Gnus development mailing list
 help / color / mirror / Atom feed
* Adding parameters to gnus-select-method causes madness
@ 2001-11-23 21:09 Karl Kleinpaste
  2001-11-24  0:42 ` Harry Putnam
  0 siblings, 1 reply; 2+ messages in thread
From: Karl Kleinpaste @ 2001-11-23 21:09 UTC (permalink / raw)


I have nnml as my primary method:

	(setq gnus-select-method '(nnml ""))

Fine so far.  I am looking into adding multiple-backend support to
nnir, so I want to add a parameter by modifying this definition to:

	(setq gnus-select-method '(nnml ""
					(nnir-search-engine 'swish++)))

because I simply set nnir's backend to gnus-select-method.

I start Gnus, and observe that Gnus now rediscovers every single nnml
group I've got, subscribes them at default level 3, abandons them
loosely inside the top level of my topic list, and `G E' shows them
having a complete description of:

        ("nnml:foo.bar" 3 nil nil
         (nnml ""))

Note that there is a bogus prefix "nnml:" on the group; this shouldn't
have happened at all, considering that this is my primary method --
method descriptors are only supposed to be included in the name for
secondary and foreign groups.

The original groups, at varying group levels and with lots of group
parameters and other information engaged, are still present.  These
"new" groups are information-free clones.  They're wrong, just plain
wrong.

Even better, go to one of the *old* groups and hit `G E'.  One sees
that it has all the appropriate stuff (what's been read, reply marks,
the works).  Again, fine so far.  Now, modify nothing, and hit `C-c
C-c' to exit...

...and the old group *disappears*, while the new group persists,
lacking all the good information that used to be part of the old.
Also, the new group now lacks the "nnml:" prefix.

*whimper*  kill -9 ...

(kill -9 several times, in fact, as I experimented with what was
happening.  Start Gnus, find something horrible happening, kill -9,
modify .gnus, start again, find something dreadful, kill -9 ...)

Now, I swear, all I did was add a server parameter.  I don't have the
first understanding why this should cause Gnus to fail to comprehend
that the groups are no different.  It didn't mark the old groups with
`*' to indicate possibly bogus content; it just cloned them all,
badly.  Gnus is so confused about the clones that it can't quite
differentiate them from the originals, once one edits their
descriptions with `G E'.  It seems that there is a very bad "eq"
comparison of servers occurring.

--karl



^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Adding parameters to gnus-select-method causes madness
  2001-11-23 21:09 Adding parameters to gnus-select-method causes madness Karl Kleinpaste
@ 2001-11-24  0:42 ` Harry Putnam
  0 siblings, 0 replies; 2+ messages in thread
From: Harry Putnam @ 2001-11-24  0:42 UTC (permalink / raw)


Karl Kleinpaste <karl@charcoal.com> writes:

> Now, I swear, all I did was add a server parameter.  I don't have the
> first understanding why this should cause Gnus to fail to comprehend
> that the groups are no different.  It didn't mark the old groups with

You've probably checked this already but recently I had a major change
suddenly happen and it was because gnus had taken a notion to rename
.newsrc.eld to .newsrc.SERVER.eld



^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2001-11-24  0:42 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-11-23 21:09 Adding parameters to gnus-select-method causes madness Karl Kleinpaste
2001-11-24  0:42 ` Harry Putnam

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).