"subscribed" has overloaded meaning in gnus?

"subscribed" has overloaded meaning in gnus?

[Benjamin Rutt]

I'm using a couple month-old Oort 0.04.  I'm a little confused as to
what 'subscribed' means as a group parameter.  To me, the concept of
'subscribed' in gnus means that you have used the command 'U' or 'u'
in the *Group* buffer on a group to subscribe it.  But it seems from
the following definition that 'subscribed' takes on another meaning:

,----[ from gnus.el ]
| (defun gnus-find-subscribed-addresses ()
|   "Return a regexp matching the addresses of all subscribed mail groups.
| It consists of the `to-address' or `to-list' parameter of all groups
| with a `subscribed' parameter."
`----

If 'subscribed' is a group parameter, then it should be documented in
the info node "(gnus)Group Parameters".  (Is there any other source of
a list of group parameters that I'm missing?)  I think the prior
meaning of 'subscribed' as a *Group* buffer concept is pretty well
entrenched, and this new meaning may confuse others as it has me.
Maybe the group parameter 'subscribed' should have its name changed to
avoid the confusion from the overloaded meanings?  It's also possible
that I'm just being picky, and everyone can understand the two
meanings without any problems.  Opinions?

BTW, I learned about 'subscribed' as a group parameter a few months
ago in this group during a discussion of MFT.  The following sets
`message-subscribed-regexps' to a regexp matching all my mailing lists
(I've named them all with the prefix "mlist-" and I've set 'to-list'
and 'to-address' on each one).  If there's a better way to do this
please let me know:

(setq gnus-parameters '(("^nnml:mlist-.*" (subscribed . t))))
(add-hook 'gnus-started-hook
	  (lambda ()
	    (setq message-subscribed-regexps
		  (gnus-find-subscribed-addresses))))

Thanks,
-- 
Benjamin

Re: "subscribed" has overloaded meaning in gnus?

[Matt Armstrong]

Benjamin Rutt <rutt+news@cis.ohio-state.edu> writes:

> I'm using a couple month-old Oort 0.04.  I'm a little confused as to
> what 'subscribed' means as a group parameter.

Yes I think it should be documented.  There was a recent post to
'ding' that documented the MFT stuff.  I don't know if this parameter
was documented there or not.

The group parameter "subscribed" simply means "consider the to-address
or to-list parameter of this group to be that of a mailing list you're
subscribed to."  The only place it is currently used in Gnus is in
gnus-find-subscribed-addresses.

Perhaps the parameter should be "mailing-list-subscribed" or
"address-subscribed" or "to-addresses-subscribed" something else?


> (setq gnus-parameters '(("^nnml:mlist-.*" (subscribed . t))))
> (add-hook 'gnus-started-hook
> 	  (lambda ()
> 	    (setq message-subscribed-regexps
> 		  (gnus-find-subscribed-addresses))))

Instead of the hook, I think this ended up being the "preferred" way
of doing it.

    (setq message-subscribed-address-functions
          '(gnus-find-subscribed-addresses))

Your method is technically wrong since gnus-find-subscribed-addresses
does not return a set of regexps (but it'll mostly work).

-- 
matt

Re: "subscribed" has overloaded meaning in gnus?

[Benjamin Rutt]

"Matt Armstrong" <matt+dated+1010595772.908bba@lickey.com> writes:

> Benjamin Rutt <rutt+news@cis.ohio-state.edu> writes:
>
>> I'm using a couple month-old Oort 0.04.  I'm a little confused as to
>> what 'subscribed' means as a group parameter.
>
> Yes I think it should be documented.  There was a recent post to
> 'ding' that documented the MFT stuff.  I don't know if this parameter
> was documented there or not.

It wasn't.  Maybe your "preferred" snippet below should be included in
that documentation, since I think it is a nice way to get your MFT
working without duplicating your list of mailing lists addresses in
another place.

> The group parameter "subscribed" simply means "consider the to-address
> or to-list parameter of this group to be that of a mailing list you're
> subscribed to."  The only place it is currently used in Gnus is in
> gnus-find-subscribed-addresses.
>
> Perhaps the parameter should be "mailing-list-subscribed" or
> "address-subscribed" or "to-addresses-subscribed" something else?

Based on your definition, I would vote for "mailing-list-subscribed".

> Instead of the hook, I think this ended up being the "preferred" way
> of doing it.
>
>     (setq message-subscribed-address-functions
>           '(gnus-find-subscribed-addresses))

Thanks, that looks much better.  I think this should go in the
recently posted MFT docs, as noted above.

> Your method is technically wrong since gnus-find-subscribed-addresses
> does not return a set of regexps (but it'll mostly work).

Hmm.  It does return a list on mine (I'm using >month-old Oort 0.04),
containing a single regexp.  Maybe that's why it has been working for
me.
-- 
Benjamin

Re: "subscribed" has overloaded meaning in gnus?

[Sriram Karra]

Benjamin Rutt <rutt+news@cis.ohio-state.edu> writes:

> "Matt Armstrong" <matt+dated+1010595772.908bba@lickey.com> writes:
>
>> Benjamin Rutt <rutt+news@cis.ohio-state.edu> writes:
>>
>>> I'm using a couple month-old Oort 0.04.  I'm a little confused as to
>>> what 'subscribed' means as a group parameter.
>>
>> Yes I think it should be documented.  There was a recent post to
>> 'ding' that documented the MFT stuff.  I don't know if this parameter
>> was documented there or not.
>
> It wasn't.  Maybe your "preferred" snippet below should be included
> in that documentation, since I think it is a nice way to get your
> MFT working without duplicating your list of mailing lists addresses
> in another place.

[ I had sent that MFT doc... ]

This is what I will do: I will document the "subscribed" parameter in
the Group Parameters node and include a cross reference or something
in the MFT docs node.

Also, it will take a while before the changes will appear in CVS...

-- 
"Have you ever fought an IDEA, Picard?"
      -- Gowron, Star Trek TNG.

Re: "subscribed" has overloaded meaning in gnus?

[Benjamin Rutt]

Sriram Karra <karra@cs.utah.edu> writes:

> [ I had sent that MFT doc... ]
>
> This is what I will do: I will document the "subscribed" parameter in
> the Group Parameters node and include a cross reference or something
> in the MFT docs node.

Sounds like an excellent idea.  Hopefully that will make it clearer to
everyone what the double-meaning of subscribed is.

> Also, it will take a while before the changes will appear in CVS...

I'm guessing you're talking about signing papers.  I am eagerly
awaiting snail mail papers to arrive in my mailbox from GNU for a gnus
contribution (gnus-dired.el) of my own.  Hopefully we won't have to
wait too much longer, but it's been almost a month for me already.
-- 
Benjamin