Article View: gnu.epoch.misc
Article #1386Epoch Frequently Asked Questions (FAQ)
From: Marc Andreessen
Date: Sun, 31 May 1992 21:48
Date: Sun, 31 May 1992 21:48
1359 lines
55021 bytes
55021 bytes
Archive-name: epoch-faq
Last-modified: 18 May 1992
Version: 1.0.1
Posting-frequency: monthly
----------------------------------------------------------------------
* Epoch FAQ (Frequently Asked Questions)
* (or at least Occasionally Wondered-About Questions)
* Maintained by Marc Andreessen (marca@ncsa.uiuc.edu)
* Last Modified Mon May 18 22:17:07 1992 -- Marc Andreessen
If you are viewing this text in a GNU Emacs buffer, you can type
"M-2 C-x $" to get an overview of just the questions. Then, when
you want to look at the text of the answers, just type "C-x $".
To search for a question numbered XXX, type "M-C-s ^XXX:", followed by
a C-r if that doesn't work; type ESC to end the search.
The latest version of this file will be always available via anonymous
FTP to ftp.ncsa.uiuc.edu (141.142.20.50) in /outgoing/marca/epoch.
Please feel free to send comments, suggestions, additions, and
corrections to marca@ncsa.uiuc.edu.
* === REALLY IMPORTANT NOTE ===
Unless otherwise attributed, information below is based only on Yours
Truly's use and experience with Epoch and is BY NO MEANS GUARANTEED to
be correct, complete, or even coherent. If it works for you, great;
if it doesn't, let me know, but there are no guarantees.
* === CHANGES IN THIS VERSION OF THE FAQ ===
Here is a list of changes made since the FAQ's last monthly
posting to gnu.epoch.misc and news.answers:
a) Addition of this section of the FAQ.
b) F-1: Two more ways to fix the backspace key.
c) G-2: Patch to make drag style always display.
d) Z-3: Only applies to X11R5 multiscreen servers, plus workaround.
* === PREREQUISITES ===
In the answers that follow, I assume at least a little familiarity
with GNU Emacs Lisp (elisp). Suggested resources that could be
considered necessary references for material in this FAQ are:
--> The GNU Emacs and Epoch manuals (available online via C-h i).
--> The GNU Elisp manual (available via anonymous FTP from
prep.ai.mit.edu (18.71.0.38) as /pub/gnu/elisp-manual-1.03.tar.Z).
--> The gnu.emacs.help FAQ in the file etc/FAQ (the latest version
is available in the gnu.emacs.help Usenet newsgroup or via FTP from
pit-manager.mit.edu (18.172.1.27) in /pub/usenet/gnu.emacs.help).
* === CAVEATS ===
Many of the answers given herein APPLY ONLY to Epoch 4.0 (and also may
only apply to 4.0 beta level 0 or later). When I know about them,
3.2'isms and 4.0-alpha'isms are mentioned also. Also, my Epoch
platform is an SGI IRIS Indigo XS running IRIX 4.0.2; other
configurations may behave differently. Please let me know of any such
differences. (Now that an official release of 4.0 has been made,
details about previous versions will be removed at some point in the
future.)
* === EPOCH AND GNU EMACS ===
A-1: What is Epoch?
Epoch is GNU Emacs on steroids: an adaptation of GNU Emacs with lots
of additional support for features made possible by the X11 windowing
system. These features include multiple editing windows, arbitrary
colors and fonts (fixed-width and proportional), selectable zones per
buffer with arbitrary display styles (font, color, underline, stipple,
pixmap), an optional separate minibuffer window, improved keyboard and
mouse handling, full 8-bit character set support, and more.
Epoch, like GNU Emacs, is very customizable; much of this FAQ
describes ways to customize Epoch for specific purposes.
A-2: How can I get Epoch?
Epoch is available via anonymous FTP to cs.uiuc.edu (128.174.252.1)
in pub/epoch-files/epoch. The main file of interest is currently
epoch-4.0p1.tar.Z.
According to Simon Marshall (S.Marshall@sequent.cc.hull.ac.uk):
Like GNU & X11 stuff, epoch is available in the UK from
uk.ac.ic.doc.src (00000510200001). You can log on; there's help to
tell you the different ways of getting source. Epoch is in
/gnu/epoch.
Epoch is also available from aix370.rrz.uni-koeln.de (134.95.132.2)
in /gnu/emacs/epoch.
Volunteers for other FTP sites outside the United States, or for
mail access to Epoch source, please step forward.
A-3: What is the latest version of Epoch?
The latest version is 4.0 patchlevel 1. Previous versions still in
use by some people include 4.0 patchlevel 0, the 4.0 alpha and beta
levels, and version 3.2.
Epoch 4.0p1 (a.k.a. 4.0 patchlevel 1) is based on GNU Emacs 18.58,
which is the most recent version of GNU Emacs.
A-4: What are the differences between the various versions of Epoch?
NOTE: Many bugs were also fixed from version to version; for the
most part, those fixes aren't mentioned here.)
Important differences between 3.2 and 4.0-alpha include:
[From initial announcement of Epoch 4.0a0.]
a) Based on 18.57 GNU Emacs. [3.2 was based on 18.55.]
b) Completely rebuilt display code: now supports proper proportional
fonts.
c) Changes from attributes to styles.
d) Minibuffer can now be in its own window (traditional Epoch) or in the
same window as the epoch screens (traditional GNU Emacs).
e) Mouse maps now based on Ken Laprade's mouse package.
f) Selections rather than cut-buffers (but cut-buffers still work).
g) Runs on more platforms.
Important differences between 4.0-alpha and 4.0-beta include:
[From 4.0-beta README.]
h) Includes code to build ASCII version of GNU Emacs 18.57.
i) Support for sharing lisp directory between Epoch and GNU Emacs
18.57 installed elsewhere.
j) "Buttons" became "zones". (See question A-5 below.)
k) Performance enhancements.
Important differences between 4.0b0 and 4.0b1 include:
[From 4.0b1 README.]
l) Based on 18.58 GNU Emacs.
m) Bug-fixes for graphical zone support.
n) Epoch no longer reads $HOME/.Xdefaults on its own; it assumes
that Epoch resources are available in the X server's resource
database. (This is ICCCM-compliant behavior, which makes Epoch
more like other X applications).
