Gnus fails to show some email messages

Gnus fails to show some email messages

[Hrvoje Niksic]

Gnus v5.6.2; nntp 5.0; nnml 1.0; nnmh 1.0; nndraft 1.0; nnfolder 1.0
XEmacs 21.0 "Myotonic" [Lucid] (sparc-sun-solaris2.5.1) of Mon Mar 23 1998 on jagor.srce.hr
200 CARNet.hr InterNetNews NNRP server INN 1.4unoff4 05-Mar-96 ready (posting ok).


I have noticed a bug in Gnus: it fails to show some email messages.
That is the second most serious bug a mail reader can exhibit (the
first most serious bug is to delete or corrupt the message without
showing it).  For instance, ~/Mail/xemacs.beta/20087 is present on
disk.  It is currently the last message in the `xemacs.beta' group,
and its NOV entry looks like this:

20087	Re: [PATCH MS-Windows] Major display glitches assault 	"Kirill M. Katsnelson" <kkm@kis.ru>	Thu, 26 Mar 1998 15:21:48 +0300	<3.0.5.32.19980326152148.00a9cd60@kis.ru>	<3.0.5.32.19980322183244.00a82180@kis.ru> <Your message of "Fri, 20 Mar 1998 14:41:33 PST."             <002201bd5451$5595b280$4432a8c0@Raj_D.squaresoft.com> <199803202354.AA217888095@mina.sr.hp.com> <3.0.5.32.19980322183244.00a82180@kis.ru>	8954	163	Xref: jagor.srce.hr xemacs.beta:20087	

The `xemacs.beta' information in `.newsrc.eld' looks like this:

("nnml:xemacs.beta" 2 ((1 . 20087)) ((reply 19020 19053 (19065 . 19066) 19077 19082 19089 19124 19130 19139 19148 (19159 . 19160) 19166 (19172 . 19173) 19179 19181 19189 19195 19198 19202 19235 19238 19258 19264 19271 19275 (19283 . 19284) 19294 19303 19316 19322 19327 19330 19350 19410 19452 19483 19485 19488 19492 19561 19568 19588 19606 19612 19624 19634 19654 19662 19701 19703 19705 19739 19751 19776 19786 19792 19800 (19807 . 19808) 19814 19822 19825 19853 19857 19862 19867 19924 19943 19965 19974 19980 20005 20031 20044)) (nnml "") ((to-address . "XEmacs Developers <xemacs-beta@xemacs.org>") (total-expire . t) (expiry-wait . 14)))


However, that message is not shown when I enter the group with `RET',
`C-u RET', `1 RET' or anything else.  When I press `1 RET', Gnus tries
to enter the group, but apparently gives up, because it just runs the
expiry process and switches me back to Group buffer.  Typing `2 RET'
generates the summary buffer with only one message, the second-to-last
one.

When I enter the group with `C-u RET', typing `j TAB' does not offer
20087 as a choice.  Typing `j 20087 RET' ends up in an apparently
infite loop, debug-on-quit giving the backtrace appended below.


I have noticed this happening since some time into Quassia, and it's
very serious.  It happens on some messages, and not on others.  I have
no clue how to debug it, and I may be forced to downgrade to Gnus 5.5
because of this bug.

If it happens to anyone else, I'd like to know about it.


The C-g backtrace follows:

