fancy splitting interactively

fancy splitting interactively

[Uwe Brauer]

[-- Attachment #1: Type: text/plain, Size: 528 bytes --]

Hello

I usually have the following setting.
(setq 
 nnimap-split-rule  'nnimap-split-fancy
 nnimap-split-inbox "INBOX"
 nnimap-split-fancy
  `(|									;macro
    ,@(bbdb-nnimap-folder-list-from-bbdb)	
   ("Content-type" "application/msword" "DOC")))

However this seems only to work when I start gnus and the message in the
INBOX are unread.

How can I run this splitting on the INBOX for all messages even read
ones?
(Thunderbird has such a feature and I have to use it regularly).

thanks

Uwe Brauer 

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 6007 bytes --]

Re: fancy splitting interactively

[Glyn Millington]

Uwe Brauer <oub@mat.ucm.es> writes:

> Hello
>
> I usually have the following setting.
> (setq 
>  nnimap-split-rule  'nnimap-split-fancy
>  nnimap-split-inbox "INBOX"
>  nnimap-split-fancy
>   `(|									;macro
>     ,@(bbdb-nnimap-folder-list-from-bbdb)	
>    ("Content-type" "application/msword" "DOC")))
>
> However this seems only to work when I start gnus and the message in the
> INBOX are unread.
>
> How can I run this splitting on the INBOX for all messages even read
> ones?
> (Thunderbird has such a feature and I have to use it regularly).

Mark the messages you want re-spooled (#)

Then do
 B r in the summary buffer


It is possible I have misunderstood what you want! 

atb

Glyn

Re: fancy splitting interactively

[Eric S Fraga]

On Saturday,  3 Jan 2015 at 11:22, Uwe Brauer wrote:
> How can I run this splitting on the INBOX for all messages even read
> ones?

Possibly:

,----[ C-h v nnimap-unsplittable-articles RET ]
| nnimap-unsplittable-articles is a variable defined in `nnimap.el'.
| Its value is (%Deleted %Seen)
| 
| Documentation:
| Articles with the flags in the list will not be considered when splitting.
| 
| This is a Gnus server variable.  See Info node `(gnus)Select Methods'.
| 
| [back]
`----

-- 
: Eric S Fraga, GnuPG: 0xFFFCF67D
: in Emacs 25.0.50.1 + Ma Gnus v0.12 + evil-git-bbc9fa3
: BBDB version 3.1.2 (2014-05-06 11:45:08 -0500)

Re: fancy splitting interactively

[Eric Abrahamsen]

Eric S Fraga <e.fraga@ucl.ac.uk> writes:

> On Saturday,  3 Jan 2015 at 11:22, Uwe Brauer wrote:
>> How can I run this splitting on the INBOX for all messages even read
>> ones?
>
> Possibly:
>
> ,----[ C-h v nnimap-unsplittable-articles RET ]
> | nnimap-unsplittable-articles is a variable defined in `nnimap.el'.
> | Its value is (%Deleted %Seen)
> | 
> | Documentation:
> | Articles with the flags in the list will not be considered when splitting.
> | 
> | This is a Gnus server variable.  See Info node `(gnus)Select Methods'.
> | 
> | [back]
> `----

Yup, this is what I've done in the past and it works a treat. Just set
it to (%Deleted) temporarily, then refresh.

Re: fancy splitting interactively

[Adam Sjøgren]

Uwe writes:

> How can I run this splitting on the INBOX for all messages even read
> ones?

You can always mark the relevant messages with # and then run B r on
them.

,----
| B r runs the command gnus-summary-respool-article, which is an
| interactive compiled Lisp function in `gnus-sum.el'.
| 
| It is bound to B r, <menu-bar> <Article> <Copy, move,... (Backend)>
| <Respool article...>.
| 
| (gnus-summary-respool-article &optional N METHOD)
| 
| Respool the current article.
| The article will be squeezed through the mail spooling process again,
| which means that it will be put in some mail newsgroup or other
| depending on `nnmail-split-methods'.
| If N is a positive number, respool the N next articles.
| If N is a negative number, respool the N previous articles.
| If N is nil and any articles have been marked with the process mark,
| respool those articles instead.
| 
| Respooling can be done both from mail groups and "real" newsgroups.
| In the former case, the articles in question will be moved from the
| current group into whatever groups they are destined to.  In the
| latter case, they will be copied into the relevant groups.
`----

You can use B q on a message to see what splitting will do to it, if you
actually do B r on it.

