weird results when setting gnus-parameters

weird results when setting gnus-parameters

[lee]

Hi,

setting gnus-parameters in my ~/.gnus yields weird results. For example:


(setq gnus-parameters
      `(
	("mail\\.0-incoming"
	 (gnus-show-threads nil)
	 (gnus-use-scoring nil)
	 (gnus-article-sort-functions '((not gnus-article-sort-by-date)))
	 (display . [not expire])
	 (gcc-self . "mail.sent")
	 (gnus-summary-line-format
	  ":%U%R%z%I(%5k) %(%[%-23,23A%]%) %11&user-date; '%s\n"))
	;;
	("mail\\.lists\\.debian.*"
	 (subscribed . t)
	 (gnus-show-threads t)
	 (gnus-use-scoring  t)
	 (display . default)
	 (gcc-self . "mail.posted")
	 (posting-style
	  (organization "my virtual residence")
	  (signature "Debian testing amd64"))
	 (gnus-summary-line-format
	  ":%U%R%z%B(%5k) %(%[%-23,23f%]%) '%s\n"))))


This does ignore the signature specified in the posting-style for a
group called "mail.lists.debian-user". It does apply the Gcc header
correctly. Why is the signature setting ignored?

It also seems to (sometimes) ignore (display . default), as gnus is
asking me how many articles to get when entering some group that should
have that setting enabled.

The documentation says:


,---- [ info (gnus) Group Parameters ]
|    All clauses that matches the group name will be used, but the last
| setting "wins".  So if you have two clauses that both match the group
| name, and both set, say `display', the last setting will override the
| first.
`----


It definitely doesn't work that way. I have done some experimentation
with the order of entries in gnus-parameters and started thinking that
either not the last setting wins but the first one or that the first
setting that matches wins. However, the results then contradicted that
and are inconclusive, and I can't figure out how it works.

What I expected, according to the documentation, is that I could set
parameters for all groups first and then create "exceptions" in later
clauses, like:


(setq gnus-parameters
      `(
	(".*"
	 (gnus-show-threads nil)
	 (gnus-use-scoring nil)
	 (gnus-article-sort-functions '((not gnus-article-sort-by-date)))
	 (display . [not expire])
	 (gcc-self . "mail.sent")
	 (gnus-summary-line-format
	  ":%U%R%z%I(%5k) %(%[%-23,23A%]%) %11&user-date; '%s\n"))
	;;
	("mail\\.lists.*"
	 (subscribed . t)
	 (gnus-show-threads t)
	 (gnus-use-scoring t)
	 (display . default)
	 (gcc-self . "mail.posted")
	 (posting-style
	  (organization "my virtual residence"))
	 (gnus-summary-line-format
	  ":%U%R%z%B(%5k) %(%[%-23,23f%]%) '%s\n"))
        ;;
	("mail\\.lists\\.debian.*"
	 (subscribed . t)
	 (gnus-show-threads t)
	 (gnus-use-scoring  t)
	 (display . default)
	 (gcc-self . "mail.posted")
	 (posting-style
	  (organization "my virtual residence")
	  (signature "Debian testing amd64"))
	 (gnus-summary-line-format
	  ":%U%R%z%B(%5k) %(%[%-23,23f%]%) '%s\n"))))


It doesn't really work that way. Is there a bug with this, or how does
it actually work?


-- 
http://www.asciiribbon.org/
http://tools.ietf.org/html/rfc1855
http://www.caliburn.nl/topposting.html

Re: weird results when setting gnus-parameters

[Tassilo Horn]

lee <lee@yun.yagibdah.de> writes:

Hi!

> The documentation says:
>
> ,---- [ info (gnus) Group Parameters ]
> |    All clauses that matches the group name will be used, but the last
> | setting "wins".  So if you have two clauses that both match the group
> | name, and both set, say `display', the last setting will override the
> | first.
> `----
>
> It definitely doesn't work that way. I have done some experimentation
> with the order of entries in gnus-parameters and started thinking that
> either not the last setting wins but the first one or that the first
> setting that matches wins.

Indeed, the way posting styles and group parameters were applied used to
be not exactly as documented.  This has been changed not long time ago.
Maybe that fix isn't in Gnus 5.13, so you might want to try out Gnus
from git.

Bye,
Tassilo

Re: weird results when setting gnus-parameters

[lee]

Tassilo Horn <tsdh@gnu.org> writes:

> lee <lee@yun.yagibdah.de> writes:
>
> Hi!
>
>> The documentation says:
>>
>> ,---- [ info (gnus) Group Parameters ]
>> |    All clauses that matches the group name will be used, but the last
>> | setting "wins".  So if you have two clauses that both match the group
>> | name, and both set, say `display', the last setting will override the
>> | first.
>> `----
>>
>> It definitely doesn't work that way. I have done some experimentation
>> with the order of entries in gnus-parameters and started thinking that
>> either not the last setting wins but the first one or that the first
>> setting that matches wins.
>
> Indeed, the way posting styles and group parameters were applied used to
> be not exactly as documented.  This has been changed not long time ago.
> Maybe that fix isn't in Gnus 5.13, so you might want to try out Gnus
> from git.

Awww that's what I was trying to avoid ... Maybe there is no good way
around that when you want to use gnus. Hmmm ...

