Re: What's a killed article ?

[Francis Moreau]

Ted Zlatanov <tzz@lifelogs.com> writes:

> On Thu, 18 Dec 2008 21:00:52 +0100 harven <harven@free.fr> wrote: 
>
> h> From the top of my head, some terms, either ambiguous or defined too
> h> late in the manual
>
> (note the below are thoughts that can go into the manual eventually, but
> I think it's too early)
>
> h> - MAIL BACK END (info "(gnus)Comparing mail back ends") The list of
> h>   all available mail backends should be given sooner The following
> h>   list appears in 1.5.3 (`nnml', `nnbabyl', `nnfolder',
> h>   `nnmbox',`nnmh', and `nnmaildir') but nnimap is also mentioned in
> h>   3.22 and others are alluded to at 6.3.13 It would be nice if real
> h>   life config examples were given for each of them and if their
> h>   different features and limitations were summarized somewhere.
>
> I think a full backend list is useful and should probably be referenced
> at the front of the manual ("here's what Gnus can do...") but it should
> not live there.  OTOH, the full list of protocols that Gnus can handle
> is urgently needed by new users and should probably be in the first
> section (with references to the later chapters that define them).  I
> agree about configuration examples too.
>
> "back end" is defined early in the Terminology appendix, by the way.
> That's where updated definitions we discuss here will go.  Do you think
> the definition is poor?
>
> h> For example, I tried to use nnmbox to browse some archive and got the
> h> following: Article mail.misc:1190 out of order. This seems specific
> h> to that backend, and still mysterious for me.
>
> I don't know what causes that, I don't use nnmbox, sorry.
>
> h> - IDLE (info "(gnus)Daemons")
> h>   what does exactly idle means ? Can really emacs be idle ? 
> h>   Is emacs idle between that short time separating two keystrokes ? 
> h>   Fetching mail freezes emacs, some warning and comments 
> h>   would be welcome here.
>
> Unfortunately "IDLE" has an IMAP meaning too.  Specifically for Daemons,
> "idle" means that Emacs is not busy running ELisp code.  Emacs can
> indeed be idle, and it can do things automatically at those times.  It's
> not idle between consecutive ELisp function calls, but it's idle between
> two keystrokes (font-lock uses that, for example).
>
> Asynchronous mail retrieval is indeed a badly needed feature.  What
> should we say/promise/warn?  It's basically a statement of "we know
> Emacs as a whole freezes while you get your mail, sorry."
>
> h> - GNUS-SECONDARY-SELECT-METHODS
> h>   and gnus-select-method. A single select method but many secondary
> h>   methods ? Why  differentiate ? 
>
> For better or worse, Gnus has made the decision that the primary select
> method will have unqualified group names.  Thus, if you have a group
> "mail", it really refers to "primary_select_method:mail".  All secondary
> select methods are full (including the select method's name).  I think
> it's best to use nnil as the primary and have all the groups look the
> same, fully qualified, but I'm sure many people like the brevity of the
> primary select method.

I agree too.

>From what I undestood, primary select method does not have qualified
group names because if you decide to change the primary method, the
groups will use this new method automatically. But since articles
numbers are not the same on different nntp servers, it's going to
break a lot of things. So changing the primary method is not
something I should do anyway.

With secondary select methods, the above doesn't seem to be possible.
-- 
Francis