,----
| B q runs the command gnus-summary-respool-query, which is an
| interactive compiled Lisp function in `gnus-sum.el'.
| 
| It is bound to B q, <menu-bar> <Article> <Copy, move,... (Backend)>
| <Query respool>.
| 
| (gnus-summary-respool-query &optional SILENT TRACE)
| 
| Query where the respool algorithm would put this article.
`----


  Best regards,

    Adam

-- 
 "Vegetarian progressive grindcore"                           Adam Sjøgren
                                                         asjo@koldfront.dk

Re: fancy splitting interactively

[Uwe Brauer]

[-- Attachment #1: Type: text/plain, Size: 365 bytes --]

>> "Adam" == Adam Sjøgren <asjo@koldfront.dk> writes:

   > Uwe writes:
   >> How can I run this splitting on the INBOX for all messages even read
   >> ones?

   > You can always mark the relevant messages with # and then run B r on
   > them.

This is not what I want. I don't want to mark them myself, this is a
issue which should be done by gnus....

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 6007 bytes --]

Re: fancy splitting interactively

[Uwe Brauer]

[-- Attachment #1: Type: text/plain, Size: 1057 bytes --]

>> "Eric" == Eric Abrahamsen <eric@ericabrahamsen.net> writes:

   > Eric S Fraga <e.fraga@ucl.ac.uk> writes:
   >> On Saturday,  3 Jan 2015 at 11:22, Uwe Brauer wrote:
>> How can I run this splitting on the INBOX for all messages even read
>> ones?
   >> 
   >> Possibly:
   >> 
   >> ,----[ C-h v nnimap-unsplittable-articles RET ]
   >> | nnimap-unsplittable-articles is a variable defined in `nnimap.el'.
   >> | Its value is (%Deleted %Seen)
   >> | 
   >> | Documentation:
   >> | Articles with the flags in the list will not be considered when splitting.
   >> | 
   >> | This is a Gnus server variable.  See Info node `(gnus)Select Methods'.
   >> | 
   >> | [back]
   >> `----

   > Yup, this is what I've done in the past and it works a treat. Just set
   > it to (%Deleted) temporarily, then refresh.

Thanks for pointing this variable out to me, I see that it is not
customize-able via customize, so I have to use:


(setq nnimap-unsplittable-articles '(%Deleted))

By refresh you mean 
gnus-group-get-new-news??

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 6007 bytes --]

Re: fancy splitting interactively

[Adam Sjøgren]

Uwe writes:

> This is not what I want. I don't want to mark them myself, this is a
> issue which should be done by gnus....

I'm sure a superuser like you can figure out how to mark all articles
with M P a before pressing B r.


  Best regards,

    Adam

-- 
 "We're not unreasonable, I mean, noone's going to eat        Adam Sjøgren
  your eyes"                                             asjo@koldfront.dk

Re: fancy splitting interactively

[Uwe Brauer]

[-- Attachment #1: Type: text/plain, Size: 417 bytes --]

>> "Adam" == Adam Sjøgren <asjo@koldfront.dk> writes:

   > Uwe writes:
   >> This is not what I want. I don't want to mark them myself, this is a
   >> issue which should be done by gnus....

   > I'm sure a superuser like you can figure out how to mark all articles
   > with M P a before pressing B r.

Thanks for this brilliant solution, which I would not have discovered
myself. 


regards

Uwe 

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 6007 bytes --]

Re: fancy splitting interactively

[Eric Abrahamsen]

Uwe Brauer <oub@mat.ucm.es> writes:

>>> "Eric" == Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>
>    > Eric S Fraga <e.fraga@ucl.ac.uk> writes:
>    >> On Saturday,  3 Jan 2015 at 11:22, Uwe Brauer wrote:
>>> How can I run this splitting on the INBOX for all messages even read
>>> ones?
>    >> 
>    >> Possibly:
>    >> 
>    >> ,----[ C-h v nnimap-unsplittable-articles RET ]
>    >> | nnimap-unsplittable-articles is a variable defined in `nnimap.el'.
>    >> | Its value is (%Deleted %Seen)
>    >> | 
>    >> | Documentation:
>    >> | Articles with the flags in the list will not be considered when splitting.
>    >> | 
>    >> | This is a Gnus server variable.  See Info node `(gnus)Select Methods'.
>    >> | 
>    >> | [back]
>    >> `----
>
>    > Yup, this is what I've done in the past and it works a treat. Just set
>    > it to (%Deleted) temporarily, then refresh.
>
> Thanks for pointing this variable out to me, I see that it is not
> customize-able via customize, so I have to use:
>
>
> (setq nnimap-unsplittable-articles '(%Deleted))
>
> By refresh you mean 
> gnus-group-get-new-news??

Yes, any entry point into the "fetch new news" process will run the
splitting routine.

I think the idea is that you aren't really meant to run Gnus regularly
without the %Seen flag in that variable. That means (as far as I
understand it) that *every* time you get new news, *all* of your
existing messages will be run through the splitting process again. That
could be heavy, though to be honest when I've done this in the past I
don't remember it taking nearly as long as I thought it would.

