Wrong behaviour while fetching message with IMAP

Wrong behaviour while fetching message with IMAP

[Julien Danjou]

[-- Attachment #1: Type: text/plain, Size: 507 bytes --]

Clicking on a message id in a mail, does very weird thing it seems.

Simplified backtrace:

* gnus-read-header("<92164B15-45D1-4D76-9AF6-8E1E098C04BD@telia.com>")
  gnus-summary-insert-by-id("<92164B15-45D1-4D76-9AF6-8E1E098C04BD@telia.com>")
  gnus-summary-refer-article("92164B15-45D1-4D76-9AF6-8E1E098C04BD@telia.com")

read-header returns data, whereas it should probably returns nil. Then a
mail with no info is inserted into the summary.

-- 
Julien Danjou
❱ http://julien.danjou.info

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: Wrong behaviour while fetching message with IMAP

[Lars Ingebrigtsen]

Julien Danjou <julien@danjou.info> writes:

> Clicking on a message id in a mail, does very weird thing it seems.
>
> Simplified backtrace:
>
> * gnus-read-header("<92164B15-45D1-4D76-9AF6-8E1E098C04BD@telia.com>")

Interesting.  This basically boils down to the two following commands:

17:16:09 7071 EXAMINE "ding"
17:16:16 7072 UID SEARCH HEADER Message-Id "<92164B15-45D1-4D76-9AF6-8E1E098C04BD@telia.com>"

And this is what my IMAP server is returning:

* SEARCH
7072 OK Search completed (0.032 secs).

I.e., it didn't find it.  I think.  But the article is clearly in the
ding mailbox, and fetching messages by Message-ID used to work.  I mean,
that's what's used when moving articles, too...

And that code hasn't changed for a while.

Anybody?  This is with a Dovecot IMAP server.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Wrong behaviour while fetching message with IMAP

[Andrew Cohen]

>>>>> "Lars" == Lars Ingebrigtsen <larsi@gnus.org> writes:

    Lars> Julien Danjou <julien@danjou.info> writes:
    >> Clicking on a message id in a mail, does very weird thing it
    >> seems.
    >> 
    >> Simplified backtrace:
    >> 
    >> *
    >> gnus-read-header("<92164B15-45D1-4D76-9AF6-8E1E098C04BD@telia.com>")

    Lars> Interesting.  This basically boils down to the two following
    Lars> commands:

    Lars> 17:16:09 7071 EXAMINE "ding" 17:16:16 7072 UID SEARCH HEADER
    Lars> Message-Id "<92164B15-45D1-4D76-9AF6-8E1E098C04BD@telia.com>"

    Lars> And this is what my IMAP server is returning:

    Lars> * SEARCH 7072 OK Search completed (0.032 secs).

    Lars> I.e., it didn't find it.  I think.  But the article is clearly
    Lars> in the ding mailbox, and fetching messages by Message-ID used
    Lars> to work.  I mean, that's what's used when moving articles,
    Lars> too...

    Lars> And that code hasn't changed for a while.

    Lars> Anybody?  This is with a Dovecot IMAP server.


I just tried it and it worked for me (I haven't updated from git since
Jan 24 though). Just to be clear: I emailed myself a message with the
message-id of an existing email on my imap server in the body. Clicking
on the message-id in this new email adds the old one to the summary
buffer and opens the email in the article buffer. I also stepped through
nnimap-find-article-by-message-id while this was happening to see that
the search was performed and the message was found. Both messages were
in the same imap group (since  IMAP SEARCH doesn't work across groups I
suppose that the messages would have to be in the same group).

Regards,
Andy

Re: Wrong behaviour while fetching message with IMAP

[James Cloos]

>>>>> "AC" == Andrew Cohen <cohen@andy.bu.edu> writes:
>>>>> "Lars" == Lars Ingebrigtsen <larsi@gnus.org> writes:

Lars> Interesting.  This basically boils down to the two following
Lars> commands:

Lars> 17:16:09 7071 EXAMINE "ding" 17:16:16 7072 UID SEARCH HEADER
Lars> Message-Id "<92164B15-45D1-4D76-9AF6-8E1E098C04BD@telia.com>"

Lars> And this is what my IMAP server is returning:

Lars> * SEARCH 7072 OK Search completed (0.032 secs).

Lars> I.e., it didn't find it.

AC> I just tried it and it worked for me (I haven't updated from git since
AC> Jan 24 though).

I'm also a bit behind; last updated on the 25th.

I tried it while reading Andrew's message.  When I hit return with the
point in the References: header on Lars' m-id gnus jumped to that aricle.
When I returned to Andrew's and hit the older m-id in reference (which
already has been expired here) it reported this:

  apply: Format specifier doesn't match argument type

I suppose 'article not found' or 'message-id not found' would be
better.  ☺

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: Wrong behaviour while fetching message with IMAP

[Lars Ingebrigtsen]

James Cloos <cloos@jhcloos.com> writes:

>   apply: Format specifier doesn't match argument type

`(setq debug-on-error t)', repeat the bug and post the resulting
backtrace.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Wrong behaviour while fetching message with IMAP

[James Cloos]

>>>>> "LI" == Lars Ingebrigtsen <larsi@gnus.org> writes:

LI> James Cloos <cloos@jhcloos.com> writes:
>> apply: Format specifier doesn't match argument type

LI> `(setq debug-on-error t)', repeat the bug and post the resulting
LI> backtrace.