Important differences between 4.0b1 and 4.0p0 include:
[From preliminary announcement of 4.0p0.]
o) Epoch will read $HOME/.Xdefaults on its own if the '-ud' command
line option is used; otherwise, it uses the X resource database
only.
p) Additional support for 8-bit fonts and hypertext info in the
contrib directory.
Important differences between 4.0p0 and 4.0p1 include:
[From announcement of 4.0p1.]
q) Added support for pasting from Epoch to Open Windows clients.
r) Ability to set cursor foreground/background.
s) Ability to load standard X bitmap files as well as XPM pixmap
files.
t) Replacement contrib/hyper/info.el for broken 4.0p0 version.
A-5: I recently moved to Epoch 4.0b0 (or later), and elisp code that
worked before has stopped working. Why?
Change c) for Epoch 4.0b0 is probably the culprit; the term "zone" was
chosen to replace the term "button", changing the names of all the
button-related function names and variables. There are two ways to
deal with this change:
METHOD #1:
Put the following command at the head of each file of elisp code
that uses buttons:
(require 'button)
(This simply does a series of fset's to replace the "button" names
with the new "zone" names.)
METHOD #2:
Convert all references of "button" to "zone". You can use the
function convert-current-buffer in epoch-lisp/convert-buttons.el
(in the Epoch distribution) to do this automatically.
NOTE: There may be additional problems; in particular, since
attributes were eliminated in Epoch 4.0a0, any use of them will
have to be eliminated also. This is less simple than switching
from buttons to zones. Fortunately, most Epoch packages on the
net now support 4.0. Therefore, I won't bother to come up with an
answer for this one (unless someone actually needs to know).
A-6: I recently moved to Epoch 4.0b1 (or later), and my X resources are
suddenly being ignored. Why?
Change n) for Epoch 4.0b1 above has caused some trouble for people
moving to 4.0b1 (or later). Previous to 4.0b1, Epoch would always
extract its X default resources from the user's .Xdefaults file by
hand; this is no longer the case. If Epoch 4.0b1 (or later) isn't
picking up resources like previous versions did on your system, prior
to starting Epoch do this:
xrdb -merge $HOME/.Xdefaults
This makes the X server's database pick up and maintain all of the
resource definitions found in $HOME/.Xdefaults.
If you are running Epoch 4.0p0 or later you can use the '-ud'
command line flag to force Epoch to manually scan $HOME/.Xdefaults.
A second problem is that Epoch 4.0b1 (or later) uses the first word
in the command that executed Epoch (argument 0) to search the X
resource database for Epoch resources. Thus, if you invoke Epoch
with the name:
epoch-4.0.1
...Epoch will use the string 'epoch-4.0.1' to search the resource
database, thereby missing the resources specified by the name
'epoch'. The workaround is to use the '-rn' command-line option
like so:
epoch-4.0.1 -rn epoch
If you are running Epoch 4.0p1 or later you can use the '-name' flag
(which is standard for X applications) in place of '-rn', like so:
epoch-4.0.1 -name epoch
[From Martin Boyer (mboyer@ireq-robot.hydro.qc.ca).]
A-7: What's the difference between Epoch and GNU Emacs version 19?
To quote directly from the GNU 'status' file (obtained via anonymous
FTP to prep.ai.mit.edu (18.71.0.38) as /pub/gnu/status):
-- BEGIN QUOTE -----------------------------------------------------
Emacs 18 maintenance continues for simple bug fixes. Version 19
approaches release, counting among its new features before and after
change hooks, source-level debugging of Emacs Lisp programs, X
selection processing, including CLIPBOARD selections, scrollbars,
support for European character sets, floating point numbers,
per-buffer mouse commands, interfacing with the X resource manager,
mouse-tracking, lisp-level binding of function keys, and multiple X
windows (`screens' to emacs).
Thanks go to Alan Carroll and the people who worked on Epoch for
generating initial feedback to a multi-windowed emacs. Emacs 19
supports two styles of multiple windows, one with a separate screen
for the minibuffer, and another with a minibuffer attached to each
screen.
A couple of other features of Version 19 are buffer allocation, which
uses a new mechanism capable of returning storage to the system when a
buffer is killed, and a new input system--all input now arrives in the
form of Lisp objects.
Other features under consideration for later releases of Version 19
include:
** Associating property lists with regions of text in a buffer.
** Multiple font, color, and pixmaps defined by those properties.
** Different visibility conditions for the regions, and for various
windows showing one buffer.
** Hooks to be run if point or mouse moves outside a certain range.
** Incrementally saving undo history in a file, so that recover-file
also reinstalls buffer's undo history.
** Static menu bars, and better pop-up menus.
-- END QUOTE -------------------------------------------------------
Epoch currently supports much of the new functionality of GNU Emacs
v19, including some features that apparently won't be immediately
available in GNU Emacs v19 (notably, associating property lists with
regions of text, and multiple fonts/colors/pixmaps defined by those
properties).
A postscript to this answer is a message that Richard Stallman
(Mr. Gnu himself) posted to gnu.emacs.help in Apr 1992:
-- BEGIN QUOTE -----------------------------------------------------
Emacs 19 won't support text highlighting, but a later version will.
We have code basically working for attaching arbitrary properties to
parts of the text, including choice of font and color.
This feature is designed for attributes which are considered part of
the text itself. What Epoch provides is highlighting that is part
of the presentation of the text, not logically part of the text itself.
That is useful also; I hope we will find the wherewithal to implement
that along with the other highlighting features.
-- END QUOTE -------------------------------------------------------
* === CONFIGURING EPOCH ===
B-1: How can my .emacs file deal with both GNU Emacs and Epoch?
An easy way to handle both GNU Emacs and Epoch within the same .emacs
file is to define a variable 'running-epoch' as follows:
(defvar running-epoch (boundp 'epoch::version))
Then, the rest of your .emacs file can look like this:
(if running-epoch
(progn
... [Epoch-specific code here] ...
))
... [Emacs code here] ...
(After this, when I refer to 'your .emacs file', read that as 'the
Epoch-specific portion of your .emacs file', under the assumption that
you have both general and Epoch-specific parts as detailed above.)
If you have a large amount of Epoch-specific code it may be better to
put it in a separate file (e.g., "epoch-specific.el"), possibly
byte-compile it, then load it from your .emacs file like so:
(if (boundp 'epoch::version) (load "epoch-specific"))
B-2: What if I want to have separate sections for Epoch 3.2 and 4.0 in
my .emacs file?