What's meant to happen is that *unseen* messages are split once, on
their way in, and then left where they are.

It's been noted on this list that this can produce unwanted results if
you're accessing your mail from several devices, and some other device
marks a message as read before Gnus gets to it -- that way Gnus won't
know to split it. I'm not sure what the proper solution is there, but I
still don't think it's running day-to-day with %Seen removed from
`nnimap-unsplittable-articles'.

Eric

[bug] (was: fancy splitting interactively)

[Uwe Brauer]

[-- Attachment #1.1: Type: text/plain, Size: 360 bytes --]

>> "Adam" == Adam Sjøgren <asjo@koldfront.dk> writes:

   > Uwe writes:
   >> This is not what I want. I don't want to mark them myself, this is a
   >> issue which should be done by gnus....

   > I'm sure a superuser like you can figure out how to mark all articles
   > with M P a before pressing B r.

The following bug showed up when Using B r 

[-- Attachment #1.2: spool-bug.txt --]
[-- Type: text/plain, Size: 1537 bytes --]

Debugger entered--Lisp error: (wrong-type-argument char-or-string-p nil)
  decode-coding-string(nil utf-8)
  nnimap-decode-gnus-group(nil)
  nnimap-request-accept-article(nil "UCMgmail" t)
  gnus-request-accept-article(nil (nnimap "UCMgmail" (nnimap-address "imap.gmail.com") (nnimap-server-port 993) (nnimap-authinfo-file "~/.authinfo") (nnimap-stream ssl)) t t)
  eval((gnus-request-accept-article nil (quote (nnimap "UCMgmail" ... ... ... ...)) t t))
  nnimap-request-move-article(19859 "INBOX" "UCMgmail" (gnus-request-accept-article nil (quote (nnimap "UCMgmail" ... ... ... ...)) t t) t nil)
  gnus-request-move-article(19859 "nnimap+UCMgmail:INBOX" "UCMgmail" (gnus-request-accept-article nil (quote (nnimap "UCMgmail" ... ... ... ...)) t t) t nil)
  gnus-summary-move-article(nil nil (nnimap "UCMgmail" (nnimap-address "imap.gmail.com") (nnimap-server-port 993) (nnimap-authinfo-file "~/.authinfo") (nnimap-stream ssl)))
  #<compiled-function (from gnus-summary-respool-article) (&optional n method) "...(32)" [gnus-newsgroup-name n method error "No method given for respooling" assoc symbol-name gnus-find-method-for-group gnus-methods-using respool gnus-summary-move-article nil gnus-summary-copy-article] 4 ("/home/oub/xemacs/site-lisp/packages/ma-gnus-09/lisp/gnus-sum.elc" . 293934) (list current-prefix-arg (let* ... ...)) 0x2f4b>(nil (nnimap "UCMgmail" (nnimap-address "imap.gmail.com") (nnimap-server-port 993) (nnimap-authinfo-file "~/.authinfo") (nnimap-stream ssl)))
  call-interactively(gnus-summary-respool-article)

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 6007 bytes --]

Re: fancy splitting interactively

[Eric S Fraga]

On Sunday,  4 Jan 2015 at 11:12, Eric Abrahamsen wrote:

[...]

> I think the idea is that you aren't really meant to run Gnus regularly
> without the %Seen flag in that variable. That means (as far as I
> understand it) that *every* time you get new news, *all* of your
> existing messages will be run through the splitting process again. That
> could be heavy, though to be honest when I've done this in the past I
> don't remember it taking nearly as long as I thought it would.

Indeed but it works well when you have more than one device
(esp. phones) accessing the inbox, as you note later in your post, *and*
if you have a Inbox 0 approach!

My approach, when I use gnus splitting, is that I do not expect any
emails in my INBOX once gnus has processed them.  I do access emails on
my phone frequently so I need those that have been read elsewhere split
as well.  In my case, *all* emails in my INBOX get split off to
somewhere, whether to a special folder or to a "general" catch-all
folder for emails that I cannot classify.

I hope this makes sense...