Signaling: (quit)
  (mail-header-references prev)
)
  (gnus-parent-id (mail-header-references prev))
)
  (setq last-id id id (gnus-parent-id (mail-header-references prev)))
)
  (while (and id (setq prev ...)) (setq last-id id id (gnus-parent-id ...)))
)
  (let (last-id prev) (while (and id ...) (setq last-id id id ...)) last-id)
)
  gnus-root-id("<3.0.5.32.19980326152148.00a9cd60@kis.ru>")
  (setq last-id (gnus-root-id id))
)
  (let ((dep gnus-newsgroup-dependencies) headers thread last-id) (setq last-id (gnus-root-id id)) (setq headers (list ... ...)) (let (... sub) (while threads ... ... ...) (if thread ... ...) (when thread ...)))
)
  gnus-remove-thread("<3.0.5.32.19980326152148.00a9cd60@kis.ru>")
  (setq thread (gnus-remove-thread id))
)
  (if (not gnus-show-threads) (setq thread (list ...)) (setq thread (gnus-remove-thread id)))
)
  (let ((buffer-read-only nil) old-pos current thread data) (if (not gnus-show-threads) (setq thread ...) (setq thread ...)) (setq old-pos (gnus-point-at-bol)) (setq current (save-excursion ...)) (when (stringp ...) (let ... ... ... ...)) (let (threads) (let ... ... ... ...) (gnus-data-enter-list current data ...) (setq gnus-newsgroup-threads ...)))
)
  gnus-rebuild-thread("<3.0.5.32.19980326152148.00a9cd60@kis.ru>")
  (progn (when (and ... old-header) (when ... ... ...)) (when old-header (mail-header-set-number header ...)) (setq gnus-newsgroup-sparse (delq ... gnus-newsgroup-sparse)) (setq gnus-newsgroup-ancient (delq number gnus-newsgroup-ancient)) (push number gnus-newsgroup-limit) (gnus-rebuild-thread (mail-header-id header)) (gnus-summary-goto-subject number nil t))
)
  (if header (progn (when ... ...) (when old-header ...) (setq gnus-newsgroup-sparse ...) (setq gnus-newsgroup-ancient ...) (push number gnus-newsgroup-limit) (gnus-rebuild-thread ...) (gnus-summary-goto-subject number nil t)))
)
  (when header (when (and ... old-header) (when ... ... ...)) (when old-header (mail-header-set-number header ...)) (setq gnus-newsgroup-sparse (delq ... gnus-newsgroup-sparse)) (setq gnus-newsgroup-ancient (delq number gnus-newsgroup-ancient)) (push number gnus-newsgroup-limit) (gnus-rebuild-thread (mail-header-id header)) (gnus-summary-goto-subject number nil t))
)
  (let ((header ...) (number ...) pos d) (when header (when ... ...) (when old-header ...) (setq gnus-newsgroup-sparse ...) (setq gnus-newsgroup-ancient ...) (push number gnus-newsgroup-limit) (gnus-rebuild-thread ...) (gnus-summary-goto-subject number nil t)) (when (and ... ...) (when ... ...) (when ... ...) (setq gnus-newsgroup-unselected ...)) (and header (mail-header-number header)))
)
  gnus-summary-insert-subject(20087 nil t)
  (and (not data) force (gnus-summary-insert-subject article (and ... force) t) (setq data (gnus-data-find article)))
)
  (let ((b ...) (data ...)) (and (not data) force (gnus-summary-insert-subject article ... t) (setq data ...)) (goto-char b) (if (not data) (progn ... nil) (goto-char ...) article))
)
  gnus-summary-goto-subject(20087 t)
  (if (gnus-summary-goto-subject article force) (gnus-summary-display-article article all-headers) (gnus-message 4 "Couldn't go to article %s" article) nil)
)
  (if (and (stringp article) (string-match "@" article)) (gnus-summary-refer-article article) (when (stringp article) (setq article ...)) (if (gnus-summary-goto-subject article force) (gnus-summary-display-article article all-headers) (gnus-message 4 "Couldn't go to article %s" article) nil))
)
  (prog1 (if (and ... ...) (gnus-summary-refer-article article) (when ... ...) (if ... ... ... nil)) (gnus-summary-position-point))
)
  (lambda (article &optional all-headers force) "Fetch ARTICLE (article number or Message-ID) and display it if it exists.\nIf ALL-HEADERS is non-nil, no header lines are hidden." (interactive (list ... current-prefix-arg t)) (prog1 (if ... ... ... ...) (gnus-summary-position-point)))("20087" nil t)
  call-interactively(gnus-summary-goto-article)


Here is another backtrace:

Signaling: (quit)
  (while (and id (setq prev ...)) (setq last-id id id (gnus-parent-id ...)))
)
  (let (last-id prev) (while (and id ...) (setq last-id id id ...)) last-id)
)
  gnus-root-id("<3.0.5.32.19980326152148.00a9cd60@kis.ru>")
  (setq last-id (gnus-root-id id))
)
  (let ((dep gnus-newsgroup-dependencies) headers thread last-id) (setq last-id (gnus-root-id id)) (setq headers (list ... ...)) (let (... sub) (while threads ... ... ...) (if thread ... ...) (when thread ...)))
)
  gnus-remove-thread("<3.0.5.32.19980326152148.00a9cd60@kis.ru>")
  (setq thread (gnus-remove-thread id))
)
  (if (not gnus-show-threads) (setq thread (list ...)) (setq thread (gnus-remove-thread id)))
)
  (let ((buffer-read-only nil) old-pos current thread data) (if (not gnus-show-threads) (setq thread ...) (setq thread ...)) (setq old-pos (gnus-point-at-bol)) (setq current (save-excursion ...)) (when (stringp ...) (let ... ... ... ...)) (let (threads) (let ... ... ... ...) (gnus-data-enter-list current data ...) (setq gnus-newsgroup-threads ...)))
)
  gnus-rebuild-thread("<3.0.5.32.19980326152148.00a9cd60@kis.ru>")
  (progn (when (and ... old-header) (when ... ... ...)) (when old-header (mail-header-set-number header ...)) (setq gnus-newsgroup-sparse (delq ... gnus-newsgroup-sparse)) (setq gnus-newsgroup-ancient (delq number gnus-newsgroup-ancient)) (push number gnus-newsgroup-limit) (gnus-rebuild-thread (mail-header-id header)) (gnus-summary-goto-subject number nil t))
)
  (if header (progn (when ... ...) (when old-header ...) (setq gnus-newsgroup-sparse ...) (setq gnus-newsgroup-ancient ...) (push number gnus-newsgroup-limit) (gnus-rebuild-thread ...) (gnus-summary-goto-subject number nil t)))
)
  (when header (when (and ... old-header) (when ... ... ...)) (when old-header (mail-header-set-number header ...)) (setq gnus-newsgroup-sparse (delq ... gnus-newsgroup-sparse)) (setq gnus-newsgroup-ancient (delq number gnus-newsgroup-ancient)) (push number gnus-newsgroup-limit) (gnus-rebuild-thread (mail-header-id header)) (gnus-summary-goto-subject number nil t))
)
  (let ((header ...) (number ...) pos d) (when header (when ... ...) (when old-header ...) (setq gnus-newsgroup-sparse ...) (setq gnus-newsgroup-ancient ...) (push number gnus-newsgroup-limit) (gnus-rebuild-thread ...) (gnus-summary-goto-subject number nil t)) (when (and ... ...) (when ... ...) (when ... ...) (setq gnus-newsgroup-unselected ...)) (and header (mail-header-number header)))
)
  gnus-summary-insert-subject(20087 nil t)
  (and (not data) force (gnus-summary-insert-subject article (and ... force) t) (setq data (gnus-data-find article)))
)
  (let ((b ...) (data ...)) (and (not data) force (gnus-summary-insert-subject article ... t) (setq data ...)) (goto-char b) (if (not data) (progn ... nil) (goto-char ...) article))
)
  gnus-summary-goto-subject(20087 t)
  (if (gnus-summary-goto-subject article force) (gnus-summary-display-article article all-headers) (gnus-message 4 "Couldn't go to article %s" article) nil)
)
  (if (and (stringp article) (string-match "@" article)) (gnus-summary-refer-article article) (when (stringp article) (setq article ...)) (if (gnus-summary-goto-subject article force) (gnus-summary-display-article article all-headers) (gnus-message 4 "Couldn't go to article %s" article) nil))
)
  (prog1 (if (and ... ...) (gnus-summary-refer-article article) (when ... ...) (if ... ... ... nil)) (gnus-summary-position-point))
)
  (lambda (article &optional all-headers force) "Fetch ARTICLE (article number or Message-ID) and display it if it exists.\nIf ALL-HEADERS is non-nil, no header lines are hidden." (interactive (list ... current-prefix-arg t)) (prog1 (if ... ... ... ...) (gnus-summary-position-point)))("20087" nil t)
  call-interactively(gnus-summary-goto-article)