The variable 'epoch::version' is a string that contains the current
Epoch version ID. For instance, evaluating epoch::version under
Emacs4.0b0 (the first beta level) gives:
"Epoch 4.0 Beta Patchlevel 0"
A simple way to determine if version 4.x is running is to do something
like this:
(string-match " 4" epoch::version)
This will result in 't' if the string " 4" is a substring of
epoch::version (as would be the case when epoch::version is
"Epoch 4.0 ..."), and 'nil' otherwise.
B-3: How can I configure Epoch to support full 8-bit character sets?
The variable 'ctl-arrow', which is a buffer-local variable, controls
how control characters are displayed. By setting this variable to
something non-nil and non-t, you get full 8-bit characters. To get
8-bit characters in all buffers, put this in your .emacs file:
(setq-default ctl-arrow 'foo)
[Thanks to Pierre Bondono (bondono@vnet.ibm.com) and Tor Lillqvist
(tml@tik.vtt.fi) for the following:]
There is also a package in the Epoch contrib directory that deals
with 8-bit characters better than the above technique. See the
directory contrib/keriven in the Epoch distribution.
* === WINDOWS ===
C-1: How do I get the minibuffer back at the bottom of my edit window?
Start Epoch with the command-line option '-nm'. Alternately, in your
general .Xdefaults file, specify:
epoch.nonlocal.minibuf: False
(These techniques only work in Epoch 4.0.)
C-2: How do I prevent the initial edit window from appearing? (In
other words, how do I get only the minibuffer upon startup?)
There is an undocumented variable that controls whether or not an
initial edit window will be provided. (This variable is documented as
of 4.0p0.) By default, of course, one is; to stop this from
happening, put the following in your .emacs:
(setq inhibit-initial-screen-mapping t)
This will have no effect if you're using local minibuffers (i.e. when
the minibuffer is a part of each edit window), as described above.
[Thanks to Chris Love (love@cs.uiuc.edu).]
C-3: How do I stop the windows from automatically raising upon entry?
The easiest way to do this, if you're running Epoch 4.0b0 or later, is
to specify the following in your .emacs file:
(setq auto-raise-screen 'minibuf)
If you're running a version of Epoch prior to Epoch 4.0b0, the
following code also works:
(setq auto-raise-screen nil)
(push (function (lambda () (raise-screen (minibuf-screen))))
*select-screen-hook*)
Both of the above solutions still make the minibuffer raise when an
Epoch edit window is entered; if you don't want the minibuffer to
raise itself either, just do the following instead:
(setq auto-raise-screen nil)
C-4: How do I change the cursor?
Another noticeable change from GNU Emacs is the cursor glyph (which is
a pencil shape). To change this to a more normal-appearing arrow
shape, put this in your .emacs file:
(cursor-glyph 68)
(cursor-glyph 68 (epoch::minibuf-screen))
(push '(cursor-glyph . 68) epoch::screen-properties)
NOTE: You can specify a different cursor glyph for the minibuffer,
if you like.
There are also X resources for the screen and minibuffer cursor
glyphs; equivalent to the above elisp code would be the following (in
your .Xdefaults file):
epoch.screen.cursor.glyph: 68
epoch.minibuf.cursor.glyph: 68
For either of these approaches, any valid cursor ID number will do; to
see the available cursors and their ID numbers, take a look at the
file /usr/include/X11/cursorfont.h.
(See below for details on how to customize the cursor glyph based on
the mode of a screen's initial buffer.)
C-5: How do I define edit screen attributes on a per-mode basis? Or,
how do I make my C++ edit windows come up with 48 lines and a big
round dot for a cursor?
You can define styles for new edit screens created with
find-file-other-screen (which is, by default, bound to 'C-z 4 C-f' and
'C-z 4 f') based on the initial buffer's mode like so:
(setq epoch-mode-alist
(append (list
(cons 'c-mode
(list
(cons 'geometry "80x52")
(cons 'cursor-glyph 94)))
(cons 'LaTeX-mode
(list
(cons 'geometry "90x48")
(cons 'cursor-glyph 58)))
(cons 'c++-mode
(list
(cons 'geometry "80x48")
(cons 'cursor-glyph 38))))
epoch-mode-alist))
C-6: How do I make Epoch manage edit window and icon names based on the
name of the current buffer(s)?
The base Epoch distribution doesn't support this, but there have been
a number of small packages written that accomplish this. The one
Yours Truly uses was written by Derek Ney (derek@boingo.med.jhu.edu)
and is up for anonymous FTP at ftp.ncsa.uiuc.edu (141.142.20.50) as
/outgoing/marca/epoch/derek-title.el.Z.
C-7: How do I iconify my entire Epoch session at once?
By default, 'C-z i' is bound to iconify-screen, which only iconifies
the current edit window. To iconify everything, define a small
function iconify-everything in your .emacs file like so:
(defun iconify-everything ()
"Iconify entire Epoch session at once."
(interactive)
(dolist (s (screen-list))
(iconify-screen s))
(iconify-screen 0)
)
Then bind it to a key sequence (e.g. 'C-z C-i') like so:
(global-set-key "\C-z\C-i" 'iconify-everything)
C-8: Does Epoch support scrollbars?
No; native support for scrollbars is a technical challenge that so far
no one has taken up for either GNU Emacs or Epoch.
[Thanks to Pierre Bondono (bondono@vnet.ibm.com) for the following:]
Related to this, there is a ``scrolling concept'' inside the IMOUSE
packages that may be of interest. IMOUSE can be found in
contrib/boyer in the Epoch distribution or via anonymous FTP from
ireq-robot.hydro.qc.ca (131.195.2.130); see question H-1 below.
C-9: My workstation's screen area is overwhelmed by all the windows I
create in my Epoch editing sessions. How can I squeeze more windows
into my working environment?
This is only a half-facetious question; there's only room on most
workstation screens (usually 1024x768 or 1280x1024 pixels in area) for
a couple of windows with decent-sized fonts. To help alleviate the
discomfort of cramped space, try the 'tvtwm' window manager. (I sound
like a cough-drop commercial.)