-- 
: Eric S Fraga, GnuPG: 0xFFFCF67D
: in Emacs 25.0.50.1 + Ma Gnus v0.12 + evil-git-bbc9fa3
: BBDB version 3.1.2 (2014-05-06 11:45:08 -0500)

Re: fancy splitting interactively

[Eric Abrahamsen]

Eric S Fraga <e.fraga@ucl.ac.uk> writes:

> On Sunday,  4 Jan 2015 at 11:12, Eric Abrahamsen wrote:
>
> [...]
>
>> I think the idea is that you aren't really meant to run Gnus regularly
>> without the %Seen flag in that variable. That means (as far as I
>> understand it) that *every* time you get new news, *all* of your
>> existing messages will be run through the splitting process again. That
>> could be heavy, though to be honest when I've done this in the past I
>> don't remember it taking nearly as long as I thought it would.
>
> Indeed but it works well when you have more than one device
> (esp. phones) accessing the inbox, as you note later in your post, *and*
> if you have a Inbox 0 approach!
>
> My approach, when I use gnus splitting, is that I do not expect any
> emails in my INBOX once gnus has processed them.  I do access emails on
> my phone frequently so I need those that have been read elsewhere split
> as well.  In my case, *all* emails in my INBOX get split off to
> somewhere, whether to a special folder or to a "general" catch-all
> folder for emails that I cannot classify.
>
> I hope this makes sense...

Yeah, that definitely makes sense, and if you're willing to adopt the
no-messages-in-the-INBOX approach then it all works out quite nicely. I
still think someone clever should come up with a more certain solution
to the problem, though. Multiple devices -- like multiple email
addresses -- are becoming the norm rather than the exception, and in
both cases Gnus has a bit of catch-up to do...

Re: fancy splitting interactively

[Eric S Fraga]

On Monday,  5 Jan 2015 at 22:34, Eric Abrahamsen wrote:

[...]

> still think someone clever should come up with a more certain solution
> to the problem, though. Multiple devices -- like multiple email
> addresses -- are becoming the norm rather than the exception, and in
> both cases Gnus has a bit of catch-up to do...

Funnily enough, my view is that it is all the other MUAs that need
catching up... :)


-- 
: Eric S Fraga, GnuPG: 0xFFFCF67D
: in Emacs 25.0.50.1 + Ma Gnus v0.12 + evil-git-bbc9fa3
: BBDB version 3.1.2 (2014-05-06 11:45:08 -0500)

Re: fancy splitting interactively

[Adam Sjøgren]

Eric writes:

> On Monday,  5 Jan 2015 at 22:34, Eric Abrahamsen wrote:

>> still think someone clever should come up with a more certain solution
>> to the problem, though. Multiple devices -- like multiple email
>> addresses -- are becoming the norm rather than the exception, and in
>> both cases Gnus has a bit of catch-up to do...

> Funnily enough, my view is that it is all the other MUAs that need
> catching up... :)

I agree with Eric!


Seriously, though, what would make Gnus "catch-up"?


  /A

-- 
 "Eternal evil                                                Adam Sjøgren
  Yeah it stares behind the desk"                        asjo@koldfront.dk

Re: fancy splitting interactively

[Eric Abrahamsen]

asjo@koldfront.dk (Adam Sjøgren) writes:

> Eric writes:
>
>> On Monday,  5 Jan 2015 at 22:34, Eric Abrahamsen wrote:
>
>>> still think someone clever should come up with a more certain solution
>>> to the problem, though. Multiple devices -- like multiple email
>>> addresses -- are becoming the norm rather than the exception, and in
>>> both cases Gnus has a bit of catch-up to do...
>
>> Funnily enough, my view is that it is all the other MUAs that need
>> catching up... :)

I usually agree completely!

> I agree with Eric!
>
>
> Seriously, though, what would make Gnus "catch-up"?

Okay, maybe "catch up" isn't quite right, but these are two areas where
Gnus isn't miles ahead, like it usually is. For the
multiple-email-addresses issue, it seems like we ought to have something
baked into message.el that lets you set up different identities -- sort
of like posting styles, but for blank composed messages. There's a
package called gnus-alias in Melpa that looks like it does pretty much
that, maybe it could be considered for inclusion.

The multiple device problem is trickier. The idea of Gnus-specific marks
was floated, and Lars was also talking about something called "Gnus
Cloudy" which might also address this. It's not an area I'm very
familiar with, and I guess I don't know what the best solution is...

Re: fancy splitting interactively

[Russ Allbery]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> Okay, maybe "catch up" isn't quite right, but these are two areas where
> Gnus isn't miles ahead, like it usually is. For the
> multiple-email-addresses issue, it seems like we ought to have something
> baked into message.el that lets you set up different identities -- sort
> of like posting styles, but for blank composed messages.

I don't understand why you'd need something like posting styles for blank
composed messages?  Posting styles work fine for blank composed messages.
I've been using Gnus with multiple identities for more than a decade.
Just move to a group that matches the identity you want to post with, and
do C-u a, or compose the message while you're reading a group.

I suppose the UI could be a bit better since moving around your folders to
select the identity may be a bit awkward, but this works quite well for
me.