------------------ Environment follows ------------------

(setq gnus-message-archive-group
      '((concat
	 (if
	     (message-news-p)
	     "sent-news" "sent-mail")
	 (format-time-string "-%b-%Y"
			     (current-time)))))
(setq gnus-secondary-select-methods
      '((nnml "")))
(setq gnus-large-newsgroup 2000)
(setq gnus-use-long-file-name t)
(setq gnus-kill-files-directory "~/.scores/")
(setq gnus-use-trees t)
(setq gnus-novice-user nil)
(setq gnus-expert-user t)
(setq gnus-group-change-level-function
      'gnus-topic-change-level)
(setq gnus-visual t)
(setq gnus-article-display-hook
      '(latin2-translate-receive gnus-article-hide-headers-if-wanted gnus-article-hide-boring-headers gnus-article-treat-overstrike gnus-article-maybe-highlight gnus-article-display-x-face gnus-article-hide-pgp gnus-article-highlight gnus-article-emphasize))
(setq gnus-build-sparse-threads
      'some)
(setq gnus-summary-goto-unread nil)
(setq gnus-view-pseudo-asynchronously t)
(setq gnus-thread-sort-functions
      '(gnus-thread-sort-by-number gnus-thread-sort-by-total-score))
(setq gnus-summary-mode-hook
      '(gnus-xmas-switch-horizontal-scrollbar-off gnus-xmas-setup-summary-toolbar gnus-xmas-summary-menu-add))
(setq gnus-summary-exit-hook
      '(gnus-topic-update-topic))
(setq gnus-group-mode-hook
      '(gnus-topic-mode gnus-xmas-setup-group-toolbar gnus-xmas-group-menu-add))
(setq gnus-group-catchup-group-hook
      '(gnus-topic-update-topic))
(setq gnus-exit-gnus-hook
      '(#<compiled-function nil "...(18)" [frame-live-p toolbar-news-frame frame-list delete-frame nil] 2>))
(setq gnus-visible-headers "^From:\\|^Newsgroups:\\|^Subject:\\|^Date:\\|^Followup-To:\\|^Reply-To:\\|^Organization:\\|^Summary:\\|^Keywords:\\|^To:\\|^Cc:\\|^Posted-To:\\|^Mail-Copies-To:\\|^Apparently-To:\\|^Gnus-Warning:\\|^Resent-From:\\|^X-Newsreader:\\|^X-Mailer:\\|^X-URL:\\|^X-WWW\\|^X-[yz]")
(setq gnus-article-x-face-command
      'gnus-xmas-article-display-xface)
(setq gnus-prompt-before-saving t)
(setq gnus-default-article-saver
      'gnus-summary-save-in-file)
(setq gnus-article-mode-hook
      '(gnus-xmas-article-menu-add turn-off-scroll-in-place))
(setq gnus-save-newsrc-file nil)
(setq gnus-save-killed-list nil)
(setq gnus-subscribe-newsgroup-method
      'gnus-subscribe-hierarchically)
(setq gnus-subscribe-hierarchical-interactive t)
(setq gnus-check-bogus-groups-hook
      '(gnus-topic-clean-alist))
(setq gnus-asynchronous nil)
(setq gnus-buffer-configuration
      '((article
	 (vertical 1.0
		   (horizontal 0.25
			       (summary 0.75 point)
			       (tree 1.0))
		   (article 1.0)))
	(group
	 (vertical 1.0
		   (group 1.0 point)
		   (if gnus-carpal
		       '(group-carpal 4))))
	(summary
	 (vertical 1.0
		   (summary 1.0 point)
		   (if gnus-carpal
		       '(summary-carpal 4))))
	(server
	 (vertical 1.0
		   (server 1.0 point)
		   (if gnus-carpal
		       '(server-carpal 2))))
	(browse
	 (vertical 1.0
		   (browse 1.0 point)
		   (if gnus-carpal
		       '(browse-carpal 2))))
	(message
	 (vertical 1.0
		   (message 1.0 point)))
	(pick
	 (vertical 1.0
		   (article 1.0 point)))
	(info
	 (vertical 1.0
		   (info 1.0 point)))
	(summary-faq
	 (vertical 1.0
		   (summary 0.25)
		   (faq 1.0 point)))
	(edit-article
	 (vertical 1.0
		   (article 1.0 point)))
	(edit-form
	 (vertical 1.0
		   (group 0.5)
		   (edit-form 1.0 point)))
	(edit-score
	 (vertical 1.0
		   (summary 0.25)
		   (edit-score 1.0 point)))
	(post
	 (vertical 1.0
		   (post 1.0 point)))
	(reply
	 (vertical 1.0
		   (article-copy 0.5)
		   (message 1.0 point)))
	(forward
	 (vertical 1.0
		   (message 1.0 point)))
	(reply-yank
	 (vertical 1.0
		   (message 1.0 point)))
	(mail-bounce
	 (vertical 1.0
		   (article 0.5)
		   (message 1.0 point)))
	(pipe
	 (vertical 1.0
		   (summary 0.25 point)
		   (if gnus-carpal
		       '(summary-carpal 4))
		   ("*Shell Command Output*" 1.0)))
	(bug
	 (vertical 1.0
		   ("*Gnus Help Bug*" 0.5)
		   ("*Gnus Bug*" 1.0 point)))
	(score-trace
	 (vertical 1.0
		   (summary 0.5 point)
		   ("*Score Trace*" 1.0)))
	(score-words
	 (vertical 1.0
		   (summary 0.5 point)
		   ("*Score Words*" 1.0)))
	(category
	 (vertical 1.0
		   (category 1.0)))
	(compose-bounce
	 (vertical 1.0
		   (article 0.5)
		   (message 1.0 point)))))
(setq gnus-topic-mode-hook
      '(gnus-xmas-topic-menu-add))
(setq nnmail-split-methods
      'nnmail-split-fancy)
(setq nnmail-use-long-file-names t)
(setq nnmail-expiry-wait
      'immediate)
(setq nnmail-spool-file
      '("/var/mail/hniksic" "~/Mail/incoming"))
(setq nnmail-split-fancy
      '(|
	("mailing-list" "wget" "wget.list")
	("x-mailing-list" "xemacs-beta-discuss" "xemacs.beta.discuss")
	("x-mailing-list" "xemacs-beta" "xemacs.beta")
	("x-mailing-list" "xemacs-admin" "xemacs.admin")
	("x-mailing-list" "xemacs-announce" "xemacs.announce")
	(from "amarok" "amarok")
	("x-mailing-list" "amarok" "amarok")
	("sender" "owner-xemacs-nt" "xemacs.nt")
	("sender" "hr-translation" "hr.translation")
	("x-mailing-list" "emacs-custom" "custom")
	("sender" "owner-mp3" "mp3")
	(to "gnu-prog" "gnu.prog")
	(to "xemacs-developers" "xemacs.developers")
	("sender" "owner-ding-announce" "ding.announce")
	("sender" "owner-gnome" "gnome")
	("x-mailing-list" "gnome-list" "gnome")
	("sender" "kvake@bagan.srce.hr" "kvake.discuss")
	("sender" "owner-ansible" "ansible")
	("x-resent-by" "info-gnu-prog-request@prep" "gnu.prog")
	("received" "ding@ifi" "ding.discuss")
	("sender" "owner-ding" "ding.discuss")
	("from" "bugs@gnus.org" "gnus-bugfixing")
	("to" "bug-wget" "wget.reports")
	("subject" "wget\\|w?fetch\\|geturl" "wget.reports")
	("from" mail "mailer-daemon")
	("from" "majordomo@pc-hrvoje.srce.hr" "majordomo")
	("from" "owner-mp3@pc-hrvoje" "majordomo")
	(to "paketi" "paketi")
	(from "wing" "private")
	("to\\|cc" "hniksic\\|niksic" "private")
	"spam"))
(setq message-elide-elipsis "[...]\n")
(setq message-use-followup-to t)
(setq message-post-method
      '(nnspool ""))
(setq message-setup-hook
      '(latin2-translate-receive))
(setq message-mode-hook
      '(message-xmas-maybe-fontify
	(lambda nil
	  (interactive)
	  (local-set-key [
			 (meta tab)
			 ]
			 'lisp-complete-symbol)
	  (local-set-key "\030\013"
			 'message-kill-to-signature)
	  (local-set-key "\003w"
			 'insert-wget-url))))
(setq message-header-setup-hook
      '(gnus-inews-insert-archive-gcc gnus-inews-insert-gcc))
(setq message-signature
      'sigs-random-signature)
(setq message-default-headers "X-Attribution: Hrvoje\nX-Face: Mie8:rOV<\\c/~z{s.X4A{!?vY7{drJ([U]0O=W/<W*SMo/Mv:58:*_y~ki>xDi&N7XG\n        KV^$k0m3Oe/)'e%3=$PCR&3ITUXH,cK>]bci&<qQ>Ff%x_>1`T(+M2Gg/fgndU%k*ft\n        [(7._6e0n-V%|%'[c|q:;}td$#INd+;?!-V=c8Pqf}3J\n")
(setq message-default-news-headers "Mail-Copies-To: never\n")