Tvtwm, a superset of the twm window manager, allows windows to be
placed on a virtual root window that can be of any size (for example,
3000 by 3000 pixels). This virtual root can be navigated via tvtwm's
'panner', a reduced-size overview of your entire environment.
Tvtwm can be found on your local comp.sources.x archive or in the
X11R4 and X11R5 distributions. Try anonymous FTP to ftp.uu.net
(137.39.1.9) in /usenet/comp.sources.x.
* === COLORS AND FONTS ===
D-1: How do I interactively select my foreground color.
First, define a style for each of your desired foreground colors in
your .emacs file, like so:
(setq red-style (make-style))
(set-style-foreground red-style "red")
Then, define a short command 'goto-style' like so:
(defun goto-style (newstyle)
"Set current buffer to a given font NEWSTYLE."
(setq buffer-style newstyle)
(redraw-display))
Next, bind commands to keys like so:
(global-set-key "\C-c1" (definteractive (goto-style red-style)))
This binds a command that changes the current foreground color to red
to the key sequence C-c 1.
The following function allows you to interactive specify a new color
for the current buffer's foreground:
(defun set-foreground (newfg)
"Set current foreground color to NEWFG."
(interactive "sForeground color: ")
(set-style-foreground buffer-style newfg)
(redraw-display))
This function is activated by typing:
M-x set-foreground
Or, it can be bound to a key like so:
(global-set-key "\C-z\C-s" 'set-foreground)
D-2: How do I switch fonts interactively?
To change fonts, you don't have to define new styles. Just define a
short command 'goto-font', like so:
(defun goto-font (newfont)
"Set current buffer to a given font NEWFONT."
(font newfont)
(redraw-display))
Then, bind commands to keys like so:
(global-set-key "\C-c1" (definteractive (goto-font "8x13")))
This changes the current edit window's font to "8x13" when the key
sequence 'C-c 1' is entered.
The following function allows you to interactively input a font name:
(defun set-font (newfont)
"Set current font to NEWFONT."
(interactive "sFont name: ")
(font newfont)
(redraw-display))
This function is activated by typing:
M-x set-font
Or, it can be bound to a key like so:
(global-set-key "\C-z\C-s" 'set-font)
If the font you specify with either of these methods isn't valid for
your X server (i.e. if the font doesn't exist in a directory listed in
your X server's font path and in that directory's fonts.dir file, or
isn't available from the font server if you're running X11R5), you
will get this message in your minibuffer:
Bad font name
Font names are ordinary X font names; for example, a 14-point non-bold
non-italic Helvetica font can be referred to as:
"*helvetica-medium-r-normal--14*"
The standard X client 'xlsfonts' will give you a list of the fonts
registered with your X server. The standard X client 'xfontsel' lets
you interactively put together an X font specification string based on
individual attributes (e.g., family, size).
D-3: How can I associate a certain foreground color with a file
so that every time I load that file my foreground color is
automatically set?
You can easily do this by using GNU Emacs' ability to handle file
variables. For example, if you have defined style 'red-style' to
specify a red foreground color, and want file 'foobar.c' to always
have the foreground color red, add this to the bottom of 'foobar.c':
^L
/*
Local Variables:
buffer-style: red-style
End:
*/
NOTE: '^L' is a newline character; you can enter one in an edit
buffer by typing 'C-q C-l'.
Then, whenever 'foobar.c' is loaded, the variable 'buffer-style' will
be set to 'red-style', and as a result the foreground color for that
buffer will be red.
D-4: How do I design my own really funky technicolor modeline?
Using the tried-and-true Learning By Example(tm) educational method,
I'll just give the (heavily commented) elisp code I use to define my
own customized modeline:
;; Define styles for the modeline.
(setq cyan-style (make-style))
(set-style-foreground cyan-style "#6fcfef")
(setq red-style (make-style))
(set-style-foreground red-style "red")
(setq yellow-style (make-style))
(set-style-foreground yellow-style "yellow")
;; Make display-time display day and date also.
(setq display-time-day-and-date t)
;; Display time, day, and date in modeline.
(display-time)
;; List full filename, with path.
(set-default 'mode-line-buffer-identification
'(buffer-file-name ("%f") ("%b")))
;; This actually turns inverse-video _off_ for the modeline.
(setq mode-line-inverse-video t)
;; Set the modeline itself.
(setq default-mode-line-format
(list red-style "--"
;; Buffer name.
yellow-style 'mode-line-buffer-identification
red-style "-"
;; Buffer status (read-only, unchanged, changed).
cyan-style "%*%*"
red-style "-"
;; Position in buffer.
cyan-style "%[%3p%]"
red-style "--"
;; Current time.
yellow-style "%M"
red-style "--("
;; Mode name.
cyan-style 'mode-name 'minor-mode-alist "%n"
'mode-line-process
red-style ")-%-"))
D-5: How do I replace the mouse region selection underlining with a
different style?
You can override the variable 'motion::style' with your own style,
like so:
(setq motion::style (make-style))
Now, you can set the various elements of this style as you please.
Several examples follow.
To have just the text foreground color change on mouse selection, do:
(set-style-foreground motion::style "yellow")
(set-style-background motion::style (background))
NOTE: Replace "yellow" with any desired X color name.
To have the selected region appear as reverse video, try:
(set-style-foreground motion::style (background))
(set-style-background motion::style (foreground))
To have the selected region appear with an overlayed diagonal stipple
pattern, try:
(set-style-background motion::style "red")
(set-style-background-stipple motion::style
(make-bitmap 4 4 "\167\273\335\356"))
Also note that, if you like, you can just modify the style bound by
default to motion::style instead of actually creating a new style.
The style bound to motion::style has default settings as follows:
(set-style-foreground motion::style (foreground))
(set-style-background motion::style (background))
(set-style-underline motion::style (foreground))
So, for instance, to get yellow underlining, you could just do:
(set-style-underline motion::style "yellow")
* === MENUS ===
E-1: Does Epoch support popup menus or menubars?
The base Epoch distribution does not have native support for popup
menus or menubars. However, there are other ways to use menus with
Epoch.
METHOD #1:
If you're running the window manager GWM (available via anonymous FTP
to export.lcs.mit.edu (18.24.0.12) in /contrib/gwm), you can use the
Epoch contrib 'wm-menu' package (found in the contrib directory in the
Epoch distribution) to have GWM provide menus that pass messages to
Epoch.
METHOD #2:
If you're running a window manager that has user-specifiable menus (as
all of the common window managers do), you can use the xsendevent to
send events (e.g. keystrokes) to Epoch by activating window
manager-produced menu entries.
An example menu (in twm format) is as follows:
menu "epoch"
{
"Epoch Menu" f.title
"Undo" ! "xse -window ClickWindow 'Ctrl<Key>x' '<Key>u'"
"Quit" ! "xse -window ClickWindow 'Ctrl<Key>x' 'Ctrl<Key>c'"
}
Button3 = c : root : f.menu "epoch"
The epoch menu is activated when Ctrl-Btn3 is pressed on the root
window. You theoretically will have to point at an Epoch window and
click with the left mouse button (after activating a menu entry) for a
command to take effect. However, I [Marc] haven't been able to
actually get the ClickWindow option to work; xsendevent (version 2.0)
always wants to send the event to the window that contains the
pointer. So what I do instead is define a twm titlebutton for the
epoch menu like so:
menu "epoch"
{
"Epoch Menu" f.title
"Undo" ! "xse 'Ctrl<Key>x' '<Key>u'"
"Quit" ! "xse 'Ctrl<Key>x' 'Ctrl<Key>c'"
}
LeftTitleButton "~/epochbitmap.bm" = f.menu "epoch"
("~/epochbitmap.bm" is a little X bitmap that displays as the button's
label.) Now pulling down the menu and activating an entry will send
the keystrokes to the current window. This is not an ideal solution,
since all windows (Epoch or not) in the session will have the menu
attached to them, but it does work and is quite convenient for use
with Epoch.
To bind arbitrary commands to menu entries in this manner, you need to
define a unique keystroke pattern (with global-set-key) for each
command and then send that keystroke pattern with xsendevent via the
menu. See the section on rebind-key (below) for more information on
this technique. (Of course, you are also free to have xsendevent send
a key sequence like 'ESC x u n d o RET' directly to Epoch.)
You can get xsendevent at export.lcs.mit.edu (18.24.0.12) in /contrib
or at ftp.uu.net (137.39.1.9) in /packages/X/contrib.
[From Philippe Bondono (bondono@vnet.ibm.com).]
METHOD #3:
A completely different technique involves using a separate X client
that posts a menu and has Epoch (or for that matter, GNU Emacs)
receive commands from it. One example of this technique is
/gnu/emacs/buttons.tar.Z from anonymous FTP to aix370.rrz.uni-koeln.de
(134.95.132.2).
METHOD #4:
Another method uses xmenu or xmenu2, which are available at any
comp.sources.x archive. (If you don't know of such an archive, try
anonymous FTP to ftp.uu.net (137.39.1.9) in /usenet/comp.sources.x.)
After you have compiled xmenu or xmenu2, define the following function
in your .emacs file:
(defun interpret-output (process output)
(command-execute (car (read-from-string output))))
Then you can define other functions that make specific menus, with
menu entries bound to commands, like the following example shows:
(defun make-menu ()
(interactive)
(set-process-filter (start-process "xmenu" nil
"/usr/local/bin/xmenu" "-heading" "Epoch Menu"
"Undo=undo")
'interpret-output))
This example uses /usr/local/bin/xmenu to post a one-entry menu to the
screen; the menu entry's label is "Undo" and the action it triggers in
Epoch is 'undo'. Specify as many "Label=action" pairs as you like as
part of the start-process command.
[From Philippe Bondono (bondono@vnet.ibm.com).]
If you use XView/Open Windows, Dev Joneja (dj7@cunixf.cc.columbia.edu)
has written a menu/button client similar to xmenu; it's available via
anonymous FTP from ftp.ncsa.uiuc.edu (141.142.20.50) as
/outgoing/marca/epoch/xvmenu.tar.Z.
Bob Weiner (rsw@cs.brown.edu) says the following concerning menus
within Hyperbole:
----------------------------------------------------------------------
If you use Hyperbole, you know what the Smart Key system is, context
sensitive key bindings for many Emacs modes and subsystems. There is
a Smart Menu system that goes along with this but has not been
released because it needs some additional menus and some integration
work. It is window manager and window system independent, works under
Emacs, with a slightly customized version that can highlight selected
menu items and display menus in a separate Epoch screen, and is
written entirely in Elisp, so no patching is needed. It provides much
easier subsystem access for novices and experienced users and has been
in use at a few sites for the last several years.
I'd like to get it in a form for release but won't have the time for
several months. If an experienced Elisp programmer wanted to
integrate it with Hyperbole and make any other changes, I [Bob, not
Marc] would be willing to send it out and coordinate on and test the
final changes.
----------------------------------------------------------------------
See the list of FTP sites below for more information on Hyperbole.
* === KEYBOARD ===
F-1: How do I make the backspace key work properly?
This is actually a GNU Emacs question, but it's included in this FAQ
because it's probably more important to more people than all the
other questions here. (One of the great mysteries of GNU Emacs is
why the backspace key _still_ doesn't do its job.)
Here are four ways to fix this:
METHOD #1:
Put this in your .emacs file:
(rebind-key "BackSpace" nil "\C-?")
This is the cleanest Epoch-specific solution. It rebinds
the X representation for the backspace key directly to that
of the delete key.
[Thanks to Joe Wells (jbw@bigbird.bu.edu).]
METHOD #2:
Put this in your .emacs file:
(global-set-key "\C-^bdc" 'backward-delete-char-untabify)
(rebind-key "BackSpace" nil "\C-^bdc")
This rebinds the X representation for the backspace key to the
appropriate command for deleting a character backward. Like the
first method, this is Epoch-specific.
If you don't like to have your backspace key also untabify (i.e.,
convert tags to spaces on the fly), use this instead of the previous
global-set-key:
(global-set-key "\C-^bdc" 'backward-delete-char)
METHOD #3:
Put this in your .emacs file:
(global-set-key "\C-h" 'backward-delete-char-untabify)
(global-set-key "\C-xh" 'help-command) ; override mark-whole-buffer
This makes C-h (the backspace key as well as the Control-h key
sequence) delete characters backward, and shifts responsibility for
help to C-x h. This fix will work for GNU Emacs as well as Epoch.
If you don't like to have your backspace key also untabify (i.e.,
convert tabs to spaces on the fly), use this instead of the previous
global-set-key for C-h:
(global-set-key "\C-h" 'backward-delete-char)
METHOD #4:
Outside of Epoch, you can change the X representation of the
BackSpace key into a Delete by doing this:
xmodmap -e "keysym BackSpace = Delete"
This approach has the disadvantage of not being done inside
of Epoch for those who prefer customizing lisp code to
customizing their system environments.
[Thanks to Joe Wells (jbw@bigbird.bu.edu).]
F-2: How do I make the keys marked "Page Up" and "Page Down" on an IBM
Selectric-style keyboard do their jobs?
Using the key rebinding facility discussed below, you can put the
following two lines in your .emacs file:
(rebind-key "Prior" nil "\M-v") ; Note Prior, not Page Up.
(rebind-key "Next" nil "\C-v") ; Note Next, not Page Down.
Similarly, if you want "Home" and "End" to go to the beginning and end
of the current buffer respectively, do:
(rebind-key "Home" nil "\M-<")
(rebind-key "End" nil "\M->")
F-3: How do I bind keys in the numeric keypad on an IBM Selectric-style
keyboard to their obvious functions (arrow movement, home, end, etc.)?
Put these lines in your .emacs file:
(rebind-key "KP_Home" nil "\M-<")
(rebind-key "KP_Up" nil "\C-p")
(rebind-key "KP_Prior" nil "\M-v")
(rebind-key "KP_Left" nil "\C-b")
(rebind-key "KP_Begin" nil "\C-l")
(rebind-key "KP_Right" nil "\C-f")
(rebind-key "KP_End" nil "\M->")
(rebind-key "KP_Down" nil "\C-n")
(rebind-key "KP_Next" nil "\C-v")
NOTE: This makes the keypad arrow keys work, the keypad PgUp/PgDown
work, and the keypad Home/End go to the beginning/end of the
buffer.
Other keys appropriate for rebinding in a Selectric-style numeric
keypad include KP_Divide, KP_Multiply, KP_Subtract, KP_Add, KP_Enter,
KP_Insert, and KP_Delete; see below for more details.
F-4: What general facilities does Epoch provide for rebinding function
keys, and how do I use them?
The rebind-key function allows you to rebind any keycode to a new
string. An example is probably best to show how this works. To make
the key marked "Page Up" (on IBM Selectric-style keyboards) actually
move the current buffer up a page (like M-v), do:
(rebind-key "Prior" nil "\M-v")
Note that "Prior" is the actual X keysym name for the "Page Up" key;
to see the X keysym names, refer to /usr/include/X11/keysymdef.h (or
$OPENWINHOME/include/X11/keysymdef.h on OpenWindows systems).
(Disregard the leading XK_ in keysymdef.h's definitions for Epoch's
purposes; thus, X11's "XK_Prior" becomes Epoch's "Prior". To find out
which keys are actually active for your keyboard and server, use the
command "xmodmap -pk".)
The standard function keys are named F1 through F12, the function keys
on the left side of a Sun-style keyboard are named L1 through L10, and
so on. The keypad apparently cannot be addressed by KP_0 through KP_9
but must be referenced by names like KP_Left, KP_Up, and so on.
A simple two-step approach for binding function keys to commands
involves binding a key to an arbitrary GNU Emacs key encoding (in the
following example, C-^ k 1) and then binding that encoding to the
command. An example that binds the function key marked "End" to an
arbitrary command (in this case, end-of-line) follows:
(rebind-key "End" 0 "\C-^k1")
(global-set-key "\C-^k1" 'end-of-line)
Note that the above is only an example of the two-step method; if you
really want "End" to do 'end-of-line, do the obvious instead:
(rebind-key "End" 0 "\C-e")
The rebind-key command allows modifiers (e.g. shift, control) to be
specified; the following example binds Control-"End" to an arbitrary
command (in this case, end-of-buffer):
(rebind-key "End" 'control "\C-^Ck1")
(global-set-key "\C-^Ck1" 'end-of-buffer)
For an example of these principles in action, see the file
'amc/keys.el' in the 3.2 contrib directory (which is still up for
anonymous ftp at cs.uiuc.edu as a separate .tar.Z file) or at
ftp.ncsa.uiuc.edu in /outgoing/marca/epoch.
[Thanks to Alan Carroll (carroll@cs.uiuc.edu).]
F-5: I want Epoch to disregard accidental presses of function keys that
insert spurious characters (e.g., "-1~") into the buffer. Does Epoch
perchance have some magical way to make this happen?
Perchance, Epoch does. The variable epoch::function-key-mapping, if
set to nil, makes Epoch disregard such keys, unless they're explicitly
rebound with rebind-key. You can set this in your .emacs file like
so:
(setq epoch::function-key-mapping nil)
[Thanks to Alan Carroll (carroll@cs.uiuc.edu).]
F-6: How do I turn the keyboard bell off?
The easy answer is to put the following line in your .emacs file:
(setq epoch::bell-volume -50)
Another possibility, if you like visual bells (i.e., having the
entire edit window flash reverse video instead of hearing an audible
beep), is to put the following in your .emacs file:
(epoch::set-bell t)
The malicious answer, for those who, like myself, never want to hear
another beep for the rest of time, is to modify src/x11term.c in the
Epoch distribution; just change the line that reads:
XBell (xs->display,volume);
To:
/* XBell (xs->display,volume); */
Then recompile.
* === HIGHLIGHTING ===
G-1: How do I highlight regions of text in a buffer with different styles?
While the current Epoch distribution contain plenty of support for
zones and styles, the enabling mechanisms that make syntax-directed
highlighting possible, no intrinsic support for such highlighting is
provided. This situation may change in subsequent releases of Epoch.
Three packages that provide varying degrees of support for
syntax-directed highlighting are:
tek-epoch-stuff-1.1.tar.Z (available from archive.cis.ohio-state.edu
or ftp.ncsa.uiuc.edu, see below), which supports comment highlighting
in source code as well as support for the Info documentation browser,
various mail and news packages, and manual pages.
marc-hilite.tar.Z (available from ftp.ncsa.uiuc.edu, see below), which
is an extension to tek-epoch-stuff made by Yours Truly to allow
per-mode regexp-based highlighting. Thus, you can have all
#define's/#endif's in red, setq's and defun's in green,
int/float/double's in yellow, '-->'s in purple, and so on.
(A very preliminary version of a reimplementation of marc-hilite,
called Lightbrite, can be found here also; it includes partial
documentation. Hopefully Lightbrite proper will be completed
sometime this summer.)
hilit.el.Z (most recent version was posted to gnu.emacs.souces and is
available from ftp.ncsa.uiuc.edu, see below), which provides
source-code comment and regexp highlighting.
G-2: When I use one of the highlighting packages to highlight comments,
why doesn't the mouse region selection style always take precedence
over the comment style?
This question refers to the following: if you highlight all comments
in blue and the mouse region selection style is the default
(underlining), the comments will be blue and not underlined when they
are within the selection zone.
For the full technical explanation, refer to the Epoch online info
section entitled Epoch->Zones->Zone Plotting. In short, Epoch has to
have some way to decide which style takes precedence for a given
character when that character is in two or more zones (since merging
styles in that case is neither supported nor particularly logical).
Essentially, the algorithm Epoch uses to decide which style is used
will only have the 'proper' results (i.e., the underlining will take
precedence) if the selected region lies entirely within a comment;
the selected area will be underlined and the rest of the comment will
be blue. In a different case, e.g. if the selected region starts
before the comment, the comment will be blue and not underlined even
though it's selected.
Following is a patch devised by Renaud Marlet
(Renaud.Marlet@sophia.inria.fr) that may be useful to people who
want to try hacking Epoch's C source to make the drag style always
come out on top; as usual, your mileage may vary. (This patch will
work as-is with 4.0b0 or later; if you're still using a 4.0 alpha
level you'll have to change "drag-zone" to "drag-button".)
-- BEGIN PATCH -----------------------------------------------------
*** button.c.old Sat Mar 28 01:55:11 1992
--- button.c Sat May 9 23:14:19 1992
***************
*** 695,700 ****
--- 695,709 ----
register Lisp_Object buttons, next_b = Qnil;
register int first = -1,last = -1;
+ #if 1
+ /* DRAG-BUTTON */
+ struct Lisp_Button *drag_button;
+ Lisp_Object drag_button_value;
+
+ drag_button_value = XSYMBOL(intern("drag-zone"))->value;
+ drag_button = XBUTTON(BUTTONP(drag_button_value) ? drag_button_value : Qnil);
+ #endif
+
/* Possible start in list */
if (b_first) next_b = *b_first;
if (NIL(next_b)) next_b = b->buttons;
***************
*** 722,727 ****
--- 731,742 ----
if (b_end) *b_end = last;
if (b_return) *b_return = buttons;
if (b_first) *b_first = buttons;
+
+ #if 1
+ /* DRAG-BUTTON */
+ if (button == drag_button)
+ return button->style;
+ #endif
for (buttons2 = button->next; !NIL(buttons2); buttons2 = button2->next)
{
-- END PATCH -------------------------------------------------------
* === EPOCH-SPECIFIC ELISP PACKAGES ===
H-1: Where can I find Epoch-specific elisp packages?
Try the following anonymous FTP sites for Epoch-specific elisp
packages:
aix370.rrz.uni-koeln.de (134.95.132.2)
/gnu/emacs
archive.cis.ohio-state.edu (128.146.8.52)
/pub/gnu/emacs/elisp-archive/elisp
[This is main elisp-archive site; it is mirrored at
ftp.uu.net (137.39.1.9) in /languages/elisp, among other places.
Not much Epoch code exists here; in particular, a much more
recent version of hilit.el.Z is at ftp.ncsa.uiuc.edu, below.]
ftp.cs.buffalo.edu (128.205.32.3)
/pub
[This is the distribution site for the new Dired, which will
be included in Emacs v19; it includes support for Epoch.]
ftp.ncsa.uiuc.edu (141.142.20.50)
/outgoing/marca/epoch
[This is Yours Truly's archive of Epoch code; the
README file gives an overview of what I consider to be
'essential' Epoch packages, and all of those packages
are available there. If I get out of date on any of
these and you notice it, please let me know.]
icsi-ftp.berkeley.edu (128.32.201.55)
/pub/elisp
ireq-robot.hydro.qc.ca (131.195.2.130)
/pub/emacs/lisp
[This is the distribution site for the IMOUSE package.]
wilma.cs.brown.edu (128.148.31.66)
/pub/hyperbole
[This is the distribution site for the Hyperbole hypertext
system, which includes support for Epoch. Also available
separately from here is the wrolo 'rolodex' package, one of the
components of Hyperbole.]
Also see the contrib directory in the actual Epoch distribution.
Also watch the gnu.emacs.sources Usenet newsgroup, as well as the
epoch newsgroup/mailing list (see below).
Please volunteer the names of any sites not on this list.
H-2: Where can I get a PostScript version of the Epoch manual?
Use anonymous ftp to cs.uiuc.edu (128.174.252.1) in
/pub/epoch-files/epoch; get epoch-4.0.epoch-man.ps.Z.
* === OTHER RESOURCES ===
I-1: What Epoch newsgroups/mailing lists are out there?
The Usenet newsgroup gnu.epoch.misc and the mailing list
epoch@cs.uiuc.edu are one and the same. To join the mailing list,
send a request to epoch-request@cs.uiuc.edu. (It is best to read the
newsgroup if you have access to it, to save on network resources and
the mailing list maintainer's time.)
For more general information about Epoch and GNU Emacs, see the list
of resources at the start of this file.
* === ADVANCED QUESTIONS ===
[NOTE: This section is at the bottom since the questions involve actual
source-code hacking or problems with specific machine configurations
that Yours Truly cannot personally verify and whose solutions will
probably involve some sort of hacking.]
Z-1: Under tvtwm, when I create a new screen, it is mapped relative to
the origin on the virtual desktop, not the window I currently have on
the desktop. How do I fix this?
This problem is present in Epoch4.0 beta level 0, although it appears
to be fixed as of beta level 1. The simple solution for beta level 0
is to change the line in src/screen.c from
size_hints.flags |= (gmask & (XValue|YValue)) ? USPosition : PPosition;
to
size_hints.flags |= PPosition;
And set:
UsePPosition "on"
in your .twmrc or .tvtwmrc file.
[From Philippe Bondono (bondono@vnet.ibm.com).]
Z-2: Epoch 4.0b1 (or later) (as well as GNU Emacs 18.58) built under
HP-UX 8.0 and 8.05 apparently don't recognize the Meta key as Alt.
How can this be fixed?
Bob Fisher (bob@fisher.depaul.edu) suggests the following command,
executed once per session, prior to starting Epoch:
xmodmap -e "clear mod1" \
-e "add mod1 = Meta_R" \
-e "add mod1 = Meta_L"
Z-3: Why do all of Epoch's colors go away when Epoch is run on an X11R5
server with multiple screens?
Dana Chee (dana@thumper.bellcore.com) reports that X11R5 has a new
SCREEN_RESOURCES property that allows/forces colors to only appear for
color screens, but Epoch doesn't look at this property. This bug
will only bite you if you're running on a multi-screen system. If it
happens, there is a workaround; see the xrdb manpage and the xrdb
option '-all'.
Z-4: Even though Epoch 4.0b1 includes fixes to keep the cursor from
disappearing (as it does on occasion, especially in a shell, in
Epoch 4.0b0), the cursor still disappears sometimes. Is there a fix
for this?
First, Epoch 4.0p0 has additional fixes for this problem which
presumably will supersede the fix given below.
According to Michael Thome (mthome@bbn.com), activating the XFlush
code at the end of the CursorToggle routine in 4.0b1's x11term.c seems
to fix the problem. (In other words, '#if 0' on line 737 of x11term.c
should be changed to '#if 1'.)
Z-5: I want Epoch to iconify itself immediately upon starting. To this
end, in my .emacs file I have (epoch::iconify-screen). But most
(~95%) of the time this doesn't work. How can I make it work all the
time?
The technical explanation appears to be that the initial Epoch
minibuffer and edit screen are unmapped after creation to set some
properties for ICCCM purposes and then remapped, and this interferes
with the iconification process.
One workaround (that may or may not work for you) is to add a sit-for
command to your .emacs file right before epoch::iconify-screen; for
example:
(sit-for 15)
(epoch::iconify-screen)
[Thanks to Bobby Bodenheimer (bobby@hot.caltech.edu), Simon Kaplan
(kaplan@cs.uiuc.edu), and Chris Love (love@cs.uiuc.edu).]
Z-6: I'm running Epoch on an IBM RISC System/6000 under AIX 3.1 or 3.2
and the arrow keys don't work. How do I fix this?
Marc's solution is to hack the Epoch source code; GNU Emacs and Epoch
have, for a reason which I have not yet determined, decided to exclude
AIX users specifically from those who are entitled to use arrow keys.
Go into src/x11term.c and search for 'case KeyPress'. Shortly
thereafter you will see '#ifndef AIX'. Remove this directive and its
corresponding #endif, and the arrow keys will work.
It should also be possible (although I have not tried it since I don't
use RS/6000's currently) to rebind the arrow keys with rebind-key to
C-n/C-p/C-b/C-f without modifying the source code.
Z-7: Every once in a while Epoch seemingly loses control of the mouse;
I can no longer move from window to window or move the cursor within
a window with the mouse. How do I fix this?
A temporary fix for this problem is to enter the following elisp code
directly into Epoch (via M-x eval-expression):
(setq epoch::event-handler 'event::handler)
[Thanks to Chris Love (love@cs.uiuc.edu).]
* === CONTRIBUTORS ===
--> Thanks to the following people for contributing to this FAQ
(whether knowingly or not :-) as well as anyone else I may have
inadvertently forgotten to list:
Bobby Bodenheimer (bobby@hot.caltech.edu)
Philippe Bondono (bondono@vnet.ibm.com)
Martin Boyer (mboyer@ireq-robot.hydro.qc.ca)
Alan Carroll (carroll@cs.uiuc.edu)
Nhi Casey (nhi@ariel.aero.org)
Dana Chee (dana@thumper.bellcore.com)
Alice Chen (alice@innerdoor.austin.ibm.com)
Bob Fisher (bob@fisher.depaul.edu)
Jerry Graves (jerry@math.ep.utexas.edu)
Dev Joneja (dj7@cunixf.cc.columbia.edu)
Simon Kaplan (kaplan@cs.uiuc.edu)
Tor Lillqvist (tml@tik.vtt.fi)
Chris Love (love@cs.uiuc.edu)
Simon Marshall (S.Marshall@sequent.cc.hull.ac.uk)
Rob McCool (robm@ncsa.uiuc.edu)
Michael Thome (mthome@bbn.com)
Bob Weiner (rsw@cs.brown.edu)
Joe Wells (jbw@bigbird.bu.edu)
Thanks particularly to Jerry Graves for much FAQ work.
Thanks also to all the people who have spent time and effort
developing Epoch and associated packages.
----------------------------------------------------------------------
--
Marc Andreessen
Software Development Group
National Center for Supercomputing Applications
marca@ncsa.uiuc.edu
Technical Details
Message-ID:
<MARCA.92May31164809@wintermute.ncsa.uiuc.edu>Path:
arc1-ca.pugleaf.net!sparky!uunet!wupost!m.cs.uiuc.edu!ux1.cso.uiuc.edu!news.cso.uiuc.edu!ux3.cso.uiuc.edu!marcaShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: fc7ee,c07bd08b3c1ccc5a,start X-Google-Attributes: gidfc7ee,public X-Google-Thread: f7ded,c07bd08b3c1ccc5a,start X-Google-Attributes: gidf7ded,public X-Google-ArrivalTime: 1992-05-31 09:04:36 PST Newsgroups: gnu.epoch.misc,news.answers Path: arc1-ca.pugleaf.net!sparky!uunet!wupost!m.cs.uiuc.edu!ux1.cso.uiuc.edu!news.cso.uiuc.edu!ux3.cso.uiuc.edu!marca From: marca@ncsa.uiuc.edu (Marc Andreessen) Subject: Epoch Frequently Asked Questions (FAQ) Message-ID: <MARCA.92May31164809@wintermute.ncsa.uiuc.edu> Summary: Epoch FAQ for June '92 Sender: usenet@news.cso.uiuc.edu (Net Noise owner) Reply-To: marca@ncsa.uiuc.edu Followup-To: gnu.epoch.misc Organization: Nat'l Center for Supercomputing Applications Approved: news-answers-request@mit.edu Date: Sun, 31 May 1992 21:48:09 GMT Keywords: Epoch FAQ Lines: 1359