-- 
Russ Allbery (eagle@eyrie.org)              <http://www.eyrie.org/~eagle/>

Re: fancy splitting interactively

[Eric Abrahamsen]

Russ Allbery <eagle@eyrie.org> writes:

> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>
>> Okay, maybe "catch up" isn't quite right, but these are two areas where
>> Gnus isn't miles ahead, like it usually is. For the
>> multiple-email-addresses issue, it seems like we ought to have something
>> baked into message.el that lets you set up different identities -- sort
>> of like posting styles, but for blank composed messages.
>
> I don't understand why you'd need something like posting styles for blank
> composed messages?  Posting styles work fine for blank composed messages.
> I've been using Gnus with multiple identities for more than a decade.
> Just move to a group that matches the identity you want to post with, and
> do C-u a, or compose the message while you're reading a group.
>
> I suppose the UI could be a bit better since moving around your folders to
> select the identity may be a bit awkward, but this works quite well for
> me.

Right, that's pretty much what I meant -- if I want to send an email as
Eric Abrahamsen, International Man of Mystery, complete with custom
headers and signature, I don't want to have to navigate to the "Man of
Mystery" group and compose from there. There probably isn't such a
group, and even if there were, that's not really what I *mean*.

We'd have to figure out how these identities would interact with
composing from a group, but that could actually be interesting. A
group's posting styles could say "use identity B when composing from
this group", for example. Or, if you compose from a group with "C-u a",
but also select an identity, the group's posting styles could say
"override header X on whatever identity was selected."

In short, in several cases I have gone through multiple groups and set
the same headers on all those groups' posting styles. It would make more
sense to me to specify that all these groups use the same identity.

Just an idea!

Eric

Re: fancy splitting interactively

[Russ Allbery]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> Right, that's pretty much what I meant -- if I want to send an email as
> Eric Abrahamsen, International Man of Mystery, complete with custom
> headers and signature, I don't want to have to navigate to the "Man of
> Mystery" group and compose from there. There probably isn't such a
> group, and even if there were, that's not really what I *mean*.

Yeah, that makes sense.

> We'd have to figure out how these identities would interact with
> composing from a group, but that could actually be interesting. A
> group's posting styles could say "use identity B when composing from
> this group", for example.

...but this is how Gnus already works now.  Or am I missing something?

> In short, in several cases I have gone through multiple groups and set
> the same headers on all those groups' posting styles. It would make more
> sense to me to specify that all these groups use the same identity.

Oh, I think you're just missing the fact that this is Gnus, so you don't
need to be mucking about in the properties of individual groups and can
just use elisp.  :)

(setq gnus-posting-styles
      '(("."
         (address "eagle@eyrie.org")
         (name "Russ Allbery")
         (organization "The Eyrie")
         (signature-file "~/docs/sigs/eyrie")
         (eval
          (setq gnus-message-archive-group (rra-archive-group "mail"))))

        ((or (string-match "^comp\\.lang\\.perl" gnus-newsgroup-name)
             (string-match "^nnml:project\\.perl" gnus-newsgroup-name))
         (address "rra@cpan.org")
         (signature-file "~/docs/sigs/perl"))

        ((and (string= "news.software.nntp" gnus-newsgroup-name)
              (message-news-p))
         (signature-file "~/docs/sigs/help-newsgroup"))

        ((string= "nnml:project.inn" gnus-newsgroup-name)
         (signature-file "~/docs/sigs/help-list"))))

and so forth.  All matching settings are applied in order, so put more
specific entries later.

https://www.gnu.org/software/emacs/manual/html_node/gnus/Posting-Styles.html

But you're right that mine is based on my current group, so I have to
navigate to the appropriate group to use a particular style.  Although
note the documentation of gnus-posting-styles: you can run arbitrary elisp
to figure out what style to use, among other things.  So if you can figure
out some other piece of information Gnus has available to select on, you
can already use that to determine the posting style.

But I don't think there's an easy way to give you something like Gmail's
drop-down list of identities to pick from right now.

-- 
Russ Allbery (eagle@eyrie.org)              <http://www.eyrie.org/~eagle/>

Re: fancy splitting interactively

[Glyn Millington]

Russ Allbery <eagle@eyrie.org> writes:

> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>
>> Right, that's pretty much what I meant -- if I want to send an email as
>> Eric Abrahamsen, International Man of Mystery, complete with custom
>> headers and signature, I don't want to have to navigate to the "Man of
>> Mystery" group and compose from there. There probably isn't such a
>> group, and even if there were, that's not really what I *mean*.
>
> Yeah, that makes sense.
>
>> We'd have to figure out how these identities would interact with
>> composing from a group, but that could actually be interesting. A
>> group's posting styles could say "use identity B when composing from
>> this group", for example.
>
> ...but this is how Gnus already works now.  Or am I missing something?
>
>> In short, in several cases I have gone through multiple groups and set
>> the same headers on all those groups' posting styles. It would make more
>> sense to me to specify that all these groups use the same identity.
>
> Oh, I think you're just missing the fact that this is Gnus, so you don't
> need to be mucking about in the properties of individual groups and can
> just use elisp.  :)
>
> (setq gnus-posting-styles
>       '(("."
>          (address "eagle@eyrie.org")
>          (name "Russ Allbery")
>          (organization "The Eyrie")
>          (signature-file "~/docs/sigs/eyrie")
>          (eval
>           (setq gnus-message-archive-group (rra-archive-group "mail"))))
>
>         ((or (string-match "^comp\\.lang\\.perl" gnus-newsgroup-name)
>              (string-match "^nnml:project\\.perl" gnus-newsgroup-name))
>          (address "rra@cpan.org")
>          (signature-file "~/docs/sigs/perl"))
>
>         ((and (string= "news.software.nntp" gnus-newsgroup-name)
>               (message-news-p))
>          (signature-file "~/docs/sigs/help-newsgroup"))
>
>         ((string= "nnml:project.inn" gnus-newsgroup-name)
>          (signature-file "~/docs/sigs/help-list"))))
>
> and so forth.  All matching settings are applied in order, so put more
> specific entries later.
>
> https://www.gnu.org/software/emacs/manual/html_node/gnus/Posting-Styles.html
>
> But you're right that mine is based on my current group, so I have to
> navigate to the appropriate group to use a particular style.  Although
> note the documentation of gnus-posting-styles: you can run arbitrary elisp
> to figure out what style to use, among other things.  So if you can figure
> out some other piece of information Gnus has available to select on, you
> can already use that to determine the posting style.
>
> But I don't think there's an easy way to give you something like Gmail's
> drop-down list of identities to pick from right now.


Isn't this what Gnus-alias does? 

http://www.emacswiki.org/emacs/GnusAlias

I've never tried it!


atb



Glyn

Re: fancy splitting interactively

[Eric S Fraga]

On Tuesday,  6 Jan 2015 at 09:25, Eric Abrahamsen wrote:

[...]

> The multiple device problem is trickier. The idea of Gnus-specific marks
> was floated, and Lars was also talking about something called "Gnus
> Cloudy" which might also address this. It's not an area I'm very
> familiar with, and I guess I don't know what the best solution is...

I am not sure what the problem is?  Gnus handles marks understood by
other MUAs properly, as far as I can tell.

My annoyance is that server side splitting is incredibly clumsy (in my
experience) compared with gnus's fancy splitting capabilities and so I
hate having to read emails on other devices with having split anything.

Of course, once the emails have been processed once by gnus, I can
access all the appropriate folders on other devices (assuming IMAP, of
course).

So, again, my view is that other MUAs need to catch up...

Am I missing something?

-- 
: Eric S Fraga, GnuPG: 0xFFFCF67D
: in Emacs 25.0.50.1 + Ma Gnus v0.12 + evil-git-bbc9fa3
: BBDB version 3.1.2 (2014-05-06 11:45:08 -0500)

Re: fancy splitting interactively

[Eric S Fraga]

On Tuesday,  6 Jan 2015 at 18:05, Eric S Fraga wrote:

[...]

> My annoyance is that server side splitting is incredibly clumsy (in my
> experience) compared with gnus's fancy splitting capabilities and so I
> hate having to read emails on other devices with having split anything.

"with" -> "without" (oops)

-- 
: Eric S Fraga, GnuPG: 0xFFFCF67D
: in Emacs 25.0.50.1 + Ma Gnus v0.12 + evil-git-bbc9fa3
: BBDB version 3.1.2 (2014-05-06 11:45:08 -0500)

Re: fancy splitting interactively

[Eric Abrahamsen]

Eric S Fraga <e.fraga@ucl.ac.uk> writes:

> On Tuesday,  6 Jan 2015 at 09:25, Eric Abrahamsen wrote:
>
> [...]
>
>> The multiple device problem is trickier. The idea of Gnus-specific marks
>> was floated, and Lars was also talking about something called "Gnus
>> Cloudy" which might also address this. It's not an area I'm very
>> familiar with, and I guess I don't know what the best solution is...
>
> I am not sure what the problem is?  Gnus handles marks understood by
> other MUAs properly, as far as I can tell.

The OP was talking about the fact that, by default, Gnus won't split
read messages. So if you've read it on your phone or whatever, it won't
get split. Unless you do the zero INBOX thing.

> My annoyance is that server side splitting is incredibly clumsy (in my
> experience) compared with gnus's fancy splitting capabilities and so I
> hate having to read emails on other devices with having split anything.

No argument here! 

> Of course, once the emails have been processed once by gnus, I can
> access all the appropriate folders on other devices (assuming IMAP, of
> course).
>
> So, again, my view is that other MUAs need to catch up...
>
> Am I missing something?