-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
"Psychos _do not_ explode when sunlight hits them."

Re: Gnus fails to show some email messages

[Stefanie Teufel]

On Thu, Mar 26, 1998 at 01:53:55PM +0100, Hrvoje Niksic wrote:

> 
> I have noticed this happening since some time into Quassia, and it's
> very serious.  It happens on some messages, and not on others.  I have
> no clue how to debug it, and I may be forced to downgrade to Gnus 5.5
> because of this bug.
> 
> If it happens to anyone else, I'd like to know about it.

It happens to me from time to time. I never mentioned it before,
because I thought it's my fault and I felt to embarrassed to admit my
stupidity. Sorry, but I cannot reproduce the failure at the moment due
to the lack of a message that triggers this behaviour in gnus.

Stefanie

Re: Gnus fails to show some email messages

[Hrvoje Niksic]

Stefanie Teufel <s.teufel@ndh.net> writes:

> On Thu, Mar 26, 1998 at 01:53:55PM +0100, Hrvoje Niksic wrote:
> 
> > 
> > I have noticed this happening since some time into Quassia, and it's
> > very serious.  It happens on some messages, and not on others.  I have
> > no clue how to debug it, and I may be forced to downgrade to Gnus 5.5
> > because of this bug.
> > 
> > If it happens to anyone else, I'd like to know about it.
> 
> It happens to me from time to time. I never mentioned it before,
> because I thought it's my fault and I felt to embarrassed to admit

I figured as much, thanks.  I have always attributed this to some kind 
of mistake, a random glitch that would get cured over time.  When it
happened a sufficient number of times, I got scared -- I have no idea
how many emails I missed.  :-(

Please everyone, check your mails for signs of this bug.

-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
A radioactive cat has eighteen half-lives.

Re: Gnus fails to show some email messages

[Bruce Stephens]

Hrvoje Niksic <hniksic@srce.hr> writes:

> Please everyone, check your mails for signs of this bug.

So the bug symptom is that Gnus fails to display (or notice in the
summary line?) mail which exists on disk?  But the mail moving bit is
working, as far as you can tell?

I'm just trying to clarify, because I haven't received a mail at work
today, which I had sent yesterday from home.  (Probably, it's still
sitting in the queue on my home machine or something.)  A grep through
the mail directories suggests that it's not there at all, so
presumably this isn't the same bug.

