annotate etc/gnus-tut.txt @ 83004:7900111db01c

Converted display hooks to be display-local. Plus many bugfixes. lib-src/emacsclient.c (window_change_signal): Renamed to pass_signal_to_emacs. (init_signal): Pass SIGINT and SIGQUIT to the emacs process. lisp/faces.el (face-valid-attribute-values): Use the window-system function, not the variable. (read-face-attribute, face-spec-set-match-display, frame-set-background-mode) (face-set-after-frame-default): Ditto. lisp/frame.el (make-frame-on-tty): Added interactive declaration (suggested by Robert J. Chassell). Use tty-create-frame-with-faces, not make-terminal-frame. src/termhooks.h (struct display_method): Renamed to display for brevity. (struct display): Added all display hook variables as members of this structure. Added next_display, reference_count, type and display_info components. (FRAME_MUST_WRITE_SPACES, FRAME_FAST_CLEAR_END_OF_LINE, FRAME_LINE_INS_DEL_OK) (FRAME_CHAR_INS_DEL_OK, FRAME_SCROLL_REGION_OK, FRAME_SCROLL_REGION_COST) (FRAME_MEMORY_BELOW_FRAME, FRAME_RIF): Updated for struct display. (FRAME_DISPLAY): New macro. (create_display, delete_display): New prototypes. src/frame.h (struct frame): Added `display' member, removed display_method. (FRAME_LIVE_P): Look at f->display, not f->output_data. src/termchar.h (struct tty_display_info): Removed display_method component. (FRAME_TTY): Use the display structure, not output_data. src/term.c (display_list): New variable. (cursor_to_hook, raw_cursor_to_hook, clear_to_end_hook, clear_frame_hook) (clear_end_of_line_hook, ins_del_lines_hook, delete_glyphs_hook) (ring_bell_hook, reset_terminal_modes_hook, set_terminal_modes_hook) (update_begin_hook, update_end_hook, set_terminal_window_hook) (insert_glyphs_hook, write_glyphs_hook, delete_glyphs_hoo, read_socket_hook) (frame_up_to_date_hook, mouse_position_hook, frame_rehighlight_hook) (frame_raise_lower_hook, set_vertical_scroll_bar_hook, condemn_scroll_bars_hook) (redeem_scroll_bar_hook, judge_scroll_bars_hook): Moved to struct display. (tty_display_method_template): Removed. (syms_of_term): Don't initialize tty_display_method_template. (ring_bell, set_terminal_modes, reset_terminal_modes, update_begin) (update_end, set_terminal_window, cursor_to, raw_cursor_to, clear_to_end) (clear_frame, clear_end_of_line, write_glyphs, insert_glyphs) (delete_glyphs, ins_del_lines): Access display hooks through the frame pointer. (Ftty_display_color_p): Use the frame given as a parameter, or else return nil. (Ftty_display_color_cells): Ditto. (get_named_tty): Renamed to get_named_tty_display, changed return type to struct display. (term_dummy_init): Renamed to initial_term_init. Create and return an initial display. (term_init): Initialize a new struct display and return a pointer to it instead of tty_display_info. Removed frame initialization kludge. (Fdelete_tty): Updated for struct display. (delete_tty): The parameter type is now struct display, not tty_display_info. Delete the display, too. (create_tty_output): New function for creating tty_output structures. (delete_tty_output): New function for deleting tty_output structures. (create_display): New function for creating and registering display structures. (delete_display): New function for deleting and unregistering display structures. src/dispextern.h: Updated prototypes. src/dispnew.c: Include frame.h before termhooks.h. (init_display): Updated term_init call to new signature. src/emacs.c: Include frame.h (for termhooks.h). src/keymap.c: Ditto. src/lread.c: Ditto. src/xsmfns.c: Ditto. src/process.c: Include frame.h before termhooks.h. src/frame.c (Fwindow_system): New function. (syms_of_frame): Initialize it. (make_terminal_frame): Open the terminal device before creating the new frame. Disable scrollbars here, term_init cannot do that anymore. (Fdelete_frame): Use the new delete_frame_hook, don't do display-specific frame deletion here. Ditto for delete_display_hook. (Fmouse_position, Fmouse_pixel_position, Fraise_frame, Flower_frame) (Fredirect_frame_focus): Access display hooks through the frame pointer. src/keyboard.c: Include frame.h before termhooks.h. (start_polling, input_polling_used, stop_polling, gobble_input): Ignore read_socket_hook. (kbd_buffer_get_event, Fset_input_mode): Access display hooks through the frame pointer. (read_avail_input): Loop through all display devices for and call all read_socket_hook functions. Check ttys even if read_socket_hook returned an error. src/sysdep.c (discard_tty_input): Ignore read_socket_hook. (stuff_char): Don't do anything if the current frame is not on a termcap display. (request_sigio, unrequest_sigio): Ignore read_socket_hook. (init_sys_modes): Always call narrow_foreground_group. Set up terminal modes and sigio even under X. src/xdisp.c (message2_nolog, message3_nolog, redisplay_internal) (set_vertical_scroll_bar, redisplay_window): Access display hooks through the frame pointer. (echo_area_display): Don't be afraid of termcap frames during an X+tty combo session. src/xfaces.c: Include termhooks.h. (Ftty_supports_face_attributes_p): Use the given frame, not selected_frame. src/xfns.c (x_set_scroll_bar_foreground, x_set_scroll_bar_background): Access display hooks through the frame pointer. (Fx_create_frame, x_create_tip_frame): Initialize the frame's display structure. src/xmenu.c: Include termhooks.h after frame.h. src/xselect.c (x_own_selection, some_frame_on_display, x_get_foreign_selection) (Fx_disown_selection_internal, Fx_get_cut_buffer_internal) (Fx_store_cut_buffer_internal, Fx_rotate_cut_buffers_internal): Don't do anything if the selected frame is not an X frame. src/xterm.c (x_display_method): Removed. (x_create_frame_display, x_delete_frame_display): New functions for handling struct display objects. (x_term_init): Set up a new struct display object, too. (x_delete_display): Delete the struct display corresponding to the X display. (x_initialize): Moved hook initialization to x_create_frame_display. src/xterm.h (x_display_method): Removed. (struct x_display_info): Added frame_display component. git-archimport-id: [e-mail]--2004/emacs--multi-tty--0--patch-44

From lars Thu Feb 23 23:20:38 1995
From: [e-mail] (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: So you want to use the new Gnus
Message-ID: <[e-mail]>

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 <[e-mail]>
;; 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: [e-mail] (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Starting up
Message-ID: <[e-mail]>

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: [e-mail] (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Where are all the groups, then?
Message-ID: <[e-mail]>

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: [e-mail] (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: I want to read my mail!
Message-ID: <[e-mail]>

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: [e-mail] (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Foreign newsgroups
Message-ID: <[e-mail]>

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: [e-mail] (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Low level changes in GNUS, or, Wrong type argument: stringp, nil
Message-ID: <[e-mail]>

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: [e-mail] (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: How do I re-scan my mail groups?
Message-ID: <[e-mail]>

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: [e-mail] (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: How do I set up virtual newsgroups?
Message-ID: <[e-mail]>

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: [e-mail] (ding)
Date: Fri Feb 24 13:40:45 1995
Subject: Bugs & stuff
Message-ID: <[e-mail]>

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 "[e-mail]" mailing list
is where to post the questions.