view etc/gnus-tut.txt @ 31384:f6cb7dfe5e7c

(vc-cvs-show-log-entry): New function. (vc-cvs-stay-local): Allow it to be a hostname regexp as well. (vc-cvs-remote-p): Renamed to vc-cvs-stay-local-p. Handle hostname regexps. Updated all callers. (vc-cvs-responsible-p): Handle directories as well. (vc-cvs-could-register): New function. (vc-cvs-retrieve-snapshot): Parse "cvs update" output, keep file properties up-to-date. (vc-cvs-checkout): Do the right thing when the workfile does not exist. (vc-cvs-registered): Use new function vc-cvs-parse-entry to do the actual work. (vc-cvs-remote-p): Allow FILE to be a directory, too. (vc-cvs-dir-state): New function. (vc-cvs-dir-state-heuristic): New function, subroutine of the above. (vc-cvs-parse-entry): New function, also to be used in vc-cvs-registered. (vc-cvs-checkout): Slight restructuring to make the control-flow more clear and to avoid running `cvs' twice. (vc-cvs-workfile-version): Removed comment that this is not reached. It is. (vc-cvs-merge): Set state to 'edited after merge. (vc-cvs-merge-news): Set workfile version to nil if not known. (vc-cvs-latest-on-branch-p): Recommented. Candidate for removal. (vc-cvs-checkin): Raise the max-correct status from 0 to 1. Make sure to switch to *vc* before looking for an error message. Use vc-parse-buffer. (vc-cvs-create-snapshot): Swap DIR and NAME. (vc-cvs-retrieve-snapshot): New function (untested). (vc-cvs-stay-local): Default to t. (vc-cvs-remote-p): New function and property. (vc-cvs-state): Stay local only if the above is t. (vc-handle-cvs): Removed. (vc-cvs-registered): Don't check vc-handle-cvs -- it should all be done via vc-handled-backends now. (vc-cvs-header): Escape Id. (vc-cvs-state, vc-cvs-fetch-status): Use with-temp-file. Use the new BUFFER=t argument to vc-do-command. (vc-cvs-print-log, vc-cvs-diff): Insert in the current buffer. (vc-cvs-state): Use vc-do-command instead of vc-simple-command. (vc-cvs-diff): Remove unused and unsupported argument CMP. (vc-cvs-registered): Obey vc-handle-cvs. (vc-cvs-registered): Use with-temp-buffer. Reorder extraction of fields and call to file-attributes because of a temporary bug in rcp.el. (vc-cvs-fetch-status): Use with-current-buffer. Merge in code from vc-cvs-hooks.el. (proto vc-cvs-registered): Require 'vc-cvs instead of 'vc-cvs-hooks. Don't require 'vc anymore. (vc-cvs-responsible-p): Use expand-file-name instead of concat and file-directory-p instead of file-exists-p. (vc-cvs-create-snapshot): New function, replacing vc-cvs-assign-name. (vc-cvs-assign-name): Remove. (vc-cvs-header): New var. Update Copyright. (vc-cvs-diff): Remove unused `backend' variable. (vc-cvs-checkout): Only toggle read-only if the buffer is setup right. (tail): Provide vc-cvs. (vc-cvs-merge-news, vc-cvs-checkout): Removed call to vc-file-clear-masterprops. (vc-cvs-state): Typo. (vc-cvs-merge-news): Return the status code rather than the error msg. (vc-cvs-state): Don't overwrite a non-heuristic state with a heuristic one. (vc-cvs-merge-news): Just use 'edited for the case with conflicts. (vc-cvs-checkin): Do a trivial parse to set the state in case of error. That allows us to get to 'needs-merge even in the stay-local case. There's still no way to detect 'needs-patch in such a setup (or to force an update for that matter). (vc-cvs-logentry-check): Remove, the default works as well. (vc-cvs-print-log, vc-cvs-diff): Run cvs asynchronously. (vc-cvs-stay-local): Renamed from vc-cvs-simple-toggle. Redocumented. (vc-cvs-state): If locality is wanted, use vc-cvs-state-heuristic. (vc-cvs-toggle-read-only): Removed. (for compiler warnings). (vc-cvs-release, vc-cvs-system-release): Remove. (vc-cvs-use-edit, vc-cvs-simple-toggle): New config variables. (vc-cvs-dired-state-info): Use `cvs-state' and slightly different status symbols. (vc-cvs-parse-status, vc-cvs-state): Move from vc-cvs-hooks.el. (vc-cvs-toggle-read-only): First cut at a function to allow a cvs-status-free vc-toggle-read-only. (vc-cvs-merge-news): Move from cvs-merge-news in vc.el. (vc-cvs-checkin): Use vc-recompute-state+vc-state instead of vc-cvs-status. Also set vc-state rather than vc-locking-user. (vc-cvs-checkout): Modify access rights directly if the user requested not to use `cvs edit'. And refresh the mode line. (if workfile' that got lost when the code was extracted from vc.el. And merged the tail with the rest of the code (not possible in the old vc.el where the tail was shared among all backends). And explicitly set the state to 'edited if `writable' is set. (vc-cvs-revert,vc-cvs-checkout): References to `vc-checkout-model' updated to `vc-cvs-update-model'. (vc-cvs-logentry-check): Function added. (vc-cvs-revert,vc-cvs-checkout): Function calls to `vc-checkout-required' updated to `vc-cvs-uses-locking'. (vc-cvs-admin): Added the query-only option as required by the vc.el file. (vc-cvs-annotate-difference): Updated to handle beginning of annotate buffers correctly. Rename `vc-uses-locking' to `vc-checkout-required'. Rename the `locked' state to `reserved'. (vc-cvs-annotate-difference): Handle possible millenium problem (merged from mainline). Split the annotate feature into a BACKEND-specific part and moved the non-BACKEND stuff to vc.el. (vc-cvs-latest-on-branch-p): Function added. (vc-cvs-revert): Merged and adapted "unedit" patch from main line. (vc-cvs-diff): Function added. (vc-cvs-checkout): Function `vc-cvs-checkout' added. Require vc when compiling. (vc-cvs-register-switches): Doc fix. (vc-annotate-color-map, vc-annotate-menu-elements): Fix custom type. (vc-cvs-print-log, vc-cvs-assign-name, vc-cvs-merge) (vc-cvs-check-headers, vc-cvs-steal, vc-cvs-revert, vc-cvs-checkin): New functions (code from vc.el). (vc-annotate-display-default): Fix interactive spec. (vc-annotate-time-span): Doc fix. Moved the annotate functionality from vc.el. (vc-cvs-admin, vc-cvs-fetch-status): Added from vc.el. (vc-cvs-system-release): Renamed from vc-cvs-backend-release. (vc-cvs-release): Moved from vc.el. (vc-cvs-backend-release): New function. (vc-cvs-dired-state-info, vc-cvs-fetch-status): Moved from vc.el and renamed.
author Gerd Moellmann <gerd@gnu.org>
date Mon, 04 Sep 2000 19:48:04 +0000
parents 03ddf0b96330
children 23a1cea22d13
line wrap: on
line source