Re: fancy splitting interactively

[Julien Cubizolles]

asjo@koldfront.dk (Adam Sjøgren) writes:


> Seriously, though, what would make Gnus "catch-up"?

I'm also very happy with the way Gnus splits however I was wondering if
someone already hacked a way to automatically construct splitting
rules. Designing them is not very difficult but for most cases, having a
function create-splitting-rule-based-on-sender or
create-splitting-rule-based-on-mailing-list with a few options (move,
copy, mark) would be convenient.

Julien.

Re: fancy splitting interactively

[Eric Abrahamsen]

Julien Cubizolles <j.cubizolles@free.fr> writes:

> asjo@koldfront.dk (Adam Sjøgren) writes:
>
>
>> Seriously, though, what would make Gnus "catch-up"?
>
> I'm also very happy with the way Gnus splits however I was wondering if
> someone already hacked a way to automatically construct splitting
> rules. Designing them is not very difficult but for most cases, having a
> function create-splitting-rule-based-on-sender or
> create-splitting-rule-based-on-mailing-list with a few options (move,
> copy, mark) would be convenient.
>
> Julien.

The Gnus registry might do some of what you want, with split-to-parent
and all that: have you looked at those functions?

Re: fancy splitting interactively

[Adam Sjøgren]

Julien writes:

> I'm also very happy with the way Gnus splits however I was wondering if
> someone already hacked a way to automatically construct splitting
> rules. Designing them is not very difficult but for most cases, having a
> function create-splitting-rule-based-on-sender or
> create-splitting-rule-based-on-mailing-list with a few options (move,
> copy, mark) would be convenient.

I tend to use bbdb for splitting on people.

I.e. using

          (: (lambda ()
               (car (bbdb/gnus-split-method))))

and having gnus-private fields in bbdb.


  Best regards,

    Adam

-- 
 "Den skal ha nr 42 for å gå med mine sko                     Adam Sjøgren
  Men han må ha sin egen pistol"                         asjo@koldfront.dk

Re: fancy splitting interactively

[Eric Abrahamsen]

Russ Allbery <eagle@eyrie.org> writes:

> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>
>> Right, that's pretty much what I meant -- if I want to send an email as
>> Eric Abrahamsen, International Man of Mystery, complete with custom
>> headers and signature, I don't want to have to navigate to the "Man of
>> Mystery" group and compose from there. There probably isn't such a
>> group, and even if there were, that's not really what I *mean*.
>
> Yeah, that makes sense.
>
>> We'd have to figure out how these identities would interact with
>> composing from a group, but that could actually be interesting. A
>> group's posting styles could say "use identity B when composing from
>> this group", for example.
>
> ...but this is how Gnus already works now.  Or am I missing something?
>
>> In short, in several cases I have gone through multiple groups and set
>> the same headers on all those groups' posting styles. It would make more
>> sense to me to specify that all these groups use the same identity.
>
> Oh, I think you're just missing the fact that this is Gnus, so you don't
> need to be mucking about in the properties of individual groups and can
> just use elisp.  :)
>
> (setq gnus-posting-styles
>       '(("."
>          (address "eagle@eyrie.org")
>          (name "Russ Allbery")
>          (organization "The Eyrie")
>          (signature-file "~/docs/sigs/eyrie")
>          (eval
>           (setq gnus-message-archive-group (rra-archive-group "mail"))))
>
>         ((or (string-match "^comp\\.lang\\.perl" gnus-newsgroup-name)
>              (string-match "^nnml:project\\.perl" gnus-newsgroup-name))
>          (address "rra@cpan.org")
>          (signature-file "~/docs/sigs/perl"))
>
>         ((and (string= "news.software.nntp" gnus-newsgroup-name)
>               (message-news-p))
>          (signature-file "~/docs/sigs/help-newsgroup"))
>
>         ((string= "nnml:project.inn" gnus-newsgroup-name)
>          (signature-file "~/docs/sigs/help-list"))))
>
> and so forth.  All matching settings are applied in order, so put more
> specific entries later.

Right, but again this is group-centric, rather than identity-centric,
which conceptually seems like the proper approach, to me.

> https://www.gnu.org/software/emacs/manual/html_node/gnus/Posting-Styles.html
>
> But you're right that mine is based on my current group, so I have to
> navigate to the appropriate group to use a particular style.  Although
> note the documentation of gnus-posting-styles: you can run arbitrary elisp
> to figure out what style to use, among other things.  So if you can figure
> out some other piece of information Gnus has available to select on, you
> can already use that to determine the posting style.
>
> But I don't think there's an easy way to give you something like Gmail's
> drop-down list of identities to pick from right now.