Re: weird results when setting gnus-parameters

[Tassilo Horn]

lee <lee@yun.yagibdah.de> writes:

>> Indeed, the way posting styles and group parameters were applied used
>> to be not exactly as documented.  This has been changed not long time
>> ago.  Maybe that fix isn't in Gnus 5.13, so you might want to try out
>> Gnus from git.
>
> Awww that's what I was trying to avoid ... Maybe there is no good way
> around that when you want to use gnus. Hmmm ...

You could also get the latest Emacs release 24.2.  I'm pretty sure the
Gnus version bundled with that is recent enough.

Bye,
Tassilo

Re: weird results when setting gnus-parameters

[lee]

Tassilo Horn <tsdh@gnu.org> writes:

> lee <lee@yun.yagibdah.de> writes:
>
>>> Indeed, the way posting styles and group parameters were applied used
>>> to be not exactly as documented.  This has been changed not long time
>>> ago.  Maybe that fix isn't in Gnus 5.13, so you might want to try out
>>> Gnus from git.
>>
>> Awww that's what I was trying to avoid ... Maybe there is no good way
>> around that when you want to use gnus. Hmmm ...
>
> You could also get the latest Emacs release 24.2.  I'm pretty sure the
> Gnus version bundled with that is recent enough.

Ok, I'll just get the git version and go with that. There's not much
point in being so lazy ... :)


-- 
http://www.asciiribbon.org/
http://tools.ietf.org/html/rfc1855
http://www.caliburn.nl/topposting.html

Re: weird results when setting gnus-parameters

[lee]

lee <lee@yun.yagibdah.de> writes:

> Tassilo Horn <tsdh@gnu.org> writes:
>
>> lee <lee@yun.yagibdah.de> writes:
>>
>>>> Indeed, the way posting styles and group parameters were applied used
>>>> to be not exactly as documented.  This has been changed not long time
>>>> ago.  Maybe that fix isn't in Gnus 5.13, so you might want to try out
>>>> Gnus from git.
>>>
>>> Awww that's what I was trying to avoid ... Maybe there is no good way
>>> around that when you want to use gnus. Hmmm ...
>>
>> You could also get the latest Emacs release 24.2.  I'm pretty sure the
>> Gnus version bundled with that is recent enough.
>
> Ok, I'll just get the git version and go with that. There's not much
> point in being so lazy ... :)

Ok I have the git version now :) Two problems:

How do I get bbdb to work with it?

The gnus-compile function in my ~/.gnus doesn't seem to exist
anymore. The debug window comes up and says it's a void function. Has it
been removed?


-- 
http://www.asciiribbon.org/
http://tools.ietf.org/html/rfc1855
http://www.caliburn.nl/topposting.html

Re: weird results when setting gnus-parameters

[Tassilo Horn]

lee <lee@yun.yagibdah.de> writes:

>> Ok, I'll just get the git version and go with that. There's not much
>> point in being so lazy ... :)
>
> Ok I have the git version now :) Two problems:
>
> How do I get bbdb to work with it?

Sorry, I don't use it.  If you get an error, what is it?

> The gnus-compile function in my ~/.gnus doesn't seem to exist
> anymore. The debug window comes up and says it's a void function. Has
> it been removed?

Yes, it has been removed.  Benchmarking showed that byte-compilation of
~/.gnus doesn't speed up things anyway.

Bye,
Tassilo

Re: weird results when setting gnus-parameters

[lee]

Tassilo Horn <tsdh@gnu.org> writes:

> lee <lee@yun.yagibdah.de> writes:
>
>>> Ok, I'll just get the git version and go with that. There's not much
>>> point in being so lazy ... :)
>>
>> Ok I have the git version now :) Two problems:
>>
>> How do I get bbdb to work with it?
>
> Sorry, I don't use it.  If you get an error, what is it?

It cannot be found. The load-path is set correctly to the emacs24
directories under /usr/local, and I've had "(require 'bbdb)" in my
~/.emacs. There are files apparently having to do with bbdb among those
"make install" puts under /usr/local, yet there is no file "bbdb.el" or
"bbdb.elc". --- I still have emacs23 installed, which comes with bbdb
(as in /usr/share/emacs23/site-lisp/bbdb/bbdb.el etc.).

So I guess bbdb isn't part of emacs. I pulled bbdb from the git
repository, and the documentation that comes with it says you need
something like "(require 'bbdb-loaddefs "/path/to/bbdb-loaddefs.el")" in
your ~/.emacs to enable it. --- I put that in, and it's working
now. It's a makeshift solution, though, because I have bbdb in my home
directory rather than /usr/local/share because I'm not sure where it
should actually be when it's under /usr/local.

The info docs of gnus still say "(require 'bbdb)", btw. Perhaps they
need an update?

>> The gnus-compile function in my ~/.gnus doesn't seem to exist
>> anymore. The debug window comes up and says it's a void function. Has
>> it been removed?
>
> Yes, it has been removed.  Benchmarking showed that byte-compilation of
> ~/.gnus doesn't speed up things anyway.

Cool --- and thanks for all your help :)


-- 
http://www.asciiribbon.org/
http://tools.ietf.org/html/rfc1855
http://www.caliburn.nl/topposting.html