From lars Thu Feb 23 23:20:38 1995
From: larsi@ifi.uio.no (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: So you want to use the new Gnus
Message-ID: <lars-doc1@eyesore.no>

Actually, since you are reading this, chances are you are already
using the new Gnus.  Congratulations.

This entire newsgroup you are reading is, in fact, no real newsgroup
at all, in the traditional sense.  It is an example of one of the
"foreign" select methods that Gnus may use.

The text you are now reading is stored in the "etc" directory with the
rest of the Emacs sources.  You are using the "nndoc" backend for
accessing it.  Scary, isn't it?

This isn't the real documentation.  `M-x info', `m gnus <RET>' to read
that.  This "newsgroup" is intended as a kinder, gentler way of getting
people started.

Gnus is a rewrite of GNUS 4.1, written by Masanobu Umeda.  The rewrite
was done by moi, yours truly, your humble servant, Lars Magne
Ingebrigtsen.  If you have a WWW browser, you can investigate to your
heart's delight at <URL:http://www.ifi.uio.no/~larsi/larsi.html>.

;; Copyright (C) 1995 Free Software Foundation, Inc.

;; Author: Lars Magne Ingebrigtsen <larsi@ifi.uio.no>
;; Keywords: news

;; This file is part of GNU Emacs.

;; GNU Emacs is free software; you can redistribute it and/or modify
;; it under the terms of the GNU General Public License as published by
;; the Free Software Foundation; either version 2, or (at your option)
;; any later version.

