pop3: reduce memory use while generating the mailbox cache

[public-inbox.git] / scripts / slrnspool2maildir

diff --git a/scripts/slrnspool2maildir b/scripts/slrnspool2maildir
index fb7a6f253aa8a8659220944c156f8192993117b2..8e2ba08ac523d46e0322aadb1610e57e4c59af87 100755 (executable)
--- a/scripts/slrnspool2maildir
+++ b/scripts/slrnspool2maildir
@@ -1,6 +1,6 @@
 #!/usr/bin/perl -w
-# 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) 2013-2021 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 #
 # One-off script to convert an slrnpull news spool to Maildir
 =begin usage
@@ -10,25 +10,31 @@ use strict;
 use warnings;
 use Email::Filter;
 use Email::LocalDelivery;
-sub usage { "Usage:\n".join('',grep(/\t/, `head -n 24 $0`)) }
+use File::Glob qw(bsd_glob GLOB_NOSORT);
+sub usage { "Usage:\n".join('',grep(/\t/, `head -n 12 $0`)) }
 my $spool = shift @ARGV or die usage();
 my $dir = shift @ARGV or die usage();
 -d $dir or die "$dir is not a directory\n";
 $dir .= '/' unless $dir =~ m!/\z!;
 foreach my $sub (qw(cur new tmp)) {
-       -d "$dir/$sub" or mkdir $sub or die "mkdir $dir/$sub failed: $!\n";
+       my $nd = "$dir/$sub";
+       -d $nd and next;
+       mkdir $nd or die "mkdir $nd failed: $!\n";
 }
 
-foreach my $n (grep(/\d+\z/, glob("$spool/*"))) {
+foreach my $n (grep(/\d+\z/, bsd_glob("$spool/*", GLOB_NOSORT))) {
        if (open my $fh, '<', $n) {
-               my $f = Email::Filter->new(data => eval { local $/; <$fh> });
+               my $f = Email::Filter->new(data => do { 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');
+               # 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");
+                       }
                }
 
                # triggers for the SA HEADER_SPAM rule