(PS, just played around with the tm stuff, and it's pretty good.  Gnus
can still do better, but as things are, I think Gnus+tm is better than
exmh at MIME email.)

Re: Gnus fails to show some email messages

[Hrvoje Niksic]

Bruce Stephens <b.stephens@isode.com> writes:

> Hrvoje Niksic <hniksic@srce.hr> writes:
> 
> > Please everyone, check your mails for signs of this bug.
> 
> So the bug symptom is that Gnus fails to display (or notice in the
> summary line?) mail which exists on disk?

Yes.

> But the mail moving bit is working, as far as you can tell?

What do you mean by "mail moving bit"?  The point is that the mail
exists in the nnml directory.

> I'm just trying to clarify, because I haven't received a mail at
> work today, which I had sent yesterday from home.  (Probably, it's
> still sitting in the queue on my home machine or something.)  A grep
> through the mail directories suggests that it's not there at all, so
> presumably this isn't the same bug.

It is not the same bug.  If it were, the grep would reveal the mail on 
disk.

-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
"Psychos _do not_ explode when sunlight hits them."

Re: Gnus fails to show some email messages

[Wes Hardaker]

Hrvoje> I have noticed this happening since some time into Quassia, and it's
Hrvoje> very serious.  It happens on some messages, and not on others.  I have
Hrvoje> no clue how to debug it, and I may be forced to downgrade to Gnus 5.5
Hrvoje> because of this bug.

I've reported this (or something similar) to Lars many a time.  He's
not able to reproduce it, so he can only offer suggestions to me on
tracking it down.  I've never succeeded.  I agree though, its a very
serious bug (there are a few in my opinion, like the agent not
handling a 'g' fetching new articles after a 'J s' and thus doing a
catchup when unplugged will mark all those articles between the 'J s' 
and the 'g' as read even though you never even saw them in the summary 
buffer).

Typically, when it happens to me it is unmarked though.  IE, I can
have a group with unread articles flagged in it that never show up.  It 
appears that your's are appearing as marked (read) and you can't see
them?

-- 
"Ninjas aren't dangerous.  They're more afraid of you than you are of them."

Re: Gnus fails to show some email messages

[Bruce Stephens]

Hrvoje Niksic <hniksic@srce.hr> writes:

> What do you mean by "mail moving bit"?  The point is that the mail
> exists in the nnml directory.

I have a procmail-equivalent deliver mail to local spool files, and
then the Gnus mail backend (nnml) moves this to a bunch of individual
files.  (That's the "mail moving bit" I was talking about.)

Now, if that had been at fault, then my missing article might have
been an example.  But that's apparently not so, so presumably sendmail
just decided not to deliver it for some reason, and the problem has
nothing to do with Gnus.

Re: Gnus fails to show some email messages

[Stefanie Teufel]

On Thu, Mar 26, 1998 at 07:32:37AM -0800, Wes Hardaker wrote:

> Typically, when it happens to me it is unmarked though.  IE, I can
> have a group with unread articles flagged in it that never show up.  It 
> appears that your's are appearing as marked (read) and you can't see
> them?

No in my case, they are unmarked too. I'm not sure, but it seems to me,
that this problem occurs especially in connection with Mailinglists.
Although I'm not absolutly sure, but at the moment I cannot remember
this failure happenend to my private mails.

Stefanie

Re: Gnus fails to show some email messages

[Wes Hardaker]

Here's another example, since it just happened to me again, details
below:

I just examined the situation a bit further.  The articles are in a
thread where the rest of the articles in the thread are expired.  C-u
enter shows the thread in question, where it didn't exist normally.

group buffer:
3          3/ 0: dcas                          

summary buffer thread in question (C-u enter to see them):

E ++ 644=[ 105: Ken Weiss           ] 6 prelim. Kerberos deployment plan
E ++     644=[  34: Ken Weiss           ] 5 
E +          135 [  22: Chris Lambertus     ] 1 
EA++         644=[  60: Ken Weiss           ] 2 
  ++             632 [  29: Wes Hardaker        ] 1 
  ++         638 [  82: Doreen Meyer        ] 1 

The last two lines are the articles that aren't shown normally...
If I un-expire any one of the parents (the top two lines), the thread
shows up again.  If I un-expire the third one (not a parent) it
disappears too.

Lars, it looks like my problem is quite possibly related to
threading...  I even removed all of my group parameters on the parent
topic to make sure none of those were affecting things and the bug is
still present.

