From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on dcvr.yhbt.net X-Spam-Level: X-Spam-ASN: X-Spam-Status: No, score=-4.0 required=3.0 tests=ALL_TRUSTED,BAYES_00 shortcircuit=no autolearn=ham autolearn_force=no version=3.4.2 Received: from localhost (dcvr.yhbt.net [127.0.0.1]) by dcvr.yhbt.net (Postfix) with ESMTP id 7B8611F803; Thu, 10 Jan 2019 09:19:55 +0000 (UTC) Date: Thu, 10 Jan 2019 09:19:55 +0000 From: Eric Wong To: meta@public-inbox.org Subject: [REJECT] watchmaildir: spam checking and learning improvements Message-ID: <20190110091955.7q6px5cwmiisjo6v@dcvr> References: <20190110090251.18736-1-e@80x24.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20190110090251.18736-1-e@80x24.org> List-Id: Eric Wong wrote: > It turns out spam training ("spamc -L spam") was never implemented > for -watch. So implement that, and take advantage of an > existing feature to pass FDs directly to spamc(1) instead of > reconverting the MIME object into a string. Will keep that... > While we're at it, the spam-checking call is now only performed > once per message when a Maildir supports multiple inboxes. And that. > This changes PublicInbox::Import::add (and V2Writable > equivalent) API to no longer perform spam-checking. But nope, we need to do spam-checking inside ->add; because we don't want to waste cycles re-spam-checking messages which are already in our archives on -watch restarts. And the V2Writable->add spam-checking is in the wrong place.