Totally incorrect splitting

Totally incorrect splitting

[Norman Walsh]

[-- Attachment #1: Type: text/plain, Size: 746 bytes --]

Hi,

I'm not sure how to debug this. I'm running Ma Gnus 0.8 (maybe I
should upgrade?) on GNU Emacs 24.3.1. I'm using fancy splitting.

Occasionally, messages go in completely incorrect groups.

Curiously, when I run 'B q' on them, they list the correct groups as
the targets.

I've got enough groups and enough email that I find this a bit
frightening.

I'm looking for suggestions about how to figure out what is going on...

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | If you are losing your leisure, look
http://nwalsh.com/            | out! You may be losing your
                              | soul.--Logan Pearsall Smith

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 186 bytes --]

Re: Totally incorrect splitting

[Steinar Bang]

>>>>> Norman Walsh <ndw@nwalsh.com>:

> I'm not sure how to debug this. I'm running Ma Gnus 0.8 (maybe I
> should upgrade?) on GNU Emacs 24.3.1. I'm using fancy splitting.

With what backend?

> Occasionally, messages go in completely incorrect groups.

> Curiously, when I run 'B q' on them, they list the correct groups as
> the targets.

> I've got enough groups and enough email that I find this a bit
> frightening.

FWIW I dealt with heaps of mailing lists by getting the mailing lists
subscribed to gmane.

(Granted with closed(ish) working group mailing lists, that is not an
option...)

Re: Totally incorrect splitting

[Norman Walsh]

[-- Attachment #1: Type: text/plain, Size: 1280 bytes --]

Steinar Bang <sb@dod.no> writes:

>>>>>> Norman Walsh <ndw@nwalsh.com>:
>
>> I'm not sure how to debug this. I'm running Ma Gnus 0.8 (maybe I
>> should upgrade?) on GNU Emacs 24.3.1. I'm using fancy splitting.
>
> With what backend?

nnml

I'm suspicious of gnus-registry, but it has seemed to work for years.
(Q: "When did this start happening, what had you changed?" A: "A while
ago, and I don't really recall." :-( )

(setq gnus-registry-max-entries          20000
      gnus-registry-split-strategy       nil
      gnus-registry-use-long-group-names t
      guns-registry-track-extra          '(sender subject))

> (Granted with closed(ish) working group mailing lists, that is not an
> option...)

Many of the lists are internal $EMPLOYER lists that I can't subscribe
to any other way.


                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | Everything we love, no doubt, will pass
http://nwalsh.com/            | away, perhaps tomorrow, perhaps a
                              | thousand years hence. Neither it nor
                              | our love for it is any the less
                              | valuable for that reason.--John Passmore

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 186 bytes --]

Re: Totally incorrect splitting

[Ted Zlatanov]

On Tue, 10 Dec 2013 12:02:54 -0600 Norman Walsh <ndw@nwalsh.com> wrote: 

NW> Steinar Bang <sb@dod.no> writes:
>>>>>>> Norman Walsh <ndw@nwalsh.com>:
>> 
>>> I'm not sure how to debug this. I'm running Ma Gnus 0.8 (maybe I
>>> should upgrade?) on GNU Emacs 24.3.1. I'm using fancy splitting.
>> 
>> With what backend?

NW> nnml

NW> I'm suspicious of gnus-registry, but it has seemed to work for years.
NW> (Q: "When did this start happening, what had you changed?" A: "A while
NW> ago, and I don't really recall." :-( )

NW> (setq gnus-registry-max-entries          20000
NW>       gnus-registry-split-strategy       nil
NW>       gnus-registry-use-long-group-names t
NW>       guns-registry-track-extra          '(sender subject))

You can set `gnus-verbose' to 9 and endure the noise until you catch the
problem.  Sorry I don't have a better way.

Ted

Re: Totally incorrect splitting

[Norman Walsh]

[-- Attachment #1: Type: text/plain, Size: 577 bytes --]

Ted Zlatanov <tzz@lifelogs.com> writes:
> You can set `gnus-verbose' to 9 and endure the noise until you catch the
> problem.  Sorry I don't have a better way.

Ugh. It's pretty random and not something I notice right away.

I found a message that was misfiled and re-delivered it, but it got
put in the right place the second time :-(

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | You look wise. Pray correct that
http://nwalsh.com/            | error.--Charles Lamb

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 186 bytes --]

Re: Totally incorrect splitting

[Norman Walsh]

[-- Attachment #1: Type: text/plain, Size: 1471 bytes --]

Just FYI:

Ted Zlatanov <tzz@lifelogs.com> writes:
> NW> (setq gnus-registry-max-entries          20000
> NW>       gnus-registry-split-strategy       nil
> NW>       gnus-registry-use-long-group-names t
> NW>       guns-registry-track-extra          '(sender subject))

First, note my clever spelling of "gnus" in
"guns-registry-track-extra". That was fun to find.

> You can set `gnus-verbose' to 9 and endure the noise until you catch the
> problem.  Sorry I don't have a better way.

Yes, thanks. I finally got around to tracking this down. I added a new
split rule that *never* worked and that made it a lot easier to debug.

In this case, it turned out that gnus-registry-split-fancy-with-parent
was splitting on subject and the subject was generic enough to have
already been dropped in my "misc" bucket. Consequently, my new rule
never fired.

I expect that explains my other incorrect splits as well.

I've removed "subject" from track-extra for the time being and I'll
see how that goes. On the one hand, it was sometimes handy to have
subject in the list, but I think the false splits probably outweigh
that convenience.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | There is nothing new under the sun but
http://nwalsh.com/            | there are lots of old things we don't
                              | know.--Ambrose Bierce

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Totally incorrect splitting

[Ted Zlatanov]

On Thu, 30 Jan 2014 08:58:53 -0600 Norman Walsh <ndw@nwalsh.com> wrote: 

NW> Just FYI:
NW> Ted Zlatanov <tzz@lifelogs.com> writes:
NW> (setq gnus-registry-max-entries          20000
NW> gnus-registry-split-strategy       nil
NW> gnus-registry-use-long-group-names t
NW> guns-registry-track-extra          '(sender subject))

NW> First, note my clever spelling of "gnus" in
NW> "guns-registry-track-extra". That was fun to find.

>> You can set `gnus-verbose' to 9 and endure the noise until you catch the
>> problem.  Sorry I don't have a better way.

NW> Yes, thanks. I finally got around to tracking this down. I added a new
NW> split rule that *never* worked and that made it a lot easier to debug.

NW> In this case, it turned out that gnus-registry-split-fancy-with-parent
NW> was splitting on subject and the subject was generic enough to have
NW> already been dropped in my "misc" bucket. Consequently, my new rule
NW> never fired.

NW> I expect that explains my other incorrect splits as well.

NW> I've removed "subject" from track-extra for the time being and I'll
NW> see how that goes. On the one hand, it was sometimes handy to have
NW> subject in the list, but I think the false splits probably outweigh
NW> that convenience.

I'm glad you found the problem.  That misspelling is a tricky one.

Ted