More details:

("nnml:dcas" 3
 ((1 . 53)
  (55 . 59)
  (62 . 62))
 ((expire 53
	  (55 . 59)
	  62)
  (reply 59))
 (nnml "")
 nil)

The summary buffer does not show articles 60 and 61 (it does show 54):

.overview lines:

60	Re: prelim. Kerberos deployment plan	Wes Hardaker <wjhardaker@ucdavis.edu>	24 Mar 1998 13:31:13 -0800	<sdsoo7ixni.fsf@oakdale.ucdavis.edu>	<v03102839b13db59cf4af@[128.120.226.137]>; from Ken Weiss on  Tue, Mar 24, 1998 at 11:10:53AM -0800  <v03102837b13daa7755a9@[128.120.226.137]>  <v03102839b13db59cf4af@[128.120.226.137]> <v0310283cb13dbe5f045d@[128.120.226.137]>	811	29	Xref: oakdale.ucdavis.edu dcas:60	
61	Re: prelim. Kerberos deployment plan	Doreen Meyer <dimeyer@ucdavis.edu>	Tue, 24 Mar 1998 13:34:08 -0800	<v03102808b13dd559193a@[128.120.226.189]>	<19980324112506.25184@zen.ucdavis.edu> <v03102839b13db59cf4af@[128.120.226.137]>; from Ken Weiss on Tue, Mar 24, 1998 at 11:10:53AM -0800 <v03102837b13daa7755a9@[128.120.226.137]> <v03102839b13db59cf4af@[128.120.226.137]>	3608	82	Xref: oakdale.ucdavis.edu dcas:61	

The files exist:

% scan +. 60 61
  60  03/24 Wes Hardaker       Re: prelim. Kerberos deployment plan<<>>>>> On T
  61  03/24 Doreen Meyer       Re: prelim. Kerberos deployment plan<<Ken/Chris,

-- 
"Ninjas aren't dangerous.  They're more afraid of you than you are of them."

Re: Gnus fails to show some email messages

[Stefanie Teufel]

On Thu, Mar 26, 1998 at 09:19:28AM -0800, Wes Hardaker wrote:

> Lars, it looks like my problem is quite possibly related to
> threading...  I even removed all of my group parameters on the parent
> topic to make sure none of those were affecting things and the bug is
> still present.

That describes exactly the problems, that I have. But again, strange
enough, only related to Mailingslists like XEmacs-Beta or Ding...

Stefanie

Re: Gnus fails to show some email messages

[Dan Christensen]

Wes Hardaker <wjhardaker@ucdavis.edu> writes:
> Lars, it looks like my problem is quite possibly related to
> threading...  

It is almost definitely related to threading.  I've had this problem
for a while, and if I turn off threading it goes away.  Reparenting
seems to cause it sometimes.  I think a problem with the References
lines causes Gnus to get confused. 

Dan

-- 
Dan Christensen
jdc@math.jhu.edu

Re: Gnus fails to show some email messages

[Kim-Minh Kaplan]

[-- Attachment #1: Type: text/plain, Size: 329 bytes --]

>>>>> Wes Hardaker writes:

    Wes> it looks like my problem is quite possibly related to
    Wes> threading [...]

    Wes> <v03102839b13db59cf4af@[128.120.226.137]>
    [...]
    Wes> <v03102839b13db59cf4af@[128.120.226.137]>

Definitly, this happens with references that contain duplicate ids.
This should fix it.

Kim-Minh.

[-- Attachment #2: Type: text/plain, Size: 1517 bytes --]

--- /usr/local/lib/xemacs/gnus/lisp/gnus-sum.el-5.6.3	Fri Mar 27 21:14:05 1998
+++ /usr/local/lib/xemacs/gnus/lisp/gnus-sum.el	Fri Mar 27 21:44:35 1998
@@ -2960,7 +2960,8 @@
   (let ((headers gnus-newsgroup-headers)
 	(deps gnus-newsgroup-dependencies)
 	header references generation relations
-	cthread subject child end pthread relation new-child)
+	cthread subject child end pthread relation new-child
+	refs-uniq)
     ;; First we create an alist of generations/relations, where
     ;; generations is how much we trust the relation, and the relation
     ;; is parent/child.
@@ -2972,17 +2973,21 @@
 		   (not (string= references "")))
 	  (insert references)
 	  (setq child (mail-header-id header)
+		refs-uniq (list child)
 		subject (mail-header-subject header))
 	  (setq generation 0)
 	  (while (search-backward ">" nil t)
 	    (setq end (1+ (point)))
-	    (when (search-backward "<" nil t)
-	      (unless (string= (setq new-child (buffer-substring (point) end))
-			       child)
-		(push (list (incf generation)
-			    child (setq child new-child)
-			    subject)
-		      relations))))
+	    (when (and (search-backward "<" nil t)
+		       (not (member (setq new-child
+					  (buffer-substring (point) end))
+				    refs-uniq)))
+	      (push new-child refs-uniq)
+	      (push (list (incf generation)
+			  child (setq child new-child)
+			  subject)
+		    relations)))
 	  (push (list (1+ generation) child nil subject) relations)
 	  (erase-buffer)))
       (kill-buffer (current-buffer)))

Re: Gnus fails to show some email messages

[Eze Ogwuma]

Kim-Minh Kaplan <KimMinh.Kaplan@utopia.eunet.fr> writes:

> [1  <text/plain; US-ASCII (7bit)>]
> >>>>> Wes Hardaker writes:
> 
>     Wes> it looks like my problem is quite possibly related to
>     Wes> threading [...]
> 
>     Wes> <v03102839b13db59cf4af@[128.120.226.137]>
>     [...]
>     Wes> <v03102839b13db59cf4af@[128.120.226.137]>
> 
> Definitly, this happens with references that contain duplicate ids.

