From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.io/gmane.emacs.gnus.general/40962 Path: main.gmane.org!not-for-mail From: Russ Allbery Newsgroups: gmane.emacs.gnus.general Subject: Checking new mail very slow with current CVS Date: Mon, 24 Dec 2001 22:49:00 -0800 Organization: The Eyrie Sender: owner-ding@hpc.uh.edu Message-ID: NNTP-Posting-Host: coloc-standby.netfonds.no Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Trace: main.gmane.org 1035176436 3706 80.91.224.250 (21 Oct 2002 05:00:36 GMT) X-Complaints-To: usenet@main.gmane.org NNTP-Posting-Date: Mon, 21 Oct 2002 05:00:36 +0000 (UTC) Return-Path: Original-Received: (qmail 7362 invoked from network); 25 Dec 2001 06:52:26 -0000 Original-Received: from malifon.math.uh.edu (mail@129.7.128.13) by mastaler.com with SMTP; 25 Dec 2001 06:52:26 -0000 Original-Received: from sina.hpc.uh.edu ([129.7.128.10] ident=lists) by malifon.math.uh.edu with esmtp (Exim 3.20 #1) id 16IlOy-0002fH-00; Tue, 25 Dec 2001 00:49:28 -0600 Original-Received: by sina.hpc.uh.edu (TLB v0.09a (1.20 tibbs 1996/10/09 22:03:07)); Tue, 25 Dec 2001 00:49:19 -0600 (CST) Original-Received: from sclp3.sclp.com (qmailr@sclp3.sclp.com [209.196.61.66]) by sina.hpc.uh.edu (8.9.3/8.9.3) with SMTP id AAA08275 for ; Tue, 25 Dec 2001 00:49:07 -0600 (CST) Original-Received: (qmail 7323 invoked by alias); 25 Dec 2001 06:49:07 -0000 Original-Received: (qmail 7318 invoked from network); 25 Dec 2001 06:49:07 -0000 Original-Received: from windlord.stanford.edu (171.64.13.23) by gnus.org with SMTP; 25 Dec 2001 06:49:07 -0000 Original-Received: (qmail 19207 invoked by uid 50); 25 Dec 2001 06:49:00 -0000 Original-To: ding@gnus.org Original-Lines: 67 User-Agent: Gnus/5.090004 (Oort Gnus v0.04) XEmacs/21.1 (Channel Islands, sparc-sun-solaris2.6) Precedence: list X-Majordomo: 1.94.jlt7 Xref: main.gmane.org gmane.emacs.gnus.general:40962 X-Report-Spam: http://spam.gmane.org/gmane.emacs.gnus.general:40962 I just upgraded to the current CVS version of Gnus so that I could start playing with nndiary, and I immediately noticed that checking new mail with "g" is *much* slower than it was under 5.8.7. Does anyone have any idea what's changed and if there's anything I can do about it? It's slow enough to be fairly frustrating. Here's the tail end of the message log; I've marked the two places where there's a noticeable pause: nntp read: 22k nntp read: 22k nntp read: 22k nntp read: 22k nntp read: 22k nntp read: 22k nntp read: 22k nntp read: 22k nntp read: 23k ===> Reading active file via nnml... Opening nnml server... Opening nnml server...done nnml: Reading incoming mail from maildir... nnml: Reading incoming mail (no new mail)...done Reading active file via nnml... Reading active file via nnml...done Reading active file via nndiary... Reading active file via nndiary... Reading active file via nndiary...done Reading active file from sent via nnml... Opening nnml server on sent... Opening nnml server on sent...done Reading active file from sent via nnml... Reading active file from sent via nnml...done Checking new news... ===> Checking new news...done The first pause isn't too bad, but the second pause is really noticeable, and XEmacs does four or six garbage collection passes in the middle of it. Setting gnus-verbose-backends to 10 doesn't produce any additional messages, and setting nnml-marks-is-evil to t makes no difference at all. Setting gnus-debug-on-quit and pressing C-g in the middle of the long pause gave the following backtrace: Signaling: (quit) nnmail-group-pathname("work.watch.registry" "~/Mail/") nnml-possibly-change-directory("work.watch.registry" "") nnml-request-group("work.watch.registry" "" t) gnus-activate-group("nnml:work.watch.registry" nil t) gnus-get-unread-articles(nil) gnus-group-get-new-news(nil) call-interactively(gnus-group-get-new-news) My *guess* from that is that for some reason activating every nnml group is taking quite a long time and that 5.8.7 either didn't do that or did it much more quickly. I have several hundred nnml groups. A search of the mailing list traffic reveals nothing except a few messages about marks, which seems unrelated given that I see the same behavior regardless of whether marks are enabled. Does anyone have any ideas? -- Russ Allbery (rra@stanford.edu)