[PATCH] pop3: reduce memory use while generating the mailbox cache

[PATCH] pop3: reduce memory use while generating the mailbox cache

[Eric Wong]

While the cache itself is relatively compact for 50K messages,
generating it was inefficient due to our schema and Over.pm APIs
being designed for NNTP.  While we won't change our schema for
now, we can choose better DBI APIs to use and limit our ephemeral
memory use.

This amounts to a 60% reduction in memory usage and a 5-10%
speedup against org.kernel.vger.git.0:

	{
		echo 'USER '$(uuidgen)'@org.kernel.vger.git.0'
		echo PASS anonymous
		echo STAT
		echo QUIT
	} | nc $HOST $PORT
---
 lib/PublicInbox/POP3.pm | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/lib/PublicInbox/POP3.pm b/lib/PublicInbox/POP3.pm
index 60eedea7..203c91a6 100644
--- a/lib/PublicInbox/POP3.pm
+++ b/lib/PublicInbox/POP3.pm
@@ -148,12 +148,21 @@ sub _stat_cache ($) {
 	my ($self) = @_;
 	my ($beg, $end) = (($self->{uid_dele} // -1) + 1, $self->{uid_max});
 	PublicInbox::IMAP::uid_clamp($self, \$beg, \$end);
-	my $opt = { limit => PublicInbox::IMAP::UID_SLICE };
-	my $m = $self->{ibx}->over(1)->do_get(<<'', $opt, $beg, $end);
+	my (@cache, $m);
+	my $sth = $self->{ibx}->over(1)->dbh->prepare_cached(<<'', undef, 1);
 SELECT num,ddd FROM over WHERE num >= ? AND num <= ?
 ORDER BY num ASC
 
-	[ map { ($_->{num}, $_->{bytes} + 0, $_->{blob}) } @$m ];
+	$sth->execute($beg, $end);
+	do {
+		$m = $sth->fetchall_arrayref({}, 1000);
+		for my $x (@$m) {
+			PublicInbox::Over::load_from_row($x);
+			push(@cache, $x->{num}, $x->{bytes} + 0, $x->{blob});
+			undef $x; # saves ~1.5M memory w/ 50k messages
+		}
+	} while (scalar(@$m) && ($beg = $cache[-3] + 1));
+	\@cache;
 }
 
 sub cmd_stat {