;; GNU Emacs is distributed in the hope that it will be useful,
;; but WITHOUT ANY WARRANTY; without even the implied warranty of
;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
;; GNU General Public License for more details.

;; You should have received a copy of the GNU General Public License
;; along with GNU Emacs; see the file COPYING.  If not, write to the
;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
;; Boston, MA 02111-1307, USA.

From lars Thu Feb 23 23:20:38 1995
From: larsi@ifi.uio.no (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Starting up
Message-ID: <lars-doc2@eyesore.no>

If you are having problems with Gnus not finding your server, you have
to set `gnus-select-method'.  A "method" is a way of specifying *how*
the news is to be found, and from *where*.

Say you want to read news from you local, friendly nntp server
"news.my.local.server". 

(setq gnus-select-method '(nntp "news.my.local.server"))

Quite easy, huh?

From the news spool:

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

From your mh-e spool:

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

There's a whole bunch of other methods for reading mail and news, see
the "Foreign groups" article for that.


From lars Thu Feb 23 23:20:38 1995
From: larsi@ifi.uio.no (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Where are all the groups, then?
Message-ID: <lars-doc3@eyesore.no>

If this is the first time you have used a newsreader, you won't have a
.newsrc file.  This means that Gnus will think that all the newsgroups
on the server are "new", and kill them all.

If you have a .newsrc file, the new groups will be processed with the
function in the `gnus-subscribe-newsgroup-method' variable, which is
`gnus-subscribe-zombies' by default.

This means that all the groups have been made into "zombies" - not
quite dead, but not exactly alive, either.

Jump back to the *Group* buffer, and type `A z' to list all the zombie
groups.  Look though the list, and subscribe to the groups you want to
read by pressing `u' on the one you think look interesting.

If all the groups have been killed, type `A k' to list all the killed
groups.  Subscribe to them the same way.

When you are satisfied, press `S z' to kill all the zombie groups.

Now you should have a nice list of all groups you are interested in.

(If you later want to subscribe to more groups, press `A k' to
list all the kill groups, and repeat.  You can also type `U' and be
prompted for groups to subscribe to.)


From lars Thu Feb 23 23:20:38 1995
From: larsi@ifi.uio.no (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: I want to read my mail!
Message-ID: <lars-doc4@eyesore.no>

Yes, Virginia, you can read mail with Gnus.

First you have to decide which mail backend you want to use.  You have
nnml, which is a one-file-one-mail backend, which is quite nice, but
apt to make your systems administrator go crazy and come after you
with a shotgun.

nnmbox uses a Unix mail box to store mail.  Nice, but slow.

nnmh uses mh-e folders, which is also a one-file-one-mail thingie, but
slower than nnml.  (It doesn't support NOV files.)

So if you want to go with nnmbox, you can simply say:

(setq gnus-secondary-select-methods '((nnmbox "")))

(The same for the other methods, kind of.)

You should also set `nnmail-split-methods' to something sensible: 

(setq nnmail-split-methods 
      '(("mail.junk" "From:.*Lars")
	("mail.misc "")))

This will put all mail from me in you junk mail group, and the rest in
"mail.misc". 

These groups will be subscribe the same way as the normal groups, so
you will probably find them among the zombie groups after you set
these variables and re-start Gnus.


From lars Thu Feb 23 23:20:38 1995
From: larsi@ifi.uio.no (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Foreign newsgroups
Message-ID: <lars-doc5@eyesore.no>

These are groups that do not come from `gnus-select-method'. 

Say you want to read "alt.furniture.couches" from "news.funet.fi".  You
can then either type `B news.funet.fi <RET>' to browse that server and
subscribe to that group, or you can type 
`G m alt.furniture.couches<RET>nntp<RET>news.funet.fi<RET>', if you
like to type a lot.

If you want to read a directory as a newsgroup, you can create an
nndir group, much the same way.  There's a shorthand for that,
though.  If, for instance, you want to read the (ding) list archives,
you could type `G d /ftp <RET>'.

There's lots more to know about foreign groups, but you have to read
the info pages to find out more.


