Gnus development mailing list
 help / color / mirror / Atom feed
From: keramida@ceid.upatras.gr (Giorgos Keramidas)
To: Lars Ingebrigtsen <larsi@gnus.org>
Cc: Katsumi Yamaoka <yamaoka@jpl.org>, ding@gnus.org, emacs-devel@gnu.org
Subject: Re: Emacs trunk and Gnus master are fully sync'd now
Date: Tue, 09 Jul 2013 13:08:16 +0200	[thread overview]
Message-ID: <67um8robackn33.fsf@saturn.laptop> (raw)
In-Reply-To: <8738ro52qr.fsf@randomsample.de> (David Engster's message of "Mon, 08 Jul 2013 20:24:28 +0200")

On Mon, 08 Jul 2013 20:24:28 +0200, David Engster <deng@randomsample.de> wrote:
> Giorgos Keramidas writes:
>> The messages logged point to 'unexist' for all the groups:
>
> [...]
>
>> Does `unexist' mean that the group does not exist locally?
>
> No. It means the 'unexist' range is missing, which makes Gnus think this
> Group was never synced before.
>
>> I see a few lines above that we check (and active uidvalidity unexist)
>> but if 'unexist' means that the groups does not exist in
>> `.newsrc.eld', then we are not reading the .newsrc.eld contents
>> correctly.
>
> The 'unexist' range contains article numbers which are no longer in the
> group (because they were moved or deleted).
>
> First, please make double sure that you are using at least Emacs trunk
> rev. 113257, or the latest Gnus from git.
>
> Then, to debug this further: If you hit 'G E' on a group, you see all
> the group info with the different ranges. You should see 'unexist' there
> somewhere, which can be empty or contain some article numbers or ranges.
>
> When you exit Gnus, this information should be written to the file
> newsrc.eld. Unfortunately, that file is barely human readable, but you
> can restart Emacs and load it through
>
> (load-file "~/.newsrc.eld")
>
> and then look at the variable `gnus-newsrc-alist'. Do the groups still
> have the 'unexist' ranges?

When I load a .newsrc.eld file produced by the Gnus version of emacs-24
branch the (unexist) ranges are there.  I used the following Lisp
snippet to split & group the groups based on a check for their 'unexist'
property:

    (defun group-by-unexist (newsrc-alist)
      (let ((left nil)
            (right nil))
        (mapcar (lambda (group-data)
                  (let ((name (car group-data))
                        (unexist (let ((third (nth 3 group-data)))
                                   (and (listp third)
                                        (not (null third))
                                        (assoc 'unexist third)))))
                    (if unexist
                        (setq left (cons name left))
                      (setq right (cons name right)))))
                newsrc-alist)
        (list left right)))

Running this:

    (group-by-unexist gnus-newsrc-alist)

before and after loading Gnus in Emacs 24.X shows all my mail groups in
the left group (found the 'unexist' property), and "nndrafts:drafts" in
the right group (no 'unexist' property).

Then as a second test, I started Emacs from trunk, compiled from this
changeset:

    commit 96ac507f8061d42bae399a900fe1408862e19795
    Author: Jan D. <jan.h.d@swipnet.se>
    Date:   Mon Jul 8 19:54:39 2013 +0200

        * NEWS: NS can be build with ImageMagick.

Running `M-x load-file RET ~/.newsrc.eld RET' and then:

    (group-by-unexist gnus-newsrc-alist)

correctly shows all my mail groups in the left group (they have the
'unexist' property and the `group-by-unexist' function can find it).

But then if I `M-x gnus' all groups go through an 'initial sync'.

So I think that something is broken in the way Gnus tries to read the
'unexist' value from ~/.newsrc.eld




  reply	other threads:[~2013-07-09 11:08 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-07-02 10:40 Katsumi Yamaoka
2013-07-02 16:52 ` Giorgos Keramidas
2013-07-06 15:21   ` Lars Ingebrigtsen
2013-07-06 15:44     ` David Engster
2013-07-06 18:45       ` David Engster
2013-07-06 20:03         ` David Engster
2013-07-07 13:27           ` Giorgos Keramidas
2013-07-07 13:24       ` Giorgos Keramidas
2013-07-07 19:51         ` David Engster
2013-07-08 13:16           ` Giorgos Keramidas
2013-07-08 18:24             ` David Engster
2013-07-09 11:08               ` Giorgos Keramidas [this message]
2013-07-09 15:34                 ` David Engster
2013-07-09 18:16                   ` Giorgos Keramidas
2013-07-09 19:33                     ` David Engster
2013-07-09 19:39                       ` Ted Zlatanov
2013-07-09 20:41                       ` Giorgos Keramidas
2013-07-09 21:42                         ` David Engster
2013-07-10 15:32                           ` Ted Zlatanov
2013-07-10 19:17                           ` David Engster
2013-07-11 17:12                             ` Giorgos Keramidas
2013-07-18 13:31                           ` Lars Magne Ingebrigtsen
2013-07-18 15:42                             ` Giorgos Keramidas
2013-07-18 16:55                               ` David Engster
2013-07-18 17:15                                 ` Giorgos Keramidas
2013-07-30 14:56                               ` Lars Magne Ingebrigtsen
2013-07-30 19:58                                 ` Giorgos Keramidas

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=67um8robackn33.fsf@saturn.laptop \
    --to=keramida@ceid.upatras.gr \
    --cc=ding@gnus.org \
    --cc=emacs-devel@gnu.org \
    --cc=larsi@gnus.org \
    --cc=yamaoka@jpl.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).