This also happens with messages that are completely original as I've
been finding out lately. Some of the messages are not from a mailing
list.

I've noticed that when I comment out 
(setq gnus-summary-ignore-duplicates t) 
from my .gnus and the restart the latest messages appear to come back.

[...]

-- 
Eze Ogwuma

Re: Gnus fails to show some email messages

[Kim-Minh Kaplan]

>>>>> Eze Ogwuma writes:

    Eze> I've noticed that when I comment out
    Eze> (setq gnus-summary-ignore-duplicates t)
    Eze> from my .gnus and the restart the latest messages appear to
    Eze> come back.

Are you sure that you don't have another message with the same ID ?
As a reminder here is what The Manual says about
`gnus-summary-ignore-duplicates' :

`gnus-summary-ignore-duplicates'
     When Gnus discovers two articles that have the same `Message-ID',
     it has to do something drastic. [...] If this variable is `t', it
     won't display the article--it'll be as if it never existed.

Re: Gnus fails to show some email messages

[Eze Ogwuma]

Kim-Minh Kaplan <KimMinh.Kaplan@utopia.eunet.fr> writes:

> >>>>> Eze Ogwuma writes:
> 
>     Eze> I've noticed that when I comment out
>     Eze> (setq gnus-summary-ignore-duplicates t)
>     Eze> from my .gnus and the restart the latest messages appear to
>     Eze> come back.
> 
> Are you sure that you don't have another message with the same ID ?
> As a reminder here is what The Manual says about
> `gnus-summary-ignore-duplicates' :

[...]

Yes, I'm sure.

I just used glimpse to index all my Mail files then I did a search on
several of the Message-IDs of messages that had been invisible. They
were not duplicated.

In this case I was lucky as I have Gnus put a "%" mark against each
folder with new mail. The folder was marked but the number stayed the
same and the messages were invisible.
-- 
Eze Ogwuma

Re: Gnus fails to show some email messages (and The horrible bug)

[Eze Ogwuma]

Hrvoje Niksic <hniksic@srce.hr> writes:

> Gnus v5.6.2; nntp 5.0; nnml 1.0; nnmh 1.0; nndraft 1.0; nnfolder 1.0
> XEmacs 21.0 "Myotonic" [Lucid] (sparc-sun-solaris2.5.1) of Mon Mar 23 1998 on jagor.srce.hr
> 200 CARNet.hr InterNetNews NNRP server INN 1.4unoff4 05-Mar-96 ready (posting ok).
> 
> 
> I have noticed a bug in Gnus: it fails to show some email messages.
> That is the second most serious bug a mail reader can exhibit (the
> first most serious bug is to delete or corrupt the message without
> showing it).  For instance, ~/Mail/xemacs.beta/20087 is present on
> disk.  It is currently the last message in the `xemacs.beta' group,
> and its NOV entry looks like this:

[...]

> The `xemacs.beta' information in `.newsrc.eld' looks like this:

> ("nnml:xemacs.beta" 2 ((1 . 20087)) ((reply 19020 19053

[...]

> l-expire . t) (expiry-wait . 14)))

How did you collect your mail? "g", "1 g" or "2 g"?

I've found that this problem can be reproduced by having mail boxes on 
different levels but only collecting mail at one level. 

The "%" mark gets put into the Group buffer and the message info is
written to the .overview file, but the message is invisible until mail
on that level is incorporated.

Wes Hardaker <wjhardaker@ucdavis.edu> writes:

> Hrvoje> I have noticed this happening since some time into Quassia, and it's
> Hrvoje> very serious.  It happens on some messages, and not on others.  I have
> Hrvoje> no clue how to debug it, and I may be forced to downgrade to Gnus 5.5
> Hrvoje> because of this bug.
> 
> I've reported this (or something similar) to Lars many a time.  He's
> not able to reproduce it, so he can only offer suggestions to me on
> tracking it down.  I've never succeeded.

[...]

> Typically, when it happens to me it is unmarked though.  IE, I can
> have a group with unread articles flagged in it that never show up.  It 
> appears that your's are appearing as marked (read) and you can't see
> them?



-- 
Eze Ogwuma