Group list functions (AM AA etc)

Group list functions (AM AA etc)

[Harry Putnam]

Complaint:
Seeming brokenness with Group list commands:
`A M'  does not list all matching groups from active file,
After `A A' (or g on server in server buffer) and then using `A M'
there seems to be no way to recover the full active list.  It
disappears.  None of `L', `C-u x' `A M . <RET>' or `C-x C-b'
can recover it.  

Diagnostic details:

Starting todays cvs with emacs -q -no-site-file -l ~/path.el -f gnus &
  (where path.el is nothing but the path to cvs gnus.)

cat path.el:
(add-to-list 'load-path  "/usr/local/gnus_04/lisp")
(add-to-list 'Info-default-directory-list "/usr/local/gnus_04/texi")

In group buffer press `A A' to download the full active list

Once downloaded press `A M comp\.unix' <RET>  The buffer is
immediately limited to only my *SUBSCRIBED* groups with that match.

The documentation seemingly draws a distinction between A M and A m

`A m'
     List all unread, subscribed groups with names that match a regexp
     (`gnus-group-list-matching').

`A M'
     List groups that match a regexp (`gnus-group-list-all-matching').
 
>From that I would surmise the `A M' is designed to work against `all'
groups since the keyword `subscribed' is not present.

C-h f gnus-group-list-all-matching <RET> give some confusing language
about LEVEL and LOWEST.

========================================
gnus-group-list-all-matching is an interactive compiled Lisp function
in `gnus-group'.
(gnus-group-list-all-matching LEVEL REGEXP &optional LOWEST)

List all groups that match REGEXP.
If the prefix LEVEL is non-nil, it should be a number that says which
level to cut off listing groups.
If LOWEST, don't list groups with level lower than LOWEST.

========================================

I guess this means `C-u 9  A M' should match against all and 
`C-u 1 A M' would match against only groups at level 1.

(Maybe this is a place for a clean example rather than Jargonized
programmer talk.  Since it is a common user command.)  And the info
doc doesn't mention this at all.

At any rate neither of the above do any better.  And in every usage of
`A M' the full list buffer is truncated and not recoverable by any
thing I could think of.

Thats a pretty big buffer to have to download repeatedly.  Surely
there is some way to maintain it for more than one search?  Am I just
overlooking some obvious recovery thing.

So, at least `A a' 

`A a'
     List all groups that have names that match a regexp
     (`gnus-group-apropos').

Here the behaviour is what one would expect from the description.  And
the active file buffer is left in tact.

So what is `A M' supposed to do?

Re: Group list functions (AM AA etc)

[Lars Magne Ingebrigtsen]

Harry Putnam <reader@newsguy.com> writes:

> Seeming brokenness with Group list commands:
> `A M'  does not list all matching groups from active file,
> After `A A' (or g on server in server buffer) and then using `A M'
> there seems to be no way to recover the full active list.  It
> disappears.  None of `L', `C-u x' `A M . <RET>' or `C-x C-b'
> can recover it.  

I've just tried the following.

`A A' and then  `A M alt RET'.

I got lots and lots of killed newsgroups listed that matched "alt".
So I don't seem to be able to reproduce this bug.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Group list functions (AM AA etc)

[Harry Putnam]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Harry Putnam <reader@newsguy.com> writes:
> 
> > Seeming brokenness with Group list commands:
> > `A M'  does not list all matching groups from active file,
> > After `A A' (or g on server in server buffer) and then using `A M'
> > there seems to be no way to recover the full active list.  It
> > disappears.  None of `L', `C-u x' `A M . <RET>' or `C-x C-b'
> > can recover it.  
> 
> I've just tried the following.
> 
> `A A' and then  `A M alt RET'.
> 
> I got lots and lots of killed newsgroups listed that matched "alt".
> So I don't seem to be able to reproduce this bug.

Were you able to return to a full `A A' listing after limiting to the
alt subset?

How does `A a' differ from `A M'.
Should I have the full active list in killed list?  
Do you?
Were any Z groups shown in your display.  Did you have any `alt'
groups subscribed?

All I get are subsribed groups from `A A' followed by `A M comp\.unix'
After <RET> the buffer is limited to my three Comp.* groups then there
is no way to return to the full `A A' buffer.

Re: Group list functions (AM AA etc)

[Lars Magne Ingebrigtsen]

Harry Putnam <reader@newsguy.com> writes:

> Were you able to return to a full `A A' listing after limiting to the
> alt subset?

Yup.

> How does `A a' differ from `A M'.

It also looks a group descriptions.

> Should I have the full active list in killed list?  
> Do you?

No, that depends on your setting of gnus-save-killed-list.  Mine is t.

> Were any Z groups shown in your display.  Did you have any `alt'
> groups subscribed?

Yes, yes.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Group list functions (AM AA etc)

[Harry Putnam]

Harry Putnam <reader@newsguy.com> writes:

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> 
> > Harry Putnam <reader@newsguy.com> writes:
> > 
> > > Were you able to return to a full `A A' listing after limiting to the
> > > alt subset?
> > 
> > Yup.
> 
> How?
> 
> [...]

Harry:
> > How does `A a' differ from `A M'.

Lars:
> It also looks a group descriptions.

Isn't that `A d' or do both `A a' and `A d' look at descriptions?

[...]

> > No, that depends on your setting of gnus-save-killed-list.  Mine is t.
> 
> Ditto .. here
> 
> > > Were any Z groups shown in your display.  Did you have any `alt'
> > > groups subscribed?
> > 
> > Yes, yes.

So apparently something local but with a vanilla startup after 
rm -rf  .newsrc.eld  News Mail

I get full `A A' buffer and nothing with `A M alt <RET>'

=================
Lars said:
> I've just tried the following.

> `A A' and then  `A M alt RET'.

> I got lots and lots of killed newsgroups listed that matched "alt".
> So I don't seem to be able to reproduce this bug.
================

Will that still work if you rm -f .newsrc.eld?  It should shouldn't
it?


Maybe that rm -f .newsrc.eld is the difference (no killed list)?

But after  `A A' Shouldn't `A M alt <RET> still display all the alt
groups that are present in `A A' generated buffer.  Even without an 
extensive killed list in .newsrc.eld?

M-x occur would do here, of course, but I'm wondering where `A M' gets
its information.  Not from the `A A' buffer it seems.  Where as
`A a alt <RET>' shows the massive number of alt groups in a separate
buffer with out affecting the `A A' generated buffer.  Even if
.newrc.eld is non-existent.