ b/scripts/import_gmane_spool
@@ -2,27 +2,40 @@
 # Copyright (C) 2013, Eric Wong <normalperson@yhbt.net> and all contributors
 # License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
 #
 # Copyright (C) 2013, Eric Wong <normalperson@yhbt.net> and all contributors
 # License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
 #

-# One-off script to convert an slrnpull news spool from gmane
+# One-off script to convert an slrnpull news spool from gmane, usage:
+=begin usage
+       mkdir -p $HOME/.public-inbox
+       MAINREPO=/path/to/your/repo.git
+       export RECIPIENT='list@example.com'
+       git init --bare $MAINREPO
+       export GIT_CONFIG=$HOME/.public-inbox/config
+       git config publicinbox.$LISTNAME.address $RECIPIENT
+       git config publicinbox.$LISTNAME.mainrepo $MAINREPO
+       unset GIT_CONFIG
+       ./import_gmane_spool SLRNPULL_ROOT/news/foo/bar
+=cut

 use strict;
 use warnings;
 use Parallel::ForkManager;
 use Email::Simple;
 use PublicInbox::Filter;
 use strict;
 use warnings;
 use Parallel::ForkManager;
 use Email::Simple;
 use PublicInbox::Filter;

+use PublicInbox::Config;

 use IPC::Run qw(run);
 use IPC::Run qw(run);

-my $usage = "$0 SLRNPULL_ROOT/news/foo/bar MAIN_REPO FAIL_REPO";
-my $spool = shift @ARGV or die "Usage: $usage\n";
-my $main_repo = shift @ARGV or die "Usage: $usage\n";
-my $fail_repo = shift @ARGV or die "Usage: $usage\n";
+sub usage { "Usage:\n".join("",grep(/\t/, `head -n 24 $0`)) }
+my $spool = shift @ARGV or die usage();

 my $nproc = `nproc 2>/dev/null` || 4;
 my $pm = Parallel::ForkManager->new($nproc);
 my $nproc = `nproc 2>/dev/null` || 4;
 my $pm = Parallel::ForkManager->new($nproc);

-my @args = ('public-inbox-mda', $main_repo, $fail_repo);
+defined $ENV{RECIPIENT} or die usage();
+my @args = ('public-inbox-mda');

 
 

-foreach my $n (<$spool/*>) {
+foreach my $n (glob("$spool/*")) {

        $n =~ m{/\d+\z} or next;
        $pm->start and next;
        if (open my $fh, '<', $n) {
        $n =~ m{/\d+\z} or next;
        $pm->start and next;
        if (open my $fh, '<', $n) {

-               local $/;
-               my $s = Email::Simple->new(<$fh>);
+               my $s = eval {
+                       local $/;
+                       Email::Simple->new(<$fh>);
+               };

 
                # gmane rewrites Received headers, which increases spamminess
                my @h = $s->header("Original-Received");
 
                # gmane rewrites Received headers, which increases spamminess
                my @h = $s->header("Original-Received");