From lars Thu Feb 23 23:20:38 1995
From: larsi@ifi.uio.no (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Low level changes in GNUS, or, Wrong type argument: stringp, nil
Message-ID: <lars-doc6@eyesore.no>

Gnus really isn't GNUS, even though it looks like it.  If you scrape
the surface, you'll find that most things have changed.

This means that old code that relies on GNUS internals will fail.  

In particular, `gnus-newsrc-hashtb', `gnus-newsrc-assoc',
`gnus-killed-list', the `nntp-header-' macros and the display formats
have all changed.  If you have some code lying around that depend on
these, or change these, you'll have to re-write your code.

Old hilit19 code does not work at all.  In fact, you should probably
remove all hilit code from all the Gnus hooks
(`gnus-group-prepare-hook', `gnus-summary-prepare-hook' and
`gnus-summary-article-hook').  (Well, at the very least the first
two.)  Gnus provides various integrated functions for highlighting,
which are both faster and more accurated.

There is absolutely no chance, whatsoever, of getting Gnus to work
with Emacs 18.  It won't even work on Emacsen older than Emacs
19.30/XEmacs 19.13.  Upgrade your Emacs or die.

	
From lars Thu Feb 23 23:20:38 1995
From: larsi@ifi.uio.no (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: How do I re-scan my mail groups?
Message-ID: <lars-doc8@eyesore.no>

Reading the active file from the nntp server is a drag.

Just press `M-g' on the mail groups, and they will be re-scanned.

You can also re-scan all the mail groups by putting them on level 1
(`S l 1'), and saying `1 g' to re-scan all level 1 groups.


From lars Thu Feb 23 23:20:38 1995
From: larsi@ifi.uio.no (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: How do I set up virtual newsgroups?
Message-ID: <lars-doc9@eyesore.no>

Virtual newsgroups are collections of other newsgroups.  Why people
want this is beyond me, but here goes:

Create the group by saying

`M-a my.virtual.newsgroup<RET>nnvirtual<RET>^rec\.aquaria\.*<RET>'

This will create the group "nnvirtual:my.virtual.newsgroup", which
will collect all articles from all the groups in the "rec.aquaria"
hierarchy.  

If you want to edit the regular expression, just type `M-e' on the
group line.

Note that all the groups that are part of the virtual group have to be
alive.  This means that the cannot, absolutely not, be zombie or
killed.  They can be unsubscribed; that's no problem.

You can combine groups from different servers in the same virtual
newsgroup, something that may actually be useful.  Say you have the
group "comp.headers" on the server "news.server.no" and the same group
on "news.server.edu".  If people have posted articles with Distribution
headers that stop propagation of their articles, combining these two
newsgroups into one virtual newsgroup should give you a better view of
what's going on.

One caveat, though: The virtual group article numbers from the first
source group (group A) will always be lower than the article numbers
from the second (group B).  This means that Gnus will believe that
articles from group A are older than articles from group B.  Threading
will lessen these problems, but it might be a good idea to sort the
threads over the date of the articles to get a correct feel for the
flow of the groups:

(setq gnus-thread-sort-functions '(gnus-thread-sort-by-date))

If you only want this in virtual groups, you could say something along
the lines of:

(setq gnus-select-group-hook
      (lambda ()
  	(if (eq 'nnvirtual (car (gnus-find-method-for-group 
	                          gnus-newsgroup-name)))
	    (progn
	      (make-local-variable 'gnus-thread-sort-functions)
	      (setq gnus-thread-sort-functions '(gnus-thread-sort-by-date))))))


From lars Thu Feb 23 23:20:38 1995
From: larsi@ifi.uio.no (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Bugs & stuff
Message-ID: <lars-doc7@eyesore.no>

If you want to report a bug, please type `M-x gnus-bug'.  This will
give me a precise overview of your Gnus and Emacs version numbers,
along with a look at all Gnus variables you have changed.

Du not expect a reply back, but your bug should be fixed in the next
version.  If the bug persists, please re-submit your bug report.

When a bug occurs, I need a recipe for how to trigger the bug.  You
have to tell me exactly what you do to uncover the bug, and you should
(setq debug-on-error t) and send me the backtrace along with the bug
report.  

If I am not able to reproduce the bug, I won't be able to fix it.

I would, of course, prefer that you locate the bug, fix it, and mail
me the patches, but one can't have everything.

If you have any questions on usage, the "ding@ifi.uio.no" mailing list
is where to post the questions.