But what you say is true -- we're actually probably very close to having
it, and I don't think it would take a lot of code to get from here to
there. Just a little thought about how to meld identities and posting
styles...

Re: fancy splitting interactively

[Eric Abrahamsen]

Glyn Millington <glyn.millington@gmail.com> writes:

> Russ Allbery <eagle@eyrie.org> writes:
>
>> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>>
>>> Right, that's pretty much what I meant -- if I want to send an email as
>>> Eric Abrahamsen, International Man of Mystery, complete with custom
>>> headers and signature, I don't want to have to navigate to the "Man of
>>> Mystery" group and compose from there. There probably isn't such a
>>> group, and even if there were, that's not really what I *mean*.
>>
>> Yeah, that makes sense.
>>
>>> We'd have to figure out how these identities would interact with
>>> composing from a group, but that could actually be interesting. A
>>> group's posting styles could say "use identity B when composing from
>>> this group", for example.
>>
>> ...but this is how Gnus already works now.  Or am I missing something?
>>
>>> In short, in several cases I have gone through multiple groups and set
>>> the same headers on all those groups' posting styles. It would make more
>>> sense to me to specify that all these groups use the same identity.
>>
>> Oh, I think you're just missing the fact that this is Gnus, so you don't
>> need to be mucking about in the properties of individual groups and can
>> just use elisp.  :)
>>
>> (setq gnus-posting-styles
>>       '(("."
>>          (address "eagle@eyrie.org")
>>          (name "Russ Allbery")
>>          (organization "The Eyrie")
>>          (signature-file "~/docs/sigs/eyrie")
>>          (eval
>>           (setq gnus-message-archive-group (rra-archive-group "mail"))))
>>
>>         ((or (string-match "^comp\\.lang\\.perl" gnus-newsgroup-name)
>>              (string-match "^nnml:project\\.perl" gnus-newsgroup-name))
>>          (address "rra@cpan.org")
>>          (signature-file "~/docs/sigs/perl"))
>>
>>         ((and (string= "news.software.nntp" gnus-newsgroup-name)
>>               (message-news-p))
>>          (signature-file "~/docs/sigs/help-newsgroup"))
>>
>>         ((string= "nnml:project.inn" gnus-newsgroup-name)
>>          (signature-file "~/docs/sigs/help-list"))))
>>
>> and so forth.  All matching settings are applied in order, so put more
>> specific entries later.
>>
>> https://www.gnu.org/software/emacs/manual/html_node/gnus/Posting-Styles.html
>>
>> But you're right that mine is based on my current group, so I have to
>> navigate to the appropriate group to use a particular style.  Although
>> note the documentation of gnus-posting-styles: you can run arbitrary elisp
>> to figure out what style to use, among other things.  So if you can figure
>> out some other piece of information Gnus has available to select on, you
>> can already use that to determine the posting style.
>>
>> But I don't think there's an easy way to give you something like Gmail's
>> drop-down list of identities to pick from right now.
>
>
> Isn't this what Gnus-alias does? 
>
> http://www.emacswiki.org/emacs/GnusAlias

Yes, I mentioned very package early on, as doing pretty much exactly
what I had in mind.

Does anyone have any particular opinions about maybe making that part of
message.el?

Eric

Re: fancy splitting interactively

[Julien Cubizolles]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:


> The Gnus registry might do some of what you want, with split-to-parent
> and all that: have you looked at those functions?

I didn't think of it this way, I'll give it a try.

Re: fancy splitting interactively

[Julien Cubizolles]

asjo@koldfront.dk (Adam Sjøgren) writes:


> I tend to use bbdb for splitting on people.
>
> I.e. using
>
>           (: (lambda ()
>                (car (bbdb/gnus-split-method))))
>
> and having gnus-private fields in bbdb.

Never tried it, will do, thanks.

Re: [bug]

[Lars Ingebrigtsen]

Uwe Brauer <oub@mat.ucm.es> writes:

> The following bug showed up when Using B r 
>
> Debugger entered--Lisp error: (wrong-type-argument char-or-string-p nil)
>   decode-coding-string(nil utf-8)
>   nnimap-decode-gnus-group(nil)
>   nnimap-request-accept-article(nil "UCMgmail" t)
>   gnus-request-accept-article(nil (nnimap "UCMgmail" (nnimap-address
> "imap.gmail.com") (nnimap-server-port 993) (nnimap-authinfo-file
> "~/.authinfo") (nnimap-stream ssl)) t t)

I didn't get exactly the same backtrace, but it failed for me, too.
I've now fixed this in the Gnus git repository.

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

Re: fancy splitting interactively

[Julien Cubizolles]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:


> The Gnus registry might do some of what you want, with split-to-parent
> and all that: have you looked at those functions?

I finally got around it, and it's working fine, thanks.