Mercurial > emacs
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.