stumpwm.texi.in

\input texinfo   @c -*-texinfo-*-
@c %**start of header
@setfilename stumpwm.info
@settitle StumpWM Manual
@setchapternewpage odd
@c %**end of header

@dircategory X11
@direntry
* StumpWM: (stumpwm).       A Common Lisp window manager
@end direntry

@ifinfo
This is the Stump Window Manager user manual.

Copyright @copyright{} 2000-2008 Shawn Betts
Copyright @copyright{} 2014 David Bjergaard

Permission is granted to make and distribute verbatim
copies of this manual provided the copyright notice and
this permission notice are preserved on all copies.

@ignore
Permission is granted to process this file through TeX
and print the results, provided the printed document
carries a copying permission notice identical to this
one except for the removal of this paragraph (this
paragraph not being relevant to the printed manual).

@end ignore
Permission is granted to copy and distribute modified
versions of this manual under the conditions for
verbatim copying, provided also that the sections
entitled ``Copying'' and ``GNU General Public License''
are included exactly as in the original, and provided
that the entire resulting derived work is distributed
under the terms of a permission notice identical to this
one.

Permission is granted to copy and distribute
translations of this manual into another language,
under the above conditions for modified versions,
except that this permission notice may be stated in a
translation approved by the Free Software Foundation.
@end ifinfo

@iftex
@kbdinputstyle code
@end iftex

@titlepage
@sp 10
@titlefont{The Stump Window Manager}
@author Shawn Betts, David Bjergaard

@page
@vskip 0pt plus 1filll
Copyright @copyright{} 2000-2008 Shawn Betts
Copyright @copyright{} 2014 David Bjergaard

Permission is granted to make and distribute verbatim
copies of this manual provided the copyright notice and
this permission notice are preserved on all copies.

Permission is granted to copy and distribute modified
versions of this manual under the conditions for
verbatim copying, provided also that the sections
entitled ``Copying'' and ``GNU General Public License''
are included exactly as in the original, and provided
that the entire resulting derived work is distributed
under the terms of a permission notice identical to this
one.

Permission is granted to copy and distribute
translations of this manual into another language,
under the above conditions for modified versions,
except that this permission notice may be stated in a
translation approved by the Free Software Foundation.
@end titlepage

@node Top, Introduction, (dir), (dir)

@ifinfo
This document explains how to use The Stump Window Manager.
@end ifinfo

@menu
* Introduction::
* Key Bindings::
* Commands::
* Message and Input Bar::
* Windows::
* Frames::
* Mode-line::
* Groups::
* Screens::
* Minor Modes::
* Internals::
* Interacting With Unix::
* Interacting With X11::
* Miscellaneous Commands::
* Colors::
* Hooks::
* Modules::
* Hacking::
* Advanced Configuration::
* Command and Function Index::
* Variable Index::

@detailmenu
 --- The Detailed Node Listing ---

Introduction

* Starting StumpWM::
* Basic Usage::
* Basic Concepts::
* Manipulating Frames and Windows::
* Interacting with the Lisp process::
* Init File::
* Contact the StumpWM developers::

Basic Concepts

* Screens and Heads::
* Group Basics::
* Floating Group Basics::
* Dynamic Group Basics::
* Frame Basics::
* Window Basics::
* System Trays and the Mode Line::

Manipulating Frames and Windows

* Moving Between Frames::
* Manipulating Windows::

Key Bindings

* List of Default Keybindings::
* Binding Keys::
* Modifiers::
* Remapped Keys::

Commands

* Writing Commands::
* StumpWM Types::

Message and Input Bar

* Customizing The Bar::
* Using The Input Bar::
* Programming The Message Bar::
* Programming the Input Bar::

Windows

* Window Marks::
* Customizing Window Appearance::
* Controlling Raise And Map Requests::
* Programming With Windows::
* Rule Based Window Placement::
* Window Selection Expressions::
Frames

* Interactively Resizing Frames::
* Frame Dumping::

Groups

* Customizing Groups::

Screens

* External Monitors::
* Programming With Screens::

Minor Modes

* Minor Mode Example::
* Programming With Minor Modes::

Internals

* IO Loop::

Miscellaneous Commands

* Menus::
* StumpWM's Data Directory::
* Debugging StumpWM::
* Sending a Bug Report::
* Timers::
* Getting Help::

Colors

* Behind The Scenes Look At Colors::

Modules
* Writing Modules::

Hacking

* General Advice::
* Adding Documentation and Editing This Manual::
* Using git with StumpWM::
* Sending Patches::

@end detailmenu
@end menu

@node Introduction, Key Bindings, Top, Top
@chapter Introduction
StumpWM is a manual, tiling X11 window manager written entirely in
Common Lisp. Unlike traditional window managers, StumpWM places
windows in order to maximize the amount of the screen used.  The
window layouts managed by StumpWM are defined by the user in much the
same way that windows are managed by GNU screen, or emacs.

Before StumpWM, there was ratpoison, another tiling window manager
written entirely in C.  StumpWM grew out of the authors' frustration
with writing ratpoison in C. Very quickly we realized we were building
into ratpoison lispy-emacs style paradigms. StumpWM's goals are
similar to ratpoison's but with an emphasis on customizability,
completeness, and cushiness.

@menu
* Starting StumpWM::
* Basic Usage::
* Basic Concepts::
* Manipulating Frames and Windows::
* Interacting with the Lisp process::
* Init File::
* Contact the StumpWM developers::
@end menu

@node Starting StumpWM, Basic Usage, Introduction, Introduction
@section Starting StumpWM
There are a number of ways to start StumpWM but the most straight
forward method is as follows. This assumes you have a copy of the
StumpWM source code and are using the @samp{SBCL} Common Lisp
environment.

@enumerate
@item
Install the prerequisites and build StumpWM as described in
@file{README}.  This should give you a @file{stumpwm} executable.

@item
In your @file{~/.xinitrc} file include the line
@command{/path/to/stumpwm}. Remember to replace @samp{/path/to/}
with the actual path.

@item
Finally, start X windows with @command{startx}. Cross your
fingers. You should see a @samp{Welcome To the Stump Window Manager}
message pop up in the upper, right corner. At this point, you have
successfully started StumpWM.
@end enumerate

@@@ stumpwm

@node Basic Usage, Basic Concepts, Starting StumpWM, Introduction
@section Basic Usage
Once you have StumpWM up and running, the first thing you might want
to do is start @command{emacs}. Type @kbd{C-t e}, or in other words
@kbd{Control + t} followed by @kbd{e}. Now perhaps you want an
@command{xterm}. Type @kbd{C-t c}. Now you have some programs running.

To see a list of windows StumpWM is managing, type @kbd{C-t w}. The
highlighted window is the one you're looking at right now. It's the
focused window.

All of StumpWM's keys are bound to named commands, which can be
executed not only by keys but also from the input bar. Type @kbd{C-t
;} to open a command prompt. Now type @command{time} and press
return. Note, @command{time} can also be called by typing @kbd{C-t a}.
Throughout this manual you'll find definitions for commands,
functions, and variables. Any command you see in this manual can be
executed from the input bar or bound to a key.

At this point you probably want to switch back from your new
@command{xterm} to @command{emacs}. Type @kbd{C-t C-t}. This runs the
@command{other} command. Type it again and you're back to xterm.

Perhaps you'd like to see @command{emacs} and @command{xterm}
side-by-side. Type @kbd{C-t s}. You have now split the screen into 2
@command{frames}. For more information see @ref{Frames}. To switch to the
empty frame type @kbd{C-t TAB}. Now let's pull the xterm window into
this empty frame. Type @kbd{C-t w} for a window listing. Find the
@command{xterm} window. See the number beside it? Type @kbd{C-t} followed
by @command{xterm}'s window number.

Another common activity is browsing the internet. Type @kbd{C-t !}.
The input bar pops up again. You can now run a shell command.  Let's
start a web browser: type @command{firefox} into the input bar and press
return.

Unfortunately, @command{firefox} probably isn't wide enough because it's
in one of the frames. Type @kbd{C-t Q} to remove all frames but the
current one and resize it to fit the screen.

For a full list of key bindings, see @ref{List of Default Keybindings}.

@node Basic Concepts, Manipulating Frames and Windows, Basic Usage, Introduction
@section Basic Concepts

An introduction to some of the basic concepts used by StumpWM.

@menu
* Screens and Heads::
* Group Basics::
* Floating Group Basics::
* Dynamic Group Basics::
* Frame Basics::
* Window Basics::
* System Trays and the Mode Line::
@end menu

@node Screens and Heads, Group Basics, Basic Concepts, Basic Concepts
@subsection Screens and Heads
A screen is an Xlib concept representing a section of video memory
onto which physical monitors, called ``heads'', are mapped. A screen can
be thought of as an abstract rectangle containing all the heads
arranged in a particular layout.

With most modern systems, you'll only have a single screen no matter
how many heads are connected to your computer. Each head will have its
own frame, and you can move between heads using the normal frame
movement commands.

The layout of the heads within the screen can be specified in one of two
ways: either at startup using your system's Xorg configuration files, or
on the fly using tools like XRandR. If the computer is booted with
multiple monitors attached, but without specifying a layout for them,
they will all show identical output.

StumpWM will attempt to detect the layout of the heads once at
startup, or any time a RandR command is issued.

In rarer setups you may have multiple screens, with one head per
screen. That means that you'll move between heads using screen
movement commands (`snext', `sprev', and `sother') rather than frame
movement commands.

@node Group Basics, Floating Group Basics, Screens and Heads, Basic Concepts
@subsection Group Basics
A group is usually referred to as a ``desktop'' or ``workspace'' in other
window managers. StumpWM starts with a single group, called ``Default''.
Each group has its own configuration of frames and windows that is
separate from and independent of other groups. You can't have
different groups display in different monitors: when you switch
groups, all monitors switch to that group.

Each group contains an ordered list of frames.

@node Floating Group Basics, Dynamic Group Basics, Group Basics, Basic Concepts
@subsection Floating Group Basics
Within a floating group, windows behave more like they do in traditional
window managers: rather than being arranged into frames, they each have
their own box, which can be freely resized and repositioned, and allowed
to overlap. Each window has a thicker border at the top. Left click in
this border and drag to move the window, or right click and drag to
resize it.

A modifier key can be used to perform the move and resize operations
by clicking in the window itself instead of on its top border. The
default modifier is super, and can be configured with
@var{*float-window-modifier*}.

### *float-window-modifier*

Most of the window-switching commands listed below do not function in
a floating group. You're restricted to `other', the `select-window-*'
commands, and `windowlist'.


@node Dynamic Group Basics, Frame Basics, Floating Group Basics, Basic Concepts
@subsection Dynamic Group Basics
Within a dynamic group, windows are organized into a @dfn{master window} and a
@dfn{window stack}, with one of each per head. When a new window is added to a
head within a dynamic group, that heads master window is pushed onto that heads
window stack, and the new window becomes the master window. When a head becomes
to full (ie more windows cannot be placed upon the stack) a overflow policy is
used to determine which window to move to a separate head or group. 

!!! gnew-dynamic

!!! gnewbg-dynamic

!!! rotate-windows

!!! change-layout

!!! change-default-layout

!!! change-split-ratio

!!! change-default-split-ratio

!!! retile

!!! select-floating-window

### *rotation-focus-policy*

When defining commands, anything restricted to tiling groups will also be
active in dynamic groups. To fully restrict it to tiling groups, call the
function dyn-blacklist-command on the command in question.


@node Frame Basics, Window Basics, Floating Group Basics, Basic Concepts
@subsection Frame Basics
Frames are the boxes within which windows are displayed. StumpWM
starts with a single frame per head, meaning that each monitor shows a
single window, full screen. If you want to see windows side-by-side,
you can ``split'' this frame in two, either vertically or horizontally.
These frames can be further split, creating nested boxes.

Technically speaking, frames live within a ``frame tree''. When you
split a frame, the command actually creates @emph{two} new frames
side-by-side within the original parent frame. This has implications for
the behaviour of many commands that interact with frames.

Within this frame tree model, all frames either contain other frames,
or contain windows, or are empty. The command `fclear' will hide all of a
frame's windows and show the background.

@node Window Basics, System Trays and the Mode Line, Frame Basics, Basic Concepts
@subsection Window Basics
Windows are created by programs to display their output. They take the
shape of the frame in which they are created. The windows within a
frame are ordered by how recently that window was focused. Only the
top window in the stack is visible.

@node System Trays and the Mode Line,  , Window Basics, Basic Concepts
@subsection System Trays and the Mode Line
Many users choose to sacrifice a little screen real-estate to display
some generally useful information: the current time and date, wireless
network connections, the names of open windows, etc. StumpWM allows
you to display this information in a bar across either the top or the
bottom of the screen. There are two ways to do this: using external
programs called system trays, or using StumpWM's own mode line.

System trays are a special kind of X window. They advertise to running
programs that they are available for embedding icons or notifications
from those programs. They often also display clickable icons for each
open window. Common tray programs include the GNOME panel or KDE's
kicker, or simpler programs such as stalonetray. Simply starting one
of these programs is usually enough for StumpWM to detect it, place it
correctly, and allow it to function normally.

The mode line, a concept borrowed from Emacs, is a built-in part of
StumpWM. It is essentially a string of text that can include a variety
of information about your current session, including the names of
current groups and windows. Several modules provide for
different types of information. @xref{Mode-line, Mode Line}, (and the modules
directory) for more.

@node Manipulating Frames and Windows, Interacting with the Lisp process, Basic Concepts, Introduction
@section Manipulating Frames and Windows
Frames and windows are concepts borrowed from Emacs and the GNU Screen
program, and should be familiar to users of those programs. Others may
find the terms a little confusing. In other window managers, a
``window'' usually refers to a bounded box on the screen, showing output
from a single program. StumpWM splits this into two concepts: the
``frame'' is the bounded box, the ``window'' is the visible output of a
program.

One frame can contain many windows. As new windows are created, they
appear at the top of the window-stack of the current frame. This is
also a little different from other tiling window managers, many of
which automatically create new frames for new windows.

Both frames and windows are ordered by when they were last focused. In
the following commands and documentation, the terms ``next'' and
``previous'' refer to this order. ``Other'' refers to the most-recently
focused object. Calling ``other'' commands multiple times will bounce
back and forth between the two most recent objects.

By default, StumpWM starts with a single group, called ``Default'',
which contains one full-screen frame per head. You can split individual
frames horizontally or vertically using the `hsplit' and `vsplit'
commands, bound to ``C-t S'' and ``C-t s'' by default. When a frame is
split, the next-most-recently-focused window is pulled into the new
frame. @xref{Frames}, and @ref{Windows}, for a complete listing of
commands.


@menu
* Moving Between Frames::
* Manipulating Windows::
@end menu

@node Moving Between Frames, Manipulating Windows, Manipulating Frames and Windows, Manipulating Frames and Windows
@subsection Moving Between Frames
Once you have multiple frames, you can move between them in various
ways:

@itemize
@item
@command{fnext} (@kbd{C-t o} or @kbd{C-t TAB}) jumps to the next frame in
the current group's frame list.

@item
@command{fother} (@kbd{C-t M-TAB}) jumps to the last frame that had
focus.

@item
@command{fselect} (@kbd{C-t f}) displays numbers on each visible frame: hit a
number key or click it to move to that frame.

@item
@command{move-focus} (@kbd{C-t <arrow key>}) focus the frame in the direction of
the arrow key pressed.

@item
@command{sibling} (unbound by default) focus the frame from which the
current frame was split.

@end itemize

@node Manipulating Windows,  , Moving Between Frames, Manipulating Frames and Windows
@subsection Manipulating Windows

Some commands change which window is currently focused, some move
windows between frames, and some may do both at once.

There are two general ways to move focus between windows: either
between windows belonging to the current frame, or between all windows
within the current group. Within a single frame:

@itemize
@item
@command{next-in-frame} (@kbd{C-t C-M-n}) focus the next window in the current
frame's list of windows.

@item
@command{prev-in-frame} (@kbd{C-t C-M-p}) focus the previous window in the
current frame's list of windows.

@item
@command{other-in-frame} (@kbd{C-t M-t}) focus the most recently focused window
in the current frame's list of windows.

@item
@command{frame-windowlist} (unbound by default) display a menu of windows in
the currently-focused frame, and allow the user to choose one.
Alternately, the command @command{frame-windows} will simply display the
list of window names, with no menu choice available.

@end itemize

Within the current group, the following commands will go straight to
the specified window. They will never move a window from its original
frame, and so may result in focus switching frames.

@itemize
@item
@command{next} (@kbd{C-t M-n}) focus the next window in the current group.

@item
@command{prev} (@kbd{C-t M-p}) focus the previous window in the current group.

@item
@command{other} or @command{other-window} (unbound by default) focus the
most recently focused window in the current group.

@item
@command{next-urgent} (@kbd{C-t C-u}) focus the next window that has marked
itself ``urgent''.

@item
@command{select} or @command{select-window} (@kbd{C-t '}) prompt for the
title of a window and focus it. Works with partial completion of the
title.

@item
@command{select-window-by-name} (unbound by default) prompt for the
title of a window and focus it. Requires the window title to be entered
exactly.

@item
@command{select-window-by-number} (@kbd{C-t <number>}) choose a window by
number.

@item
@command{windowlist} (@kbd{C-t "}) display a menu of windows in the
currently-focused group, and allow the user to choose one.

@end itemize

The following commands always keep the current frame focused. If the
selected window is not in the current frame, it will be pulled there
from wherever it is (hence the ``pull'' naming scheme).

@itemize
@item
@command{pull} or @command{pull-window-by-number} (@kbd{C-t C-<number>})
pull the numbered window into the current frame.

@item
@command{pull-hidden-next} (@kbd{C-t n} or @kbd{C-t SPC}) pull the next
currently undisplayed window in the window list into the current frame.

@item
@command{pull-hidden-previous} (@kbd{C-t p}) pull the previous currently
undisplayed window in the window list into the current frame.

@item
@command{pull-hidden-other} (@kbd{C-t C-t}) pull the most recently
focused, currently undisplayed window into the current frame.

@end itemize

The following commands move the current window from one frame to
another, bringing focus with them.

@itemize
@item
@command{move-window} (@kbd{C-t M-<arrow>}) move the currently focused
window in the direction indicated by the arrow key.

@item
@command{exchange-direction} (unbound by default) prompt for a
direction, then swap the currently focused window with the top window of
the frame in that direction.

@end itemize

@node Interacting with the Lisp process, Init File, Manipulating Frames and Windows, Introduction
@section Interacting with the Lisp process

Since StumpWM is a Lisp program, there is a way for you to evaluate
Lisp code directly, on the same Lisp process that StumpWM is running on.
Type @kbd{C-t :} and an input box will appear. Then type some Lisp
expression.

When you call @command{eval} this way, you will be in the STUMPWM-USER
package, which imports all the exported symbols from the main STUMPWM
package.

@table @code
@item *mode-line-border-width*
Reads the value of @var{*mode-line-border-width*}.

@item (setf *mode-line-border-width* 3)
Sets the variable @var{*mode-line-border-width*} to 3.

@item (set-prefix-key (kbd "C-M-H-s-z"))
Calls the @code{set-prefix-key} function (and sets a new keyboard prefix)
@end table

@node Init File, Contact the StumpWM developers, Interacting with the Lisp process, Introduction
@section Init File

Like other window managers, StumpWM's configuration and startup
state can be controlled by an initialization file. Unlike other
window managers, StumpWM's init is not limited to changing settings
and keybindings. The init file is itself a Common Lisp program
running in a Common Lisp environment, so you can write your own
hacks and make them a part of your StumpWM experience.

On launch, StumpWM searches for an init file of different names and
locations on your system, and will use the first one found in this
order:

@itemize
@item `~/.stumpwmrc' is the classic UNIX-style configuration name;

@item `~/.stumpwm.d/init.lisp' is an Emacs-style location and name;

@item `~/.config/stumpwm/config' is the XDG standard;

@item `/etc/stumpwmrc' is a system-wide file giving all users a
standardized environment.
@end itemize

StumpWM includes a basic `sample-stumpwmrc.lisp' in its source
directory. You can use this as a template when you're starting out:
copy it to the above name and location you prefer and edit it to
suit your preferences.

It is possible to split your initialization among multiple files,
if you call the additional files from within an init file matching
the names and locations listed above.

### *processing-existing-windows*

@node Contact the StumpWM developers,  , Init File, Introduction
@section Contact the StumpWM developers
The StumpWM home page is @url{http://stumpwm.nongnu.org/}.

The StumpWM mailing list is @email{stumpwm-devel@@nongnu.org} which
you can subscribe to at
@url{https://lists.nongnu.org/mailman/listinfo/stumpwm-devel}. It is
the preferred way of contacting developers for questions.  If you have
a bug report or patch, please open an issue or pull request at
@url{https://github.com/stumpwm/stumpwm/issues}.

The StumpWM IRC channel can be found on Freenode at
@uref{irc://irc.libera.chat/#stumpwm, @code{#stumpwm}}.

@node Key Bindings, Commands, Introduction, Top
@chapter Key Bindings
StumpWM is controlled entirely by keystrokes and Lisp commands. It
mimics GNU Screen's keyboard handling. StumpWM's default prefix key is
@kbd{C-t}.

@menu
* List of Default Keybindings::
* Binding Keys::
* Modifiers::
* AltGr Keys::
* Remapped Keys::
@end menu

@node List of Default Keybindings, Binding Keys, Key Bindings, Key Bindings
@section List of Default Keybindings
The following is a list of keybindings.

@table @kbd
@item C-t @var{d}
Select the window with the corresponding digit@tie{}@var{d}

@item C-t C-@var{d}
Pull the window with the corresponding digit@tie{}@var{d} into the current frame

@item C-t n
@itemx C-t C-n
@itemx C-t Space
Go to the next window in the window list

@item C-t p
@itemx C-t C-p
Go to the previous window in the window list

@item C-t '
Go to a window by name

@item C-t "
Select a window from a list and focus the window.

@item C-t C-g
Abort the current command. This is useful if you accidentally hit @kbd{C-t}

@item C-t i
Display information about the current window.

@item C-t f
Select a frame by number

@item C-t s
Split current frame vertically

@item C-t S
Split current frame horizontally

@item C-t k
@itemx C-t C-k
Sends a kill message to the current frame and the running program.

@item C-t K
Kills the current frame and running program; like a @command{kill -9}.

@item C-t c
@itemx C-t C-c
Run an X terminal; by default @command{xterm}

@item C-t e
@itemx C-t C-e
Run Emacs or raise it if it is already running.

@item C-t t
Sends a @kbd{C-t} to the frame; this is useful for applications like Firefox
which make heavy use of @kbd{C-t} (in Firefox's case, for opening a new
tab). This is similar to how GNU screen uses @kbd{C-a a}.

@item C-t w
@itemx C-t C-w
Prints out a list of all the windows, their number, and their name.

@item C-t RET
@itemx C-t C-RET
Show all windows and let the user select one, make that window the
focus.

@item C-t b
@itemx C-t C-b
Banish the mouse point to the lower right corner of the screen.

@item C-t a
@itemx C-t C-a
Display the current time and date, much like the Unix command @command{date}.

@item C-t C-t
Switch to the last window to have focus in the current frame.

@item C-t !
Prompt for a shell command to run via @file{/bin/sh}. All output is discarded.

@item C-t R
If the screen is split into multiple frames, one split will be
undone. If there is only one split, the effect will be the same as @kbd{C-t Q}.

@item C-t o
@itemx C-t TAB
If the screen is split into multiple frames, focus shifts to the
@command{next} frame, where it cycles to the right and then down; analogous
to @kbd{C-x o} in Emacs.

@item C-t F
Display ``Current Frame'' in the frame which has focus.

@item C-t ;
Opens the input box. StumpWM commands can be run from here, and the
input history moved through.

@item C-t :
Opens the input box, but all things typed in here will be sent to the
Common Lisp interpreter where they will be run as Lisp programs; thus,
input should be valid Common Lisp.

@item C-t C-h
@itemx C-t ?
The help. These can be customized using the @var{*help-keys*} variable.

@item C-t -
Hide all frames and show the root window.

@item C-t Q
Removes all splits and maximizes the frame with focus.

@item C-t Up
@itemx C-t Down
@itemx C-t Left
@itemx C-t Right
Shift focus to an adjacent frame in the specified direction. @kbd{C-t Up} will
shift focus up, if possible, @kbd{C-t Down} will shift downwards, etc.

@item C-t v
Prints out the version of the running StumpWM.

@item C-t #
Toggle the mark on the current window

@item C-t m
@itemx C-t C-m
Display the last message. Hitting this keybinding again displays the
message before that, and so on.

@item C-t l
@itemx C-t C-l
redisplay the current window and force it to take up the entire frame.

@item C-t G
Display all groups and windows in each group. For more information see @ref{Groups}.

@item C-t F@var{n}
Jump to the corresponding group @var{n}. @kbd{C-t F1} jumps to group 1 and so on.

@item C-t g g
Show the list of groups.

@item C-t g c
Create a new group.

@item C-t g n
@itemx C-t g C-n
@itemx C-t g SPC
@itemx C-t g C-SPC
Go to the next group in the list.

@item C-t g N
Go to the next group in the list and bring the current window along.

@item C-t g p
@itemx C-t g C-p
Go to the previous group in the list.

@item C-t g P
Go to the previous group in the list and bring the current window along.

@item C-t g '
Select a group by name or by number.

@item C-t g "
Select a group from a list and switch to it.

@item C-t g m
Move the current window to the specified group.

@item C-t g k
Kill the current group. All windows are merged into the next group.

@item C-t g A
@itemx C-t g r
Change the current group's name.

@item C-t g @var{d}
Go to the group with digit @var{d}. @kbd{C-t g 1} jumps to group 1 and so on.

@item C-t +
Make frames the same height or width in the current frame's subtree.

@item C-t h k
Describe the specified key binding.

@item C-t h f
Describe the specified function.

@item C-t h v
Describe the specified variable.

@item C-t h w
List all key sequences that are bound to the specified command

@item C-t h c
Describe the specified command.
@end table

@node Binding Keys, Modifiers, List of Default Keybindings, Key Bindings
@section Binding Keys

@@@ define-key
@@@ undefine-key
@@@ kbd
!!! set-prefix-key

@@@ make-sparse-keymap

### *root-map*
### *top-map*
### *groups-map*
### *group-top-maps*
### *exchange-window-map*
### *key-seq-color*

!!! bind
!!! unbind
!!! send-escape
@@@ grab-pointer
@@@ ungrab-pointer
### *banish-pointer-to*

@@@ final-key-p
@@@ help-key-p
@@@ cancel-key-p
### *editor-bindings*
### *numpad-map*

@node Modifiers, AltGr Keys, Binding Keys, Key Bindings
@section Modifiers

Many users have had some difficulty with setting up modifiers for
StumpWM keybindings. This is caused by a combination of how StumpWM
handles modifiers and the default modifiers list for many users' X
servers.

@itemize

@item
My ``Super'' key doesn't work!

This is most likely caused by having the Hyper and Super keys listed
as the same modifier in the modifier list.

@example
$ xmodmap
xmodmap:  up to 3 keys per modifier, (keycodes in parentheses):

shift       Shift_L (0x32),  Shift_R (0x3e)
lock        Caps_Lock (0x42)
control     Control_L (0x25),  Control_R (0x6d)
mod1        Alt_L (0x40),  Alt_R (0x71),  Meta_L (0x9c)
mod2        Num_Lock (0x4d)
mod3
mod4        Super_L (0x7f),  Hyper_L (0x80)
mod5        Mode_switch (0x5d),  ISO_Level3_Shift (0x7c)
@end example

The problem is in the line beginning with ``mod4''. The way to set up
the modifier list correctly is to have just the Super key as the mod4
modifier. The following @command{xmodmap} commands will do just that.

@example
# clear out the mod4 modifier
$ xmodmap -e 'clear mod4'
$ xmodmap
xmodmap:  up to 3 keys per modifier, (keycodes in parentheses):

shift       Shift_L (0x32),  Shift_R (0x3e)
lock        Caps_Lock (0x42)
control     Control_L (0x25),  Control_R (0x6d)
mod1        Alt_L (0x40),  Alt_R (0x71),  Meta_L (0x9c)
mod2        Num_Lock (0x4d)
mod3
mod4
mod5        Mode_switch (0x5d),  ISO_Level3_Shift (0x7c)

# add Super as a mod4 modifier
$ xmodmap -e 'add mod4 = Super_L'
$ xmodmap
xmodmap:  up to 3 keys per modifier, (keycodes in parentheses):

shift       Shift_L (0x32),  Shift_R (0x3e)
lock        Caps_Lock (0x42)
control     Control_L (0x25),  Control_R (0x6d)
mod1        Alt_L (0x40),  Alt_R (0x71),  Meta_L (0x9c)
mod2        Num_Lock (0x4d)
mod3
mod4        Super_L (0x73),  Super_L (0x7f)
mod5        Mode_switch (0x5d),  ISO_Level3_Shift (0x7c)
@end example

You can automate this by storing the commands in a file and calling
xmodmap when you start your X session.

@example
$ cat ~/.Xmodmap
clear mod4
add mod4 = Super_L
@end example

If you use @command{startx}, modify your @file{~/.xsession} or @file{~/.xinitrc} file.

@example
$ cat ~/.xsession
#!/bin/sh

xmodmap ~/.Xmodmap
exec /usr/bin/stumpwm
@end example

If you use a settings daemon from one of the major desktop
environments (Gnome,KDE, or Unity) you may be able to set keyboard
modifiers from their respective configuration GUIs.  If not,
@command{xmodmap} should always work if invoked at the right place.

@item
Handling Meta and Alt: when do I use @kbd{M-} and @kbd{A-}?

If you have no Meta keys defined (see the output of the @command{xmodmap}
command), then StumpWM will treat the @kbd{M-} prefix in keybindings to
mean Alt. However, if there are Meta keys defined, then the @kbd{M-}
prefix refers to them, and the @kbd{A-} prefix refers to Alt.

Most users will simply use @kbd{M-} to refer to their Alt keys. However,
users that define separate Meta and Alt keys will use @kbd{M-} to refer
to the former, and @kbd{A-} to refer to the latter.

@item
How can I set up a Hyper key and use it with StumpWM?

To set up a Hyper key, you need to do two things: bind a physical key
to be a Hyper key, and add that key to the modifiers list.

The following example shows how to bind the control key at the
bottom-left of most keyboards to be Hyper. This is useful if you've
made Caps Lock into a control key, and have no use for the bottom-left
key.

@example
$ xmodmap -e 'keycode 37 = Hyper_L'
$ xmodmap -e 'clear mod5'
$ xmodmap -e 'add mod5 = Hyper_L'
@end example

To use a different key for Hyper, replace the keycode ``37'' above.
Use the @command{xev} program to see the keycode that any physical key has.
Refer to the section above on setting up the Super key to see how to
automate setting the Hyper key when you start X.

Now you can use @kbd{H-} as a prefix in StumpWM bindings.

@example
(define-key *top-map* (kbd "H-RET") "fullscreen")
(define-key *top-map* (kbd "H-Left") "gprev")
(define-key *top-map* (kbd "H-Right") "gnext")
(define-key *top-map* (kbd "H-TAB") "other")
@end example

Since essentially no programs have Hyper bindings, you can safely bind
commands to the @var{*top-map*}.

@end itemize
### *all-modifiers*
### *modifiers*

@node AltGr Keys, Remapped Keys, Modifiers, Key Bindings
@section AltGr Keys

StumpWM uses the CLX client library for X11, which doesnt support the
XKB extension. As such support for Alt Graph must be handled
specially.

Preliminary support for Alt Graph in StumpWM is done by registering
AltGr as a modifier key and specially handling it when it is a member of
a key press' state. By default StumpWM treats AltGr key presses as a
normal key and not a modifier, and this behavior should not be changed
unless necessary. To enable treating AltGr as a modifier, the user must
place the following in their stumpwmrc:

@example
(register-altgr-as-modifier)
@end example

Additionally, the variable @var{*altgr-offset*} defaults to 2, which may
be appropriate for the users keyboard layout, or may not. Some users may
have an AltGr offset of 4, or potentially 6. If, after calling
@code{register-altgr-as-modifier}, keys on AltGr aren't being registered
properly, then the offset may need to be changed. The user can check
this explicitly by finding a keysym that requires AltGr and passing it
to @code{keysym-requires-altgr}. As an example, some layouts have @kbd{ł} on
@kbd{AltGr+l}. In such a case, the following example would return true
when @var{*altgr-offset*} is correct:

@example
(stumpwm::keysym-requires-altgr
  (stumpwm::stumpwm-name->keysym "lstroke"))
@end example

@@@ register-altgr-as-modifier
### *altgr-offset*

@node Remapped Keys, , AltGr Keys, Key Bindings
@section Remapped Keys

StumpWM may be configured to translate certain familiar top level
keybindings to alternative key sequences that are understood by
specific applications. For example, Emacs users are very familiar with
@kbd{C-n} and @kbd{C-p} as keybindings for scrolling down and up one
line at a time. However, most applications use these specific
keybindings for other actions. The @code{stumpwm:define-remapped-keys}
function may be used to define such application specific remapping of
keybindings.

@@@ define-remapped-keys

@example
(define-remapped-keys
    '(("(Firefox|Chrome)"
       ("C-n"   . "Down")
       ("C-p"   . "Up")
       ("C-f"   . "Right")
       ("C-b"   . "Left")
       ("C-v"   . "Next")
       ("M-v"   . "Prior")
       ("M-w"   . "C-c")
       ("C-w"   . "C-x")
       ("C-y"   . "C-v")
       ("M-<"   . "Home")
       ("M->"   . "End")
       ("C-M-b" . "M-Left")
       ("C-M-f" . "M-Right")
       ("M-f"   . "C-Right")
       ("M-b"   . "C-Left")
       ("C-k"   . ("C-S-End" "C-x")))))
@end example

The above form adds Emacs like keybindings to windows whose @var{window-class} matches ``Firefox'' or ``Chrome''. Additional application specific bindings may be included by using the specific X @var{window-class} values.

The window matching pattern can also be specified as a function which returns @code{T} if the focused window matches.

@example
;; Match any window with a window-class matching "Firefox"
(define-remapped-keys
    `((,(lambda (win)
          (string-equal "Firefox" (window-class win)))
       ("C-n"   . "Down")
       ("C-p"   . "Up")
       ("C-f"   . "Right")
       ("C-b"   . "Left")
       ("C-v"   . "Next")
       ("M-v"   . "Prior")
       ("M-w"   . "C-c")
       ("C-w"   . "C-x")
       ("C-y"   . "C-v")
       ("M-<"   . "Home")
       ("M->"   . "End")
       ("C-M-b" . "M-Left")
       ("C-M-f" . "M-Right")
       ("M-f"   . "C-Right")
       ("M-b"   . "C-Left")
       ("C-k"   . ("C-S-End" "C-x")))))
@end example

@subsection Circumventing Remapped Keys

However, if the original key binding needs to be explictly applied the
@command{send-raw-key} command may be used. It will prompt for a key
which will be passed to the application as-is. For example, if the
@command{send-raw-key} command were bound to @kbd{C-t C-q} as follows:

@example
(define-key *root-map* (kbd "C-q") "send-raw-key")
@end example

Then, pressing @kbd{C-t C-q}, while the Firefox window has focus,
would prompt asking for ``Press a key to send''. Pressing @kbd{C-n} at
the prompt will send the keystroke as-is to Firefox, causing it to
open a new window.

If more than a single key needs to be passed to the application as-is,
the variable @var{*remapped-keys-enabled-p*} may be used.
Set to nil it will disable all remapped keys.

!!! send-raw-key

### *remapped-keys-enabled-p*

@node Commands, Message and Input Bar, Key Bindings, Top
@chapter Commands
If you've used Emacs before you'll find the distinction between
commands and functions familiar. Commands are simply functions that
can be bound to keys and executed interactively from StumpWM's input
bar. Whereas, in Emacs, the special "(interactive)" declaration is
used to turn a function into a command, in StumpWM commands are made
with a separate @command{defcommand} or
@command{define-interactive-keymap} macro.

Once a command is defined, you can call it by invoking the
@command{colon} command (@kbd{C-t ;}), and typing the name
of the command. This may be sufficient for commands that aren't used
very often. To see all the currently-defined commands, invoke the
command called @command{commands}: ie press @kbd{C-t ;},
type ``commands'', and hit return.

Commonly-used commands can also be bound to a keystroke, which is much
more convenient. To do this, use the @command{define-key}
function (see @ref{Key Bindings}), giving the name of the command as a
string. For example:

@example
(define-key *root-map* (kbd "d") "exchange-direction")
@end example

You cannot give the command name as a symbol, nor can you bind a key
to a regular function defined with @command{defun}.

If the command takes arguments (see @ref{Writing Commands}), you can fix
those arguments when defining the key-binding, by including the
arguments in the same string as the command name, separated by a
space. For instance, the @command{exchange-direction} command,
which is unbound by default, requires a direction in which to exchange
windows. If you call @command{exchange-direction} directly, it
will prompt you for the direction. If you know that you often exchange
in left/right directions, and want those actions bound to keys, you
can use the following in your customization file:

@example
(define-key *root-map* (kbd "[") "exchange-direction left")
(define-key *root-map* (kbd "]") "exchange-direction right")
@end example

Multiple arguments can be included by adding them to the command
string, separated by spaces. Not all argument types can be represented
as strings, but StumpWM will do its best to convert types.

StumpWM does not implement the Emacs concept of prefix arguments.

@menu
* Writing Commands::
* StumpWM Types::
* Common Built-in Commands::
@end menu

@node Writing Commands, StumpWM Types, Commands, Commands
@section Writing Commands
StumpWM commands are written much like any Lisp function. The main
difference is in the way command arguments are specified. The
@command{defcommand} macro takes a list of arguments as its first
form (similar to the @command{defun} macro), and a corresponding
list of types as its second form. All arguments must belong to a
``type''. Each type specification has two parts: a keyword specifying
the argument type, and a string prompt that will be displayed when
asking the user to enter the argument value. A typical
@command{defcommand} might look like this:

@example
(defcommand now-we-are-six (name age)
    ((:string "Enter your name: ")
     (:number "Enter your age: "))
  (message "~a, in six years you will be ~a" name (+ 6 age)))
@end example

If @command{now-we-are-six} is called interactively via the
@command{colon} command, the user will be prompted for a string
and a number, which will then be bound to ``name'' and ``age'',
respectively, in the body of the command.

When invoking the command via a key-binding, it is possible to provide
some or all of the arguments directly:

@example
(define-key *root-map* (kbd "L") "now-we-are-six John")
@end example

In this case, hitting @kbd{C-t L} will only prompt for an age (the first
string argument is already bound to ``John''). Argument values provided
this way always bind to the earliest arguments defined: ie, it is not
possible to specify an age, but prompt the user for a name.

If the type declaration does not include a prompt (ie, it looks like
``(:type nil)'', or ``(:type)'' or just ``:type''), the argument is
considered optional. It can be provided via a key-binding invocation, as
above, but if it isn't, the user will not be prompted, and the argument
will be bound to nil.

It is possible to limit the scope under which the command will
be usable: a command can be defined to work only a specific group type;
the three currently implemented are tile groups,f loating groups, and
dynamic groups. This is done by replacing the name of the command with a
two-element list: the name of the command as a symbol, and either the symbol
tile-group or floating-group. For instance, the @command{next} command,
which only functions in tile groups, is defined this way:

@example
(defcommand (next tile-group) @dots{})
@end example

%%% defcommand
%%% defcommand-alias
### *last-command*

@section Interactive Keymaps

Interactive keymaps are a special type of command that basically
pushes another keymap on top of the current one. The new keymap will
only be removed after an exit command is run. An example is
@command{iresize}.

The macro @command{define-interactive-keymap} is used to define an
interactive keymap. The first argument is the same as
@command{defcommand}. The second argument is a list of extra
configurations that can be used for controlling the command and the
rest are the key bindings for the new command, optionally with a
@code{t} appended; this tells @code{define-interactive-keymap} to
exit the keymap upon use of that keybinding. 

For instance, a simple interactive keymap:

@example
(define-interactive-keymap my-new-command nil
  ((kbd "a") "execute-a-command")
  ((kbd "b") "execute-b-command" t))
@end example

This creates a command called @code{my-new-command} that, when called,
will activate the interactive keymap mode. In this mode, the user can
press ``a'' or ``b'' repeatedly, omitting any prefix. The default exit
commands are @code{RET}, @code{C-g} and @code{ESC}.

This creates a command called @code{my-new-command} that, when called,
will activate the interactive keymap mode. In this mode, the user can
press ``a'' or ``b'', omitting any prefix. The user can press ``a''
repeatedly, however pressing ``b'' exits the keymap. The default exit
commands are @code{RET}, @code{C-g} and @code{ESC}.

The available configuration is @code{on-enter}, @code{on-exit} and
@code{abort-if}:

@example
(defun before-foo () (message "start foo"))
(defun after-foo () (message "end foo"))
(defun foo-p () (and *bar* *baz*))
(defparameter *custom-exit-keys* '((kbd "RET") (kbd "SPC")
                                   (kbd "C-g") (kbd "ESC")))

(define-interactive-keymap foo (:on-enter #'before-foo
                                :on-exit #'after-foo
                                :abort-if #'foo-p
                                :exit-on *custom-exit-keys*))
@end example

In the above example, the message ``start foo'' will appear before
starting the interactive keymap, ``end foo'' will appear right after
the command exits; We've added SPC as an exit key with custom exit 
keys. Also, the command executes only if the variables @code{*bar*}
and @code{*baz*} are true.

%%% define-interactive-keymap
!!! call-and-exit-kmap

@node StumpWM Types, , Writing Commands, Commands
@section StumpWM Types
All command arguments must be of a defined ``StumpWM type''. The
following types are pre-defined:


@table @var
@item :y-or-n
A yes or no question returning T or NIL.
@item :variable
A lisp variable
@item :function
A lisp function
@item :command
A StumpWM command as a string.
@item :key-seq
A key sequence starting from *TOP-MAP*
@item :window-number
An existing window number
@item :number
An integer number
@item :string
A string
@item :key
A single key chord
@item :window-name
An existing window's name
@item :direction
A direction symbol. One of :UP :DOWN :LEFT :RIGHT
@item :gravity
A gravity symbol. One of :center :top :right :bottom :left :top-right :top-left :bottom-right :bottom-left
@item :group
An existing group
@item :frame
A frame
@item :shell
A shell command
@item :rest
The rest of the input yet to be parsed.
@item :module
An existing StumpWM module
@end table

Additional types can be defined using the macro
@command{define-stumpwm-type}. Emacs users who are accustomed to writing
more complicated interactive declarations using "(interactive (list
@dots{}))" forms will find that similar logic can be put into StumpWM
type definitions. The macro is called like this:

@example
(define-stumpwm-type :type-name (input prompt) body)
@end example

The keyword :type-name will then be available for use in
@command{defcommand} macros. When commands are called, the bodies
of these type definitions are called in turn to produce actual
argument values.

Type definitions produce their value in one of several ways: by
reading it from the argument line bound to a keystroke, by prompting
the user to enter a value, or by generating it programmatically.

Within the body of the type definition, the argument ``input'' is bound
to the argument line provided in the command string, and ``prompt'' to
the string prompt provided in the @command{defcommand} form. The
usual convention is to first check if an argument has been provided in
``input'' and, if it hasn't, to prompt for it using ``prompt''.

StumpWM provides several convenience functions for handling the value
of ``input'':

@itemize
@item
@command{argument-pop} (input) pop the next space-delimited word or a
double quote delimited string argument from the argument
line. Backslashes may be used to escape double quotes or backslashes
inside double quoted strings.

@item
@command{argument-pop-rest} (input) return the remainder of the
argument line as a single string, leaving input empty

@item
@command{argument-pop-or-read} (input prompt &optional
completions) either pop an argument from the argument line, or if it
is empty use ``prompt'' to prompt the user for a value

@item
@command{argument-pop-rest-or-read} (input prompt &optional
completions) either return the remainder of the argument line as a
string, leaving input empty, or use ``prompt'' to prompt the user for
a value

@end itemize

As an example, here's a new type called :smart-direction. The
existing :direction type simply asks for one of the four directions
``left'', ``right'', ``up'' or ``down'', without checking to see if there's a
frame in that direction. Our new type, :smart-direction, will look
around the current frame, and only allow the user to choose a
direction in which another frame lies. If only one direction is
possible it will return that automatically without troubling the user.
It signals an error for invalid directions; it could alternately
return a ``nil'' value in those cases, and let the command handle that.

@example
(define-stumpwm-type :smart-direction (input prompt)
  (let ((valid-dirs
         (loop  ; gather all the directions in which there's a neighbouring frame
            with values = '(("up" :up)
                            ("down" :down)
                            ("left" :left)
                            ("right" :right))
            with frame-set =
              (group-frames (window-group (current-window)))
            for dir in values
            for neighbour = (neighbour
                             (second dir)
                             (window-frame (current-window)) frame-set)
            if (and neighbour (frame-window neighbour))
            collect dir))
        (arg (argument-pop input)))  ; store a possible argument
    (cond ((null valid-dirs)  ; no directions, bail out
           (throw 'error "No valid directions"))
          (arg  ; an arg was bound, but is it valid?
           (or (second (assoc arg valid-dirs :test #'string=))
               (throw 'error "Not a valid direction")))
          ((= 1 (length valid-dirs))  ; only one valid direction
           (second (car valid-dirs)))
          (t  ; multiple possibilities, prompt for direction
           (second (assoc (completing-read input prompt valid-dirs
                                           :require-match t)
                          valid-dirs :test #'string=))))))

(defcommand smarty (dir) ((:smart-direction "Pick a direction: "))
  ;; `dir' is a keyword here
  (message "You're going ~a" (string-downcase dir)))

(define-key *root-map* (kbd "R") "smarty right")
@end example

%%% define-stumpwm-type

@node Common Built-in Commands, , Writing Commands, Commands
@section Common Built-in Commands

!!! emacs
!!! version
!!! banish
!!! ratwarp
!!! ratrelwarp
!!! ratclick
!!! restart-hard
!!! restart-soft
!!! lastmsg
!!! commands
!!! keyboard-quit
!!! quit
!!! reload
!!! echo-date
!!! eval-line
!!! command-mode
!!! list-window-properties
!!! show-window-properties
!!! time

@node Message and Input Bar, Windows, Commands, Top
@chapter Message and Input Bar

### *suppress-echo-timeout*
!!! echo
@@@ err
@@@ wrap
### *message-max-width*
### *help-max-height*
### *which-key-format*
!!! colon

%%% with-restarts-menu
@@@ restarts-menu

@menu
* Customizing The Bar::
* Using The Input Bar::
* Programming The Message Bar::
* Programming the Input Bar::
@end menu

@node Customizing The Bar, Using The Input Bar, Message and Input Bar, Message and Input Bar
@section Customizing The Bar
The bar's appearance and behavior can be modified with the following
functions and variables. See @ref{Colors} for an explanation of how to
set these color variables.

@@@ set-fg-color
@@@ set-bg-color
@@@ set-border-color
@@@ set-msg-border-width
@@@ set-font

### *message-window-padding*
### *message-window-y-padding*
### *message-window-gravity*
### *message-window-input-gravity*
### *message-window-timer*
### *timeout-wait*
### *timeout-wait-multiline*
### *input-window-gravity*

@node Using The Input Bar, Programming The Message Bar, Customizing The Bar, Message and Input Bar
@section Using The Input Bar

The following is a list of keybindings for the Input Bar. Users of Emacs
will recognize them.

@table @kbd
@item DEL
Delete the character before point (@code{delete-backward-char}).

@item M-DEL
Kill back to the beginning of the previous word (@code{backward-kill-word}).

@item C-d
@itemx Delete
Delete the character after point (@code{delete-forward-char}).

@item M-d
Kill forward to the end of the next word (@code{forward-kill-word}).

@item C-f
@itemx Right
Move forward one character (@code{forward-char}).

@item M-f
Move forward one word (@code{forward-word}).

@item C-b
@itemx Left
Move backward one character (@code{backward-char}).

@item M-b
Move backward one word (@code{backward-word}).

@item C-a
@itemx Home
Move to the beginning of the current line (@code{move-beginning-of-line}).

@item C-e
@itemx End
Move to the end of the current line (@code{move-end-of-line}).

@item C-k
Kill to the end of the line (@code{kill-line}).

@item C-u
Kill to the beginning of the line (@code{kill-to-beginning}),
the same as @kbd{C-a C-k}.

@item C-p
@itemx Up
Move to the next earlier entry saved in the command history (@command{history-back}).

@item C-n
@itemx Down
Move to the next later entry saved in the command history (@command{history-forward}).

@item RET
Submit the entered command (@command{submit}).

@item C-g
Abort the current action by closing the Input Bar (@command{abort}).

@item C-y
Paste text from clipboard into the Input Bar (@command{yank-selection}).

@item TAB
Clockwise tab complete the current string, if possible. Press @kbd{TAB} again to
cycle through completions.

@item S-TAB
Counter-clockwise tab complete the current string, if possible. Press @kbd{S-TAB} again to
cycle through completions.

@end table

@node Programming The Message Bar, Programming the Input Bar, Using The Input Bar, Message and Input Bar
@section Programming The Message Bar

@@@ echo-string
@@@ message
%%% with-message-queuing
### *queue-messages-p*
### *input-history-ignore-duplicates*
### *input-completion-style*
@@@ make-input-completion-style-cyclic
@@@ make-input-completion-style-unambiguous
!!! copy-last-message

@node Programming the Input Bar,  , Programming The Message Bar, Message and Input Bar
@section Programming the Input Bar
New input behavior can be added to the input bar by creating editing
functions and binding them to keys in the @var{*input-map*} using
@command{define-key}, just like other key bindings.

An input function takes 2 arguments: the input structure and the key pressed.

@@@ read-one-line
@@@ read-one-char
@@@ completing-read
@@@ input-insert-string
@@@ input-insert-char

### *input-map*

@node Windows, Frames, Message and Input Bar, Top
@chapter Windows

!!! next
!!! prev
!!! delete-window
!!! kill-window
!!! kill-windows-current-group
!!! kill-windows-other
!!! echo-windows
!!! other-window
!!! pull-hidden-next
!!! pull-hidden-previous
!!! pull-hidden-other
!!! pull-from-windowlist
!!! renumber
!!! meta
!!! select-window
!!! select-window-by-number
!!! select-window-by-name
!!! repack-window-numbers
!!! title
!!! windowlist
!!! windowlist-by-class
!!! fullscreen
!!! info
!!! refresh
!!! redisplay
!!! float-this
!!! unfloat-this
!!! flatten-floats
!!! unmaximize
!!! toggle-always-on-top
!!! toggle-always-show
!!! window-send-string

@@@ window-head
@@@ window-sync
@@@ window-visible-p
@@@ raise-window
@@@ focus-window


### *xwin-to-window*
### *window-format*
### *window-info-format*
### *window-name-source*
### *new-window-preferred-frame*
### *hidden-window-color*

### *honor-window-moves*


%%% define-fullscreen-in-frame-rule
@@@ add-fullscreen-in-frame-rule
@@@ remove-fullscreen-in-frame-rule
### *fullscreen-in-frame-p-window-functions*


@menu
* Window Marks::
* Customizing Window Appearance::
* Controlling Raise And Map Requests::
* Programming With Windows::
* Rule Based Window Placement::
* Window Selection Expressions::
@end menu

@node Window Marks, Customizing Window Appearance, Windows, Windows
@section Window Marks
Windows can be marked. A marked window has a # beside it in the window
list. Some commands operate only on marked windows.

!!! mark
!!! clear-window-marks
!!! pull-marked

@node Customizing Window Appearance, Controlling Raise And Map Requests, Window Marks, Windows
@section Customizing Window Appearance

### *maxsize-border-width*
### *transient-border-width*
### *normal-border-width*
### *window-border-style*

See @ref{Colors} for an explanation of how to set these color
variables.

@@@ set-win-bg-color
@@@ set-focus-color
@@@ set-unfocus-color
@@@ set-float-focus-color
@@@ set-float-unfocus-color
@@@ set-normal-gravity
@@@ set-maxsize-gravity
@@@ set-transient-gravity

!!! gravity
@@@ gravity-coords

@node Controlling Raise And Map Requests, Programming With Windows, Customizing Window Appearance, Windows
@section Controlling Raise And Map Requests
It is sometimes handy to deny a window's request to be focused. The
following variables determine such behavior.

A map request occurs when a new or withdrawn window requests to be
mapped for the first time.

A raise request occurs when a client asks the window manager to give
an existing window focus.

### *deny-map-request*
### *deny-raise-request*
### *suppress-deny-messages*

Some examples follow.

@example
;; Deny the firefox window from taking focus when clicked upon.
(push '(:class "gecko") stumpwm:*deny-raise-request*)

;; Deny all map requests
(setf stumpwm:*deny-map-request* t)

;; Deny transient raise requests
(push '(:transient) stumpwm:*deny-map-request*)

;; Deny the all windows in the xterm class from taking focus.
(push '(:class "Xterm") stumpwm:*deny-raise-request*)
@end example

@node Programming With Windows, Rule Based Window Placement, Controlling Raise And Map Requests, Windows
@section Programming With Windows

%%% define-window-slot

@@@ window-send-string
### *default-window-name*
### *window-events*
### *window-parent-events*

@node Rule Based Window Placement, Window Selection Expressions, Programming With Windows, Windows
@section Rule Based Window Placement

%%% define-frame-preference
@@@ clear-window-placement-rules

!!! remember
!!! forget
!!! dump-window-placement-rules
!!! restore-window-placement-rules
### *window-placement-rules*
@node Window Selection Expressions,  , Rule Based Window Placement, Windows
@section Window Selection Expressions
Window Selection Expressions (WSE) were inspired by SQL. The intent is
to allow writing consise code to select the windows you need and to
act upon them (or just to get the list of selected windows). The
implementation includes a set of (hopefully) consistent
concisely-named wrappers for the StumpWM functionality useful for
window set description and the act-on-matching-windows macro that
encapsulates the logic of iterating over a window set.

If we had SQL in StumpWM, we would write
@code{select window_id from windows as w where w.title = 'XTerm'}. WSE chooses to be
more Lisp-style and instead uses
@code{(act-on-matching-windows (w) (titled-p w "XTerm") w)}

The @code{act-on-matching-windows} function also allows performing some
actions, for example getting all the windows titled XTerm into the
current group:
@code{(act-on-matching-windows (w) (titled-p w "XTerm") (pull-w w))}

@@@ move-windows-to-group
%%% act-on-matching-windows
@@@ pull-w
@@@ titled-p
@@@ title-re-p
@@@ classed-p
@@@ class-re-p
@@@ typed-p
@@@ type-re-p
@@@ roled-p
@@@ role-re-p
@@@ resed-p
@@@ res-re-p
@@@ grouped-p
@@@ in-frame-p

@node Frames, Mode-line, Windows, Top
@chapter Frames
Frames contain windows. All windows exist within a frame.

Those used to ratpoison will notice that this differs from ratpoison's
window pool, where windows and frames are not so tightly connected.

!!! pull-window-by-number
!!! hsplit
!!! vsplit
!!! hsplit-equally
!!! vsplit-uniformly
!!! vsplit-equally
!!! hsplit-uniformly
!!! remove-split
!!! only
!!! curframe
!!! fnext
!!! fprev
!!! sibling
!!! fother
!!! fselect
!!! resize
!!! resize-direction
!!! balance-frames
!!! fclear
!!! move-focus
!!! move-window
!!! next-in-frame
!!! prev-in-frame
!!! other-in-frame
!!! next-urgent
!!! frame-windowlist
!!! echo-frame-windows
!!! exchange-direction
!!! expose

%%% save-frame-excursion

@@@ run-or-pull
@@@ only-one-frame-p

### *min-frame-width*
### *min-frame-height*
### *new-frame-action*
### *expose-auto-tile-fn*
### *expose-n-max*
### *frame-indicator-text*
### *frame-indicator-timer*
### *frame-number-map*

@menu
* Interactively Resizing Frames::
* Frame Dumping::
@end menu

@node Interactively Resizing Frames, Frame Dumping, Frames, Frames
@section Interactively Resizing Frames
There is a mode called @code{iresize} that lets you interactively
resize the current frame. To enter the mode use the @code{iresize}
command or type @key{C-t r}.

The following keybindings apply to the mode:
@table @kbd
@item C-p
@itemx Up
@itemx k
Shrink the frame vertically.

@item C-n
@itemx Down
@itemx j
Expand the frame vertically.

@item C-f
@itemx Right
@itemx l
Expand the frame horizontally.

@item C-b
@itemx Left
@itemx h
Shrink the frame horizontally.

@item C-g
@itemx ESC
Abort the interactive resize.

@item RET
Select the highlighted option.

@end table

!!! iresize
@@@ setup-iresize
### *resize-map*

### *resize-increment*

@node Frame Dumping,  , Interactively Resizing Frames, Frames
@section Frame Dumping
The configuration of frames and groups can be saved and restored using
the following commands.

!!! dump-desktop-to-file
!!! dump-group-to-file
!!! dump-screen-to-file
!!! restore-from-file
!!! place-existing-windows
!!! place-current-window

The configuration files are stored in the
@var{$XDG_CONFIG_HOME/stumpwm}. The file name specified is saved as a
@code{.dump} file type. For example, @code{: dump-desktop-to-file
example} may save a file in
@code{~/.local/share/stumpwm/example.dump}. @code{restore-from-file}
also adds the @code{.dump} extension by default.

@node Mode-line, Groups, Frames, Top
@chapter The Mode Line
The mode line is a bar that runs across either the top or bottom of
a head and is used to display information. By default the mode line
displays the list of windows, similar to the output @kbd{C-t w} produces.

Alternatively, external panel applications such as the GNOME panel and
KDE's kicker may be used. Simply starting one of these programs is enough
to set it as the mode line of the head it would like to be on (if the
panel is XRandR aware) or whichever head is available. In order to
avoid problems displaying menus, configure your panel application for
positioning at the top or bottom of the head rather than relying on
@var{*mode-line-position*}

The mode line can be turned on and off with the @command{mode-line}
command or the lisp function @code{stumpwm:toggle-mode-line}. Each
head has its own mode line. For example:

@example
;; turn on/off the mode line for the current head only.
(stumpwm:toggle-mode-line (stumpwm:current-screen)
                          (stumpwm:current-head))
@end example

The mode line is updated after every StumpWM command.

To display the window list and the current date on the modeline, one
might do the following:

@example
(setf stumpwm:*screen-mode-line-format*
      (list "%w | "
            '(:eval (stumpwm:run-shell-command "date" t))))
@end example

@code{(stumpwm:run-shell-command "date" t)} runs the command
@command{date} and returns its output as a string.

!!! mode-line

@@@ toggle-mode-line

### *screen-mode-line-format*
### *time-format-string-default*
### *time-modeline-string*
$$$ *new-mode-line-hook*
### *screen-mode-line-formatters*
### *window-formatters*
@@@ bar
@@@ bar-zone-color
@@@ add-screen-mode-line-formatter
@@@ enable-mode-line

The following variables control the color, position, and size of the
mode line. See @ref{Colors} for an explanation of how to set these
color variables.

### *mode-line-position*
### *mode-line-border-width*
### *mode-line-highlight-template*
### *mode-line-pad-x*
### *mode-line-pad-y*
### *mode-line-background-color*
### *mode-line-foreground-color*
### *mode-line-border-color*
### *mode-line-timeout*

@menu
* Mode-line Interaction::
@end menu

@node Mode-line Interaction, , Mode-line, Mode-line
@section Mode-line Interaction
Mode line formatters can register sections of text to be clickable by
use of the color formatters @code{:on-click} and
@code{:on-click-end}. Any text enclosed by these formatters has its
bounds saved, and when the mode line recieves a button press event these
bounds are checked against to find a clickable area, whose registered
function is then called. These formatters can be thought of as similar
to XML tags.

To disable the on-click behavior, remove the function
@code{check-for-ml-press} from the hook @var{*mode-line-click-hook*}.

To call a function by click the function must first be registered. The
function must take at least one argument, the button code. Here is an
example of a click-to-focus function and its registration:

@example
(defun ml-on-click-focus-window (code id &rest rest)
  (declare (ignore code rest))
  (when-let ((window (window-by-id id)))
    (focus-all window)))

(register-ml-on-click-id :ml-on-click-focus-window #'ml-on-click-focus-window)
@end example

This defines a function that focuses a window based upon its X11 window
ID, and registers it under the ID @code{:ml-on-click-focus-window}. Here
is an example of a mode line formatter that makes use of this function:

@example
(add-screen-mode-line-formatter #\i 'fmt-head-window-list-clickable)
(defun fmt-head-window-list-clickable (ml)
  "Using *window-format*, return a 1 line list of the windows,
space seperated and clickable."
  (flet ((fmt-w (w)
           (let ((str (format-expand *window-formatters*
                                     *window-format*
                                     w)))
             (format-with-on-click-id (if (eq w (current-window))
                                          (fmt-highlight str)
                                          str)
                                      :ml-on-click-focus-window
                                      (window-id w)))))
@verbatim
    (format nil "~{~a~^ ~}"
@end verbatim
            (mapcar #'fmt-w
                    (sort1 (head-windows (mode-line-current-group ml)
                                         (mode-line-head ml))
                           #'< :key #'window-number)))))
@end example

In the above formatter, every windows expansion is wrapped in a
:on-click/end pair, which takes the ID we registered as the function to
call and the window ID as an argument to be passed to its function. The
arguments provided to @code{:on-click} will be read but not
evaluated. The string generated will look like so:

@example
"^(:on-click :ml-on-click-focus-window 308242)window text^(:on-click-end)"
@end example

Clickable text can be nested, in which case the innermost clickable text
will take precendent. In the following example @code{:id2} will be
dispatched when clicking @code{2}, but @code{:id1} will be dispatched
when clicking @code{1} and @code{3}:

@example
"^(:on-click :id1)1^(:on-click :id2)2^(:on-click-end)3^(:on-click-end)"
@end example

If one wished for right click to delete windows, then the following
example could be placed in the .stumpwmrc:

@example
(labels ((ml-on-click-focus-or-delete-window (code id &rest rest)
           (declare (ignore rest))
           (when-let ((window (window-by-id id)))
             (let ((button (decode-button-code code)))
               (case button
                 ((:left-button)
                  (focus-all window))
                 ((:right-button)
                  (delete-window window)))))))
  (register-ml-on-click-id :ml-on-click-focus-window
                           #'ml-on-click-focus-or-delete-window))
@end example

This will replace the @code{:ml-on-click-focus-window} function, and all
uses of @code{:on-click} formatters referring to
@code{:ml-on-click-focus-window} will use the new function. 

@@@ register-ml-on-click-id
@@@ format-with-on-click-id

@node Groups, Screens, Mode-line, Top
@chapter Groups
Groups in StumpWM are more commonly known as @dfn{virtual desktops} or
@dfn{workspaces}. Why not create a new term for it?

!!! gnew
!!! gnew-float
!!! gnew-dynamic
!!! gnewbg
!!! gnewbg-float
!!! gnewbg-dynamic
!!! gnext
!!! gprev
!!! gnext-with-window
!!! gprev-with-window
!!! gother
!!! gmerge
!!! groups
!!! vgroups
!!! gselect
!!! gmove
!!! gmove-and-follow
!!! gmove-marked
!!! gkill
!!! gkill-other
!!! grename
!!! grouplist

### *list-hidden-groups* 
### *group-top-maps* 
### *default-group-name* 

@@@ add-group
@@@ group-add-head
@@@ group-add-window
@@@ group-button-press
@@@ group-current-head
@@@ group-current-window
@@@ group-delete-window
@@@ group-focus-window
@@@ group-indicate-focus
@@@ group-lost-focus
@@@ group-move-request
@@@ group-raise-request
@@@ group-remove-head
@@@ group-resize-head
@@@ group-resize-request
@@@ group-root-exposure
@@@ group-startup
@@@ group-suspend
@@@ group-sync-all-heads
@@@ group-sync-head
@@@ group-wake-up

@@@ really-raise-window
### *run-or-raise-all-groups*

@menu
* Customizing Groups::
@end menu

@node Customizing Groups,  , Groups, Groups
@section Customizing Groups

### *group-formatters*
### *group-format*

@@@ current-group

@node Screens, Minor Modes, Groups, Top
@chapter Screens
StumpWM handles multiple screens.

!!! snext
!!! sprev
!!! sother


### *run-or-raise-all-screens*

@menu
* External Monitors::
* Programming With Screens::
@end menu

@node External Monitors, Programming With Screens, Screens, Screens
@section External Monitors
StumpWM refers to each monitor as a head. Heads are logically
contained by screens. In a dual-monitor configuration, there will be
one screen with two heads. Non-rectangular layouts are supported
(frames will not be created in the 'dead zone'.) And message windows
will be displayed on the current head--that is, the head to which the
currently focused frame belongs.

In addition, StumpWM listens for XRandR events and re-configures the
heads to match the new monitor configuration. Occasionally StumpWM
will miss an XRandR event, use @code{refresh-heads} to synchronize
the head configuration.

!!! refresh-heads

@node Programming With Screens,  , External Monitors, Screens
@section Programming With Screens

@@@ current-screen

@@@ screen-current-window

@@@ current-window

### *screen-list*

@node Minor Modes, Internals, Screens, Top
@chapter Minor Modes

Like Emacs, StumpWM has the concept of minor modes. These are defined by
the macro @code{DEFINE-MINOR-MODE}. Defining a minor mode creates a
class and a set of methods specializing upon it. Minor modes are scoped
to a window, head, group, or screen, or they may be unscoped. In
addition to this minor modes may be local or global. When a minor mode
is global all new instances of the scope object will be created with the
minor mode already active in them. Minor modes define their own top
level and root level keymaps, as well as hooks that are run upon
enabling or disabling the minor mode, and a lighter to display in the
mode line.

Minor modes are mixins that get added to the appropriate scope object
when enabled. As such minor modes allow the augmenting, modifying, and
overriding of default StumpWM behavior by defining methods for the
generic functions of the scope object. For example, a minor mode may be
scoped to a window and define a method for the generic function
@code{UPDATE-DECORATION} to change how window decoration is handled for
the windows it is enabled in.

%%% define-minor-mode

@menu
* Programming With Minor Modes::
* Minor Mode Scopes::
* Minor Mode Example::
@end menu

@node Programming With Minor Modes, Minor Mode Scopes, Minor Modes, Minor Modes
@section Programming With Minor Modes

Minor modes get their power from their ability to override and augment
generic functions which are called with the minor mode's scope object as
an argument. If you find a function whose behavior you wish to augment
or override in the process of writing a minor mode, open an issue or
submit a PR to generify the function. Generification is easily done like
so:

@example
(defun somefun (a b)
  "docstring"
  (otherfun (+ a b)))
;; the above becomes
(defgeneric somefun (a b)
  (:documentation "docstring")
  (:method (a b)
    (otherfun (+ a b))))
@end example

When defining a minor mode, the programmer may desire to perform setup
for the minor mode. This is can be done in three ways. The first is to
hang a function upon the minor modes enable hook. However this runs the
risk of users potentially clobbering the initialization function, or
modifying the hook such that the initialization function is not the
first function run. The second way is to define a method for the generic
function @code{update-instance-for-different-class}. This function
should specialize upon the minor mode as the second argument. As a final
option, one can define before, after, and around methods for the generic
function @code{autoenable-minor-mode} if and only if the method does not
access any slots within the object.

One of the pitfalls of minor modes is that they are ultimately enabled
by calling @code{change-class}, which places some restrictions upon
where they can be enabled. Specifically, it is implied to be undefined
behavior if a minor mode is enabled in an object from within a method
which accesses slots of that object. While in practice this has not
proven to be an issue at the time of writing, this is undefined behavior
and future versions of SBCL may break if this is done.

When writing a minor mode, it is often useful to separate out the
desired functionality into its own mixin classes and use those in the
superclass list of the minor mode. For example:

@example
(define-minor-mode my-mode (my-mixin minor-mode) ())
@end example

This prevents issues with inheritance and dynamic mixins from cropping
up. Since minor modes are just classes, a minor mode can descend from
another minor mode. However after enabling the subclass minor mode, the
superclass minor mode cannot be enabled. However if the superclass minor
mode is enabled first, then the subclass minor mode can be enabled. The
easiest way around this is the aforementioned approach of mixins.
As an example of the inheritance issue, take the following minor mode
definitions:

@example
(define-minor-mode x () ())
(define-minor-mode y (x) ())

(enable-minor-mode 'x)
(enable-minor-mode 'y) ; both modes are enabled
;; As opposed to
(enable-minor-mode 'y)
(enable-minor-mode 'x) ; signals an error
@end example


When enabling and disabling minor modes theres a set of generic
functions in charge of determining what object to mix the minor mode in
to and whether or not to mix it.

@@@ autoenable-minor-mode

Defining a minor mode defines a main method for this generic function
which will mix the minor mode into the scope object when called and
returns T. This method specializes upon the minor mode symbol and the
scope type. The minor mode will only be enabled and the hooks run when
the function @code{ENABLE-WHEN} returns T. Any before after or around
methods for this function must not access any slots. 

@@@ autodisable-minor-mode

Defining a minor mode defines a main method for this generic function
which specializes upon the mode symbol and the mode, and removes the
minor mode from the object. Any methods for this function must not
access any slots.

@@@ enable-when

Outside of autoenabling and autodisabling minor modes, there are several
generic functions which dispatch upon minor modes and their names.

@@@ minor-mode-global-p
@@@ minor-mode-scope
@@@ minor-mode-enable-hook
@@@ minor-mode-disable-hook
@@@ minor-mode-hook

@@@ minor-mode-keymap

This function has a set of main methods defined which all call the next
method to obtain a list of top maps for every minor mode. Any extra
keymaps one wishes to add to the minor mode may be added by defining a
main method which calls @code{call-next-method} and returns a flat
list. Similarly, an around method may be used which abides by the same
rules.

@@@ minor-mode-lighter

This function operates similarly to @code{minor-mode-keymap}, with a
main method defined for every minor mode which calls
@code{call-next-method} and returns a flat list. In addition there is a
single around method defined which concatenates all these strings
together.

There are also a set of regular functions and special variables which
which may be of use when working with minor modes.

@@@ enable-minor-mode
@@@ disable-minor-mode
@@@ list-modes
@@@ list-minor-modes
@@@ current-minor-modes
@@@ minor-mode-enabled-p
@@@ find-minor-mode

### *minor-mode*
### *minor-mode-enable-hook*
### *minor-mode-disable-hook*
### *unscoped-minor-modes*

€€€ minor-mode

@node Minor Mode Scopes, Minor Mode Example, Programming With Minor Modes, Minor Modes
@section Minor Mode Scopes

Minor modes can be scoped to different objects in a rather arbitrary
manner. These scopes are defined by the macro
@code{DEFINE-MINOR-MODE-SCOPE}. Because minor modes are implemented as
mixins, the object returned by a scopes current object function must be
a class instance. 

%%% define-minor-mode-scope

%%% define-descended-minor-mode-scope

@@@ add-minor-mode-scope

When a minor mode is defined its scope is looked up and validated by the
function @code{VALIDATE-SCOPE}. This function takes a scope and a list
of superclasses, and ensures that the scope can descend from the
superclasses scopes. This restricts the valid scopes to ensure that a
minor mode scoped to @code{:GROUP} cant be a subclass of a minor mode
scoped to @code{:WINDOW}, for example. However there is a way to
override this by explicitly stating that two otherwise incompatible
scopes are compatible. This is done by defining methods for the generic
function @code{VALIDATE-SUPERSCOPE} which dispatch upon the scope
designators. Such methods should return at least one value, indicating
if the superscope is a valid parent of the scope. If multiple values are
returned, the second value must indicate whether the superscope is an
invalid parent of the scope. For example:

@verbatim
(defmethod stumpwm:validate-superscope ((c (eql :id1)) (p (eql :id2)))
  "Explicitly allow id1 to descend from id2"
  (values t nil))

(defmethod stumpwm:validate-superscope ((c (eql :id3)) (p (eql :id4)))
  "Explicitly prevent id3 from descending from id4"
  (values nil t))
@end verbatim

When defining and using scopes the type specifier is important; it is
used to determine what minor modes should be mixed into an object when
it is created. For this reason it is important when defining a minor
mode or minor mode scope to understand the type hierarchy. It may also
be in the programmers best interests to define an accompanying type.

The following scopes are predefined:
@itemize
@item
Designator: @code{:UNSCOPED}, type: @code{T}
Current object: return the global unscoped object.@*

@item
Designator: @code{:SCREEN}, type: @code{SCREEN}
Current object: return the current screen.@*

@item
Designator: @code{:GROUP}, type: @code{GROUP}
Current object: return the current group.@*

@item
Designator: @code{:TILE-GROUP}, type: @code{TILE-GROUP}
Current object: return the current group.@*

@item
Designator: @code{:DYNAMIC-GROUP}, type: @code{DYNAMIC-GROUP}
Current object: return the current group.@*

@item
Designator: @code{:FLOAT-GROUP}, type: @code{FLOAT-GROUP}
Current object: return the current group.@*

@item
Designator: @code{:TILING-NON-DYNAMIC-GROUP}, type: @code{TILE-GROUP}
Current object: return the current group when it is a non-dynamic tiling
group.@*

@item
Designator: @code{:HEAD}, type: @code{HEAD}
Current object: return the current head.@*

@item
Designator: @code{:FRAME}, type: @code{FRAME}
Current object: return the current frame when in a tiling group.@*

@item
Designator: @code{:FRAME-EXCLUDING-HEAD}, type: @code{ONLY-FRAME-NO-HEADS}
Descends from minor mode scope @code{:FRAME}.

@item
Designator: @code{:WINDOW}, type: @code{WINDOW}

Current object: return the current window.@*
All objects: collect every window from every group in the current screen.

@item
Designator: @code{:TILE-WINDOW}, type: @code{TILE-WINDOW}

Current object: return the current window.@*
All objects: collect every window from every tiling group, filtering all
floating windows.

@item
Designator: @code{:FLOAT-WINDOW}, type: @code{FLOAT-WINDOW}

Current object: return the current window.@*
All objects: colelct every window from every group, filtering all non
floating windows.

@end itemize

@node Minor Mode Example, , Minor Mode Scopes, Minor Modes
@section Minor Mode Example

A simple example of a minor mode is a version of the Emacs modes viper
or evil for StumpWM. Such a minor mode might look like this:

@verbatim
(define-minor-mode swm-evil-mode () ()
  (:scope :screen)
  (:interactive t)
  (:top-map '(("i" . "swm-evil-mode")
              ("j" . "move-focus down")
              ("k" . "move-focus up")
              ("h" . "move-focus left")
              ("l" . "move-focus right")
              ("p" . "pull-hidden-previous")
              ("n" . "pull-hidden-next")
              ("S" . "hsplit")
              ("s" . "vsplit")
              ("r" . "remove-split")
              ("g" . *groups-map*)
              ("x" . *exchange-window-map*)))
  (:lighter-make-clickable nil)
  (:lighter "EVIL"))
@end verbatim

In the above example, the minor mode @code{swm-evil-mode} is defined,
alongside a command of the same name which toggles it on and off. The
minor mode is scoped to a screen, meaning that upon activation it will
be dynamically mixed in to the screen object. The lighter is the string
@code{"EVIL"} and the lighter is not made clickable.

When defining a minor mode top map it is important to avoid multi-key
bindings that clobber the prefix key. For example, if the prefix key is
@kbd{C-t} then defining the keybinding @kbd{C-t n} in the top map of a
minor mode is an error. Instead bind the key @kbd{n} in the minor mode's
root map.

As another example we can define a frame topbar mode. This should adjust
every frame to leave extra space at the top of the frame to display a
bar of some sort. The following assumes that the functions
@code{frame-display-height} and @code{frame-display-y} are generic. 

@example
(defclass frame-topbar ()
  ((frame-topbar-height :initform 10 :accessor frame-topbar-height)))

(defmethod frame-display-height :around (group (frame frame-topbar))
  (let ((height (call-next-method)))
    (- height (frame-topbar-height frame))))

(defmethod frame-display-y :around (group (frame frame-topbar))
  (let ((y (call-next-method)))
    (+ y (frame-topbar-height frame))))

(define-minor-mode frame-bar (frame-topbar minor-mode) ()
  (:global t)
  (:scope :frame)
  (:lighter "T-BAR")
  (:interactive frame-topbar-mode))

(defmethod update-instance-for-different-class :after
    (prev (obj frame-bar) &rest rest)
  (declare (ignore prev rest))
  (when (frame-window obj)
    (let* ((group (window-group (frame-window obj)))
           (windows (frame-windows group obj)))
      (mapc #'maximize-window windows))))
@end example

In the above example, a class is defined which holds the height of the
frame topbar. Then two around methods are defined such that windows
querying the frame for their y position and height get an updated value
reflecting the topbars presence. Then a minor mode is defined which
inherits from the class we defined. It is scoped to frames and is a
global minor mode, so it will be enabled in all existing frames and any
other frames as they are created. Finally the initialization is handled
in the after method for update-instance-for-different-class, which
updates every window to have a new size which respects the topbar.

The implementation of the actual topbar is left as an exercise for the
reader. 

@node Internals, Interacting With Unix, Minor Modes, Top
@chapter Internals
@menu
* IO Loop::
* Internal Functions Documentation::
@end menu

@node IO Loop, Internal Functions Documentation, , Internals
@section IO Loop

StumpWM's internal loop is implemented by a generic multiplexing I/O
loop for listening to I/O events from multiple sources. The model is
as follows:

An I/O multiplexer is represented as an object, with which I/O
channels can be registered to be monitored for events when the I/O
loop runs. An I/O channel is any object for which the generic
functions IO-CHANNEL-IOPORT, IO-CHANNEL-EVENTS and
IO-CHANNEL-HANDLE are implemented.

IO-CHANNEL-IOPORT, given an I/O multiplexer and an I/O channel,
should return the underlying system I/O facility that the channel
operates on. The actual objects used to represent an I/O facility
depends on the Lisp implementation, operating system and the
specific I/O loop implementation, but, for example, on Unix
implementations they will likely be numeric file descriptors. The
I/O loop implementation implements IO-CHANNEL-IOPORT methods for
the facilities it understands (such as FD-STREAMs on SBCL), so
user-implemented channels should simply call IO-CHANNEL-IOPORT
recursively on whatever it operates on.

IO-CHANNEL-EVENTS, given an I/O channel, should return a list of
the events that the channel is interested in. See the
documentation for IO-CHANNEL-EVENTS for further details.

The I/O loop guarantees that it will check what events a channel
is interested in when it is first registered, and also at any time
the channel has been notified of an event. If the channel changes
its mind at any other point in time, it should use the
IO-LOOP-UPDATE function to notify the I/O loop of such
changes. The I/O loop may very well also update spuriously at
other times, but such updates are not guaranteed.

IO-CHANNEL-HANDLE is called by the I/O loop to notify a channel of
an event.

An I/O multiplexer is created with a MAKE-INSTANCE call on the
class of the desired multiplexer implementation. If the code using
the multiplexer has no certain preferences on an implementation
(which should be the usual case), the variable *DEFAULT-IO-LOOP*
points to a class that should be generally optimal given the
current Lisp implementation and operating system.

Given a multiplexer, channels can be registered with it using
IO-LOOP-ADD, unregistered with IO-LOOP-REMOVE, and updated with
IO-LOOP-UPDATE (as described above). Call IO-LOOP on the
multiplexer to actually run it.
### *default-io-loop*
### *current-io-loop*
### *current-io-channel*
@@@ io-channel-ioport
@@@ io-channel-events
@@@ io-channel-handle
@@@ io-loop-add
@@@ io-loop-remove
@@@ io-loop-update
@@@ io-loop

@node Internal Functions Documentation, , IO Loop, Internals
@section Internal Functions Documentation

### *executing-stumpwm-command*
### *suppress-abort-messages*
!!! refresh-time-zone
!!! getsel
!!! putsel
!!! copy-unhandled-error
@@@ define-stumpwm-command
!!! set-contrib-dir

@node Interacting With Unix, Interacting With X11, Screens, Top
@chapter Interacting With Unix

!!! run-shell-command

@@@ programs-in-path
@@@ pathname-is-executable-p
@@@ pathname-as-directory


@@@ run-or-raise
### *shell-program*

@@@ getenv
@@@ (setf getenv)

@node Interacting With X11, Miscellaneous Commands, Interacting With Unix, Top
@chapter Interacting With X11

@@@ set-x-selection
@@@ get-x-selection
### *default-selections*
### *x-selection*

@node Miscellaneous Commands, Colors, Interacting With X11, Top
@chapter Miscellaneous Commands
The following is a list of commands that don't really fit in any other
@@@ split-string

@@@ argument-line-end-p
@@@ argument-pop
@@@ argument-pop-rest
### *display*
@@@ input-delete-region
@@@ input-goto-char
@@@ input-point
@@@ input-substring
@@@ input-validate-region
@@@ list-directory
%%% move-to-head
@@@ no-focus
### *record-last-msg-override*
### *toplevel-io*

@menu
* Menus::
* StumpWM's Data Directory::
* Debugging StumpWM::
* Sending a Bug Report::
* Timers::
* Getting Help::
@end menu

@node Menus, StumpWM's Data Directory, , Miscellaneous Commands
@section Menus
@section Menus
There are three different types of menus in StumpWM; single selection
menus; interactive menus; and batch menus. Single-selection menus, as
the name suggests, are used to pick a single item from a list.
Interactive menus are used for marking multiple selections. Batch
menus are used for performing actions on multiple menu items. Both
batch and interactive menus share the same navigational keybindings,
which are found in the table below. These can be customized by
modifying the @var{*menu-map*} variable. Commands specific to each
menu type can be modified by @var{*single-menu-map*} and
@var{*batch-menu-map*}.

@table @kbd
@item C-p
@itemx Up
@itemx k
Highlight the previous menu option.

@item C-n
@itemx Down
@itemx j
Highlight the next menu option.

@item S-Down
Scroll the entire page down one entry.

@item S-Up
Scroll the entire page up one entry.

@item PageUp
Scroll up one page.

@item PageDown
Scroll down one page.

@item C-g
@itemx ESC
Abort the menu.

@end table

In addition, you can customize the number of items shown at a time (a
page) with the @var{*menu-maximum-height*} variable. The default
value, @code{nil}, means that there is no limit to the page size.

### *menu-map*
@@@ menu-page-up
@@@ menu-page-down
@@@ menu-up
@@@ menu-down
@@@ menu-scroll-up
@@@ menu-scroll-down
@@@ menu-abort
@@@ menu-backspace
@@@ menu-entry-apply
@@@ menu-entry-display
@@@ menu-finish
@@@ command-menu

@subsection Single Selection Menus
Single selection menus can be searched; start typing when the menu is
active, and the results are immediately filtered. Use @key{RET} to
selected the highlighted option.

@table @kbd

@item RET
Select the highlighted option.

@end table

### *single-menu-map*
@@@ select-from-menu

@subsection Batch Menus
Batch menus provide a menu that allows the user to mark items. Items
are marked by highlighting an item, then pressing a corresponding
key. The key pressed depends on the menu being shown, and the desired
action. For example, in a menu allowing users to manage windows,
windows to be closed/removed could be marked by @key{d}, and windows
to be raised could be marked by @key{r}. All available actions and
their keybindings are shown below. These can be customized with @var{*batch-menu-map*}.

@table @kbd
@item n
@item Space
Highlight the next item.

@item p
Highlight the previous item.

@item u
Unmark the selected item, then move the cursor down.

@item DEL
Unmark the selected item, then move the cursor up if it is not at the
top of the menu.

@item x
@item RET
Exit the menu and perform the actions associated with each mark.

@end table
### *batch-menu-map*
@@@ select-from-batch-menu


@node StumpWM's Data Directory, Debugging StumpWM, Menus, Miscellaneous Commands
@section StumpWM's Data Directory
If you want to store StumpWM data between sessions, the recommended
method is to store them in @file{~/.stumpwm.d/}. StumpWM supplies
some functions to make doing this easier.

### *data-dir*
@@@ data-dir-file
%%% with-data-file

@node Debugging StumpWM, Sending a Bug Report, StumpWM's Data Directory, Miscellaneous Commands
@section Debugging StumpWM

### *debug-level*
### *debug-stream*
### *debug-expose-events*
@@@ redirect-all-output

@node Sending a Bug Report, Timers, Debugging StumpWM, Miscellaneous Commands
@section Sending a Bug Report
While StumpWM's code-base is quite mature, it still contains some
bugs. If you encounter one here are some guidelines for making sure
the developers can fix it:
@itemize @bullet
@item
Include a procedure for reproducing the bug/bad behavior. Ideally this
will include numbered steps starting with instructions on how you
start StumpWM. Also include what the expected behavior was.
@item
Be as detailed as possible. Then add more detail!
@item
Make sure its not something you introduced by using an empty @file{.xinitrc}
containing only @samp{exec /path/to/stumpwm}.
@item
Make sure the bug is present even when @file{.stumpwmrc} is empty.
@item
If you are using the git version, include the hash of the master
branch, or better include the commit when you started to notice the
bug.
@item
If you have code that fixes the bug, then open a pull request at
@url{https://github.com/stumpwm/stumpwm/compare/}.
@item
If you don't have code to fix the bug, then open an issue at
@url{https://github.com/stumpwm/stumpwm/issues/new}.
@end itemize

@node Timers, Getting Help, Sending a Bug Report, Miscellaneous Commands
@section Timers
StumpWM has a timer system similar to that of @dfn{Emacs}.

@@@ idle-time
@@@ run-with-timer
@@@ cancel-timer

@defun timer-p @var{timer}
Return T if TIMER is a timer structure.
@end defun

@node Getting Help,  , Timers, Miscellaneous Commands
@section Getting Help

!!! describe-key
!!! describe-variable
!!! describe-function
!!! describe-command
!!! where-is
!!! modifiers
!!! which-key-mode

@@@ lookup-command

### *help-map*
### *help-keys*

@node Colors, Hooks, Miscellaneous Commands, Top
@chapter Colors
When specifying a color, it is possible to use its X11 Color Name
(usually in the file @file{/etc/X11/rgb.txt}).  You can also use a
six digit hex string prefixed by a '#' character in the same way that
you can specify colors in HTML.

All text printed by StumpWM is run through a coloring engine before
being displayed. All color commands start with a @samp{^} (caret)
character and apply to all text after it.

@table @code
@item ^0-9
A caret followed by a single digit number changes the foreground color
to the specified color. A @samp{*} can be used to specify the normal
color. See the color listing below.

@item ^0-90-9
A caret followed by two digits sets the foreground and background
color. The first digit refers to the foreground color and the second
digit to the background color. A @samp{*} can be used in place of
either digit to specify the normal color. See the color listing below.

@item ^B
Turn on bright colors.

@item ^b
Turn off bright colors.

@item ^n
Use the normal background and foreground color.

@item ^R
Reverse the foreground and background colors.

@item ^r
Turn off reverse colors.

@item ^[
Push the current colors onto the color stack. The current colors remain unchanged.

@item ^]
Pop the colors off the color stack.

@item ^>
Align the rest of the line to the right of the window.

@item ^f<n>
Sets the current font to the font at index n in the screen's font list.

@item ^(<modifier> &rest arguments)
Allows for more complicated color settings:
<modifier> can be one of :fg, :bg, :reverse, :bright, :push, :pop, :font and :>.
The arguments for each modifier differ:

@itemize
@item
:fg and :bg take a color as an argument, which can either be a numeric
index into the color map or a hexadecimal color in the form of "#fff"
or "#ffffff".
@item
:reverse and :bright take either t or nil as an argument. T enables
the setting and nil disables it.
@item
:push and :pop take no arguments. :push pushes the current settings onto
the color stack, leaving the current settings intact. :pop pops color
settings off the stack, updating the current settings.
@item
:font takes an integer that represents an index into the screen's
list of fonts, or, possibly, a literal font object that can
immediately be used. In a string you'll probably only want to specify
an integer.
@item
:> takes no arguments. It triggers right-alignment for the rest of the line.
@item
:on-click takes one or more arguments and registers the following text
area as clickable. The initial argument must be an ID registered using
the function REGISTER-ML-ON-CLICK-ID.
@item
:on-click-end takes no arguments. It marks the end of clickable text
begun by :on-click. 

@end itemize

@item ^^
Print a regular caret.
@end table

The default colors are made to resemble the 16 VGA colors and are:
@table @asis
@item 0 black
@item 1 red
@item 2 green
@item 3 yellow
@item 4 blue
@item 5 magenta
@item 6 cyan
@item 7 white
@end table

There are only 8 colors by default but 10 available digits. The last
two digits are left up to the user. @ref{Behind The Scenes Look At
Colors} for information on customizing colors.

@menu
* Behind The Scenes Look At Colors::
@end menu

@node Behind The Scenes Look At Colors,  , Colors, Colors
@section Behind The Scenes Look At Colors

Color indexes are stored in @var{*colors*} as a list. The default list
of colors leave 2 slots for the user to choose. If you'd like to use
@samp{Papaya Whip} and @samp{Dark Golden Rod 3} you might eval the
following:

@example
(setf *colors* (append *colors*
                       (list "PapayaWhip"
                             "DarkGoldenRod3")))
(update-color-map (current-screen))
@end example

Of course, you can change all the colors if you like.

@@@ parse-color-string
@@@ uncolorify
### *colors*
@@@ update-color-map

@node Hooks, Modules, Colors, Top
@chapter Hooks
StumpWM exports a number of hooks you can use to add customizations;
like hooks in Emacs, you add to a hook with the @code{add-hook}
function. For example:

@example
(stumpwm:add-hook 'stumpwm:*new-window-hook* 'my-new-window-custos)
@end example

adds your @code{my-new-window-custos} function to the list of
functions called when a new window appears.

%%% add-hook
%%% remove-hook
%%% remove-all-hooks
@@@ run-hook
@@@ run-hook-with-args

The following hooks are available:

$$$ *new-window-hook*
$$$ *destroy-window-hook*
$$$ *focus-window-hook*
$$$ *place-window-hook*
$$$ *start-hook*
$$$ *internal-loop-hook*
$$$ *focus-frame-hook*
$$$ *new-frame-hook*
$$$ *message-hook*
$$$ *top-level-error-hook*
$$$ *focus-group-hook*
$$$ *hooks-enabled-p*
$$$ *remove-split-hook*
$$$ *key-press-hook*
$$$ *root-click-hook*
$$$ *click-hook*
$$$ *mode-line-click-hook*
$$$ *urgent-window-hook*
$$$ *event-processing-hook*
$$$ *pre-command-hook*
$$$ *post-command-hook*
$$$ *menu-selection-hook*
$$$ *new-head-hook*
$$$ *command-mode-end-hook*
$$$ *command-mode-start-hook*
$$$ *destroy-mode-line-hook*
$$$ *quit-hook*
$$$ *restart-hook*
$$$ *selection-notify-hook*
$$$ *split-frame-hook*

@node Modules, Hacking, Hooks, Top
@chapter Modules
A module is an ASDF system that adds additional functionality to
StumpWM.  StumpWM searches for modules in the
@var{*data-dir*}@file{/modules} directory.  By default this is
@file{~/.stumpwm.d/modules}.

Officially supported modules exist in a separate repository within the
StumpWM organization on github.  You can install the latest copy by
issuing @command{make install-modules} from StumpWM's root source
directory.  This will run:
@example
git clone git@@github.com:stumpwm/stumpwm-contrib.git ~/.stumpwm.d/modules
@end example


!!! load-module
@@@ list-modules
### *load-path*
!!! add-to-load-path
@@@ init-load-path
@@@ find-module

@menu
* Writing Modules::
@end menu

@node Writing Modules,  , Modules, Modules
@section Writing Modules
Make sure to read @ref{Hacking}. If you are familiar with writing lisp
packages for ASDF then you can jump in and get started.  In either
case, quicklisp ships a @code{quickproject} package that makes setting
up a new module very easy.  After installing quicklisp (see the
README.md for a link):

We're going to put our new module in the @file{modules/} directory of
@var{*data-dir*} so that it will be immediately loadable by StumpWM.

First make the directory @file{new-module}, then from a REPL issue:
@example
(ql:quickload "quickproject")
(quickproject:make-project #p"~/.stumpwm.d/modules/new-module" :depends-on '(stumpwm) :name "new-module")
@end example
This will create:
@example
  -rw-rw-r--  1 dave dave   68 Apr  6 19:38 package.lisp
  -rw-rw-r--  1 dave dave   53 Mar 16  2014 README.txt
  -rw-rw-r--  1 dave dave  271 Mar 16  2014 new-module.asd
  -rw-rw-r--  1 dave dave 1.8K Apr  6 17:51 new-module.lisp
@end example

The file @file{new-module.lisp} will contain the actual implementation
of your module.  ASDF requires two other files in order to understand
how to load and compile your module.  They are @file{new-module.asd}
and @file{package.lisp}.  In our example, @file{new-module.asd} should
contain:
@example
(asdf:defsystem #:new-module
  :serial t
  :description "Describe new-module here"
  :author "Anne N. O'Nymous"
  :license "GPLv3"
  :depends-on (#:stumpwm)
  :components ((:file "package")
               (:file "new-module"))) ; any other files you make go here
@end example
The @file{package.lisp} will contain:
@example
(defpackage #:new-module
  (:use #:cl :stumpwm))
@end example
With these two files defined, and the implementation written in
@file{new-module.lisp}, you should be able to load your module.

Before we load it, we have to add the path to our @var{*load-path*}.
This can be accomplished by running the following from a REPL:
@example
(stumpwm:add-to-load-path "~/.stumpwm.d/modules/new-module")
@end example
You can also run this interactively with @kbd{C-t ;}, which is bound
to the @command{colon} command.

Because we've put our module in a sub-directory of the default
@var{*module-dir*}, it will automatically get added to the
@var{*load-path*} the next time StumpWM starts.  If you choose to
develop your module somewhere else (e.g.
@file{~/quicklisp/local-projects}), then you'll have add
@example
(add-to-load-path "~/quicklisp/local-projects/new-module")
@end example
to your @file{.stumpwmrc}.

When you've finished writing your module, you can distribute it
however you see fit.  If it becomes very popular, or you would like
the StumpWM devs to maintain it (and they agree), you can have your
module merged with the stumpwm-contrib repository on github, just open
a pull request to start the discussion.

@node Hacking, Advanced Configuration, Modules, Top
@chapter Hacking

@menu
* General Advice::
* Adding Documentation and Editing This Manual::
* Using git with StumpWM::
* Sending Patches::
@end menu

@node General Advice, Adding Documentation and Editing This Manual, Hacking, Hacking
@section Hacking:  General Advice

@enumerate

@item
Pay attention to file names and contents. If you're making changes to
mode-line related code, don't put it in @file{core.lisp}. If you're
introducing some completely new featureset, consider putting all of
the new code in a new file.

@item
Does a command need to be user-visible (``interactive'') or is it just
called by other commands?

@itemize
@item
If it's not going to be user-visible, you can just use the familiar
@code{(defun foo () ...)} syntax.

@item
If you want the command to be used interactively, you use StumpWM's
@code{defcommand} syntax, as in the examples below.

@example
(defcommand test (foo bar)
   ((:string "How you're going to prompt for variable foo: ")
    (:number "How you want to prompt for variable bar: "))
   "This command is a test"
   (body...))

(defcommand test2 () ()
   "This is also a test"
   (body...))

(defcommand title (args) (interactive-args)
   "Doc string"
   (body...))
@end example

So basically, inside the first set of parentheses after the
function name, you specify what (if any) arguments will be passed to
the command. The second set of parentheses tells StumpWM how to get
those arguments if they're not explicitly passed to the command. For
example,

@example
((:string "What do you want to do: "))
@end example

will read a string from the input the user provides. The quoted
text is the prompt the user will see. Of course, if you were to, say,
call the command test, as defined above, from another piece of code,
it wouldn't give the prompt as long as you fed it arguments.
@end itemize

@item
Note that all commands defined using the @code{defcommand} syntax are
available both to be called with @kbd{C-t ;} and from within other lisp
programs, as though they had been defun-ned (which, in fact, they
have).

@item
Any code that depends on external libraries or programs that some
users might not have installed should be packaged as a module and
placed in the @file{*data-dir*/modules/} directory.

@item
Don't be afraid to submit your patches to the StumpWM mailing list!
It may not immediately make it into the official git repository, but
individual users might find it useful and apply it to their own setup,
or might be willing to offer suggestions on how to improve the code.

@end enumerate

@node Adding Documentation and Editing This Manual, Using git with StumpWM, General Advice, Hacking
@section Hacking: Adding Documentation and Editing This Manual

The manual is written in @command{texinfo}, so you may want to read
that manual. The @file{stumpwm.texi.in} is processed by StumpWM with
some additional markup in the form of three letter character entries
@c need to double up the at-signs in the @code{@@@ function} to escape
@c them.
at the beginning of a line. @code{@@@@@@ function} defines functions,
@code{%%% some-macro} expands to that macro and its docstring, etc.
Contributors are strongly encouraged to add these items to this manual
whenever something new is defined in a patch. You can test if your
texinfo edits are valid by generating them with
@command{make stumpwm.info}, and viewing the new @file{stumpwm.info} with
@command{info -f /path/to/stumpwm.info}, or @command{make
stumpwm.texi} for the raw stuff.

@table @asis
@item %%% macro
@item @@@@@@ function
@item ### variable
@item $$$ hook
@item !!! StumpWM command
@end table

@node Using git with StumpWM, Sending Patches, Adding Documentation and Editing This Manual, Hacking
@section Hacking:  Using git with StumpWM

For quite a while now, StumpWM has been using the git version control
system for development. If you're using one of the official
releases, you can get the bleeding-edge source code from the official
git repository with a single command:

@example
$ git clone git@@github.com:stumpwm/stumpwm.git
@end example

After this, you'll have a complete git repository, along with the
complete revision history since the switch. Feel free to play around;
git has some important features that actually make this safe!

Before we get to that stuff, though, you're going to want to tell git
about yourself so that your information is included in your commits
and patches. The very minimum you're going to want to do is:

@example
$ git config --global user.name "Anne N. O'Nymous"
$ git config --global user.email "anonymous@@foo.org"
@end example

Be sure to check out the manual for @command{git-config}--there are several
options you might want to set, such as enabling colorized output or
changing the editor and pager you use when making commits and viewing
logs.

For the sake of argument, let's say you want to make some major
changes to both @file{user.lisp} and @file{core.lisp}, add a file
called @file{DANGEROUS_EXPERIMENT_DO_NOT_USE_OR_@-ELSE.lisp}, and
remove the manual because you're too 1337 for such things. However,
you don't want to break your entire StumpWM setup and start over.
Thankfully, you don't have to. Before you get started, issue this
command from the StumpWM source directory:

@example
$ git checkout -b experimental
@end example

You should now find yourself in a new branch, called experimental. To
confirm this, type @command{git branch}; there should be
an asterisk next to the branch you're currently viewing. At any time,
you can type @command{git checkout master} to return to your master branch,
and at any time you can have as many branches of the project as you
like. If you want to create a new branch based not on the master
branch but on your experimental branch, for example, you'd type:

@example
$ git checkout -b new-experiment experimental
@end example

This will place you in a newly-created branch called ``new-experiment''
which should be identical to your experimental branch as of the last
commit (more on that soon). If you're actually typing out the
directions, switch back to your old experimental branch like so:

@example
$ git checkout experimental
@end example

Anyway, now that you have a new branch, create that new file with the
long name, which we'll just call @file{danger.lisp} for brevity. Make whatever
changes you want to it, and when you're done, tell git about your new
file.

@example
$ git add dangerous.lisp
@end example

Now, let's pretend you're done making changes. Tell git you're done
for now:

@example
$ git commit
@end example

This will open up a prompt in your editor of choice for you to
describe your changes. Try to keep the first line short, and then add
more explanation underneath (for an example, run the command @command{git log}
and take a look at some of the longer commit explanations). Save that
file and then do this:

@example
$ git checkout master
$ ls
@end example

Then look for your new file. It's not there! That's because you've
done all of your work in another branch, which git is currently hiding
from you so that you can ``check out'' the branch called ``master.'' All
is as it should be---your master repository is still safe.

@example
$ git checkout experimental
@end example

Now, delete @file{manual.lisp} and @file{stumpwm.texi}. That's right. Wipe them off
the face of the Earth, or at least off the hard drive of your
computer. When you're done, you don't have to tell git you've deleted
them; it'll figure it out on its own (though things may not compile
properly unless you edit @file{Makefile.in} and @file{stumpwm.asd}. Anyway, go ahead
and edit @file{core.lisp} and @file{user.lisp}. Really break 'em. Run free! When
you're done, do another commit, as above, and give it a stupid title
like ``lolz i b0rked stUmpwm guys wTF!?!?!!111!'' Now try to compile.
Just try. It won't work. If it does, you're some kind of savant or
something. Keep up the good work. If you've actually managed to break
StumpWM like you were supposed to, never fear! You have two options at
this point.

One is to go back to the master branch (with another git checkout) and
just delete your experimental branch, like so:

@example
$ git branch -D
@end example

The ``@code{-D}'' means to force a delete, even if the changes you've made
aren't available elsewhere. A ``@code{-d}'' means to delete the branch if and
only if you've merged the changes in elsewhere.

The other option is to create patches for each of your commits so far,
delete the branch, and then apply any working/wanted patches in a new
branch. Create your patches (after committing) like so:

@example
$ git format-patch -o patches origin
@end example

(Before doing that you can review your changes with @command{git log origin..})

You can also use the @command{format-patch} command to create a patch of working
code to send in to the mailing list.

A developer might ask you to try out something they're working on. To
fetch their master branch, you'd do this:

@example
$ git remote add -f -m master -t master foo git://bar.org/~foo/stumpwm
@end example

Here, ``foo'' is the shorthand name you'll use to refer to that
repository in the future. To checkout a local copy of that repository,
you'd then do

@example
$ git checkout --track -b foo-master foo/master
@end example

Later you could use @command{git pull foo} to update while looking at that
branch (and note that @command{git pull} with no arguments, in the master
branch, will update your StumpWM from the official repository).

Finally, if you want to move your experimental changes into your
master branch, you'd checkout your master branch and run:

@example
$ git merge experimental
@end example

If there are file conflicts, @command{git diff} will show you where they are;
you have to fix them by hand. When you're done, do another

@example
$ git commit -a
@end example

to finalize the changes to your master branch. You can then delete
your experimental branch. Alternately, you can wait until your changes
(assuming you sent them in) make it into the official repository
before deleting your experimental branch.

@node Sending Patches,  , Using git with StumpWM, Hacking
@section Sending Patches
While patches are still welcome on the mailing list, StumpWM's
development has mostly migrated to github's issue tracker.  This means
you can open a pull request to submit a patch to StumpWM.  The
following guidelines apply to pull requests and patches sent to the
mailing list.

@itemize
@item
Make sure it applies clean to the main git repository

@item
Ensure that you aren't introducing tabs, extra blank lines, or whitespace at the end of lines.

@item
Ensure your patch doesn't contain irrelevant indenting or reformatting changes.

@item
Try to make your patch address a single issue. If your patch changes
two unrelated issues, break them into two seperate patches that can
stand on their own.

@item
Don't send intermediate patches. When you're working on a feature you
might make several commits to your local repository as you refine it
and work out the bugs. When it's polished and ready to ship, send it
as one patch! Sometimes it makes sense to send it as multiple patches
if each patch contains a discrete feature or bug fix that can stand on
its own. If one of your patches changes code that was added or
modified in an earlier patch, consider merging them together and
sending them as one.

@end itemize

@node Advanced Configuration, Command and Function Index, Hacking, Top
@chapter Advanced Configuration

### *default-package*
### *default-bg-color*
@@@ run-commands
### *startup-message*
### *list-hidden-groups*
%%% defprogram-shortcut
### *initializing*
!!! loadrc
### *ignore-wm-inc-hints*
### *max-last-message-size*
### *module-dir*
@@@ set-module-dir
### *mouse-focus-policy*
### *resize-hides-windows*
### *root-click-focuses-frame*
### *suppress-frame-indicator*
### *suppress-window-placement-indicator*
### *text-color*
### *draw-in-color*
### *timeout-frame-indicator-wait*
### *top-level-error-action*


@node Command and Function Index, Variable Index, Advanced Configuration, Top
@unnumbered Command and Function Index
@printindex fn

@node Variable Index,  , Command and Function Index, Top
@unnumbered Variable Index
@printindex vr

@bye