This is from hitting return with the point on the first reference in
your article (the one I am replying to now):

Debugger entered--Lisp error: (error "Format specifier doesn't match argument type")
  format("UID FETCH %d (UID RFC822.SIZE BODYSTRUCTURE BODY.PEEK[HEADER.FIELDS (Subject From Date Message-Id References In-Reply-To Xref To Newsgroups)])" "nil")
  apply(format ("UID FETCH %d (UID RFC822.SIZE BODYSTRUCTURE BODY.PEEK[HEADER.FIELDS (Subject From Date Message-Id References In-Reply-To Xref To Newsgroups)])" nil))
  nnimap-send-command("UID FETCH %d (UID RFC822.SIZE BODYSTRUCTURE BODY.PEEK[HEADER.FIELDS (Subject From Date Message-Id References In-Reply-To Xref To Newsgroups)])" nil)
  apply(nnimap-send-command ("UID FETCH %d (UID RFC822.SIZE BODYSTRUCTURE BODY.PEEK[HEADER.FIELDS (Subject From Date Message-Id References In-Reply-To Xref To Newsgroups)])" nil))
  nnimap-command("UID FETCH %d (UID RFC822.SIZE BODYSTRUCTURE BODY.PEEK[HEADER.FIELDS (Subject From Date Message-Id References In-Reply-To Xref To Newsgroups)])" nil)
  nnimap-get-whole-article(nil "UID FETCH %d (UID RFC822.SIZE BODYSTRUCTURE BODY.PEEK[HEADER.FIELDS (Subject From Date Message-Id References In-Reply-To Xref To Newsgroups)])")
  nnimap-request-head(#("<sa3r5byy4xn.fsf@cigue.easter-eggs.fr>" 0 38 (face gnus-header-content)) "ding@gnus.org" "lugabout")
  gnus-request-head(#("<sa3r5byy4xn.fsf@cigue.easter-eggs.fr>" 0 38 (face gnus-header-content)) "nnimap+lugabout:ding@gnus.org")
  gnus-read-header(#("<sa3r5byy4xn.fsf@cigue.easter-eggs.fr>" 0 38 (face gnus-header-content)))
  gnus-summary-insert-subject(#("<sa3r5byy4xn.fsf@cigue.easter-eggs.fr>" 0 38 (face gnus-header-content)))
  byte-code("\304 \305\211\203-	@\306\b!\203&\307\n!\211\203&\310\305\211\211\v$\210\311\312\313\"\210	A\211\204	*\314\315\316\n#\207" [gnus-override-method --cl-dolist-temp-- message-id number gnus-refer-article-methods nil gnus-check-server gnus-summary-insert-subject gnus-summary-select-article throw found t gnus-message 3 "Couldn't fetch article %s"] 6)
  gnus-summary-refer-article(#("<sa3r5byy4xn.fsf@cigue.easter-eggs.fr>" 0 38 (face gnus-header-content)))
  gnus-button-message-id(#("<sa3r5byy4xn.fsf@cigue.easter-eggs.fr>" 0 38 (face gnus-header-content)))
  gnus-article-press-button()
  gnus-widget-press-button((link :delete widget-leave-text :action gnus-widget-press-button :help-echo "Follow the link" :keymap (keymap (menu-bar keymap (Url\ Button menu-item "Url Button" (keymap "Url Button" (Copy\ URL\ to\ kill\ ring menu-item "Copy URL to kill ring" gnus-article-copy-string)))) (117 . gnus-article-copy-string)) :button-keymap nil :from #<marker (moves after insertion) at 2326 in *Article*> :to #<marker at 2364 in *Article*> :button-overlay #<overlay from 2326 to 2364 in *Article*>) nil)
  widget-apply((link :delete widget-leave-text :action gnus-widget-press-button :help-echo "Follow the link" :keymap (keymap (menu-bar keymap (Url\ Button menu-item "Url Button" (keymap "Url Button" (Copy\ URL\ to\ kill\ ring menu-item "Copy URL to kill ring" gnus-article-copy-string)))) (117 . gnus-article-copy-string)) :button-keymap nil :from #<marker (moves after insertion) at 2326 in *Article*> :to #<marker at 2364 in *Article*> :button-overlay #<overlay from 2326 to 2364 in *Article*>) :action nil)
  widget-apply-action((link :delete widget-leave-text :action gnus-widget-press-button :help-echo "Follow the link" :keymap (keymap (menu-bar keymap (Url\ Button menu-item "Url Button" (keymap "Url Button" (Copy\ URL\ to\ kill\ ring menu-item "Copy URL to kill ring" gnus-article-copy-string)))) (117 . gnus-article-copy-string)) :button-keymap nil :from #<marker (moves after insertion) at 2326 in *Article*> :to #<marker at 2364 in *Article*> :button-overlay #<overlay from 2326 to 2364 in *Article*>) nil)
  widget-button-press(2338)
  call-interactively(widget-button-press nil nil)

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: Wrong behaviour while fetching message with IMAP

[Lars Ingebrigtsen]

James Cloos <cloos@jhcloos.com> writes:

>   gnus-request-head(#("<sa3r5byy4xn.fsf@cigue.easter-eggs.fr>" 0 38 (face gnus-header-content)) "nnimap+lugabout:ding@gnus.org")

So it's fetching by Message-ID that's triggering this bug again.

I've now added extra checks in this function.  But I'm really puzzled by
Message-ID searching not working any more on certain servers.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen