Do not allow focus to drift from fullscreen client via focusstack()

It generally doesn't make much sense to allow focusstack() to navigate
away from the selected fullscreen client, as you can't even see which
client you're selecting behind it.

I have had this up for a while on the wiki as a separate patch[0], but
it seems reasonable to avoid this behaviour in dwm mainline, since I'm
struggling to think of any reason to navigate away from a fullscreen
client other than a mistake.

0: https://dwm.suckless.org/patches/alwaysfullscreen/

Fix x coordinate calculation in buttonpress.

dwm.1: fix wrong text in man page

Fix memory leaks in drw

The function drw_fontset_free in drw.c was never called.

dwm crashes when opening 50+ clients (tile layout)

Many users new to dwm find themselves caught out by being kicked out to the login manager (dwm crashing) when they open 50+ clients for demonstration purposes. The number of clients reported varies depending on the resolution of the monitor.

The cause of this is due to how the default tile layout calculates the height of the next client based on the position of the previous client. Because clients have a minimum size the (ty) position can exceed that of the window height, resulting in (m->wh - ty) becoming negative. The negative height stored as an unsigned int results in a very large height ultimately resulting in dwm crashing.

This patch adds safeguards to prevent the ty and my positions from exceeding that of the window height.

drawbar: Don't shadow sw global

This jarred me a bit while reading the code, since "sw" usually refers
to the global screen geometry, but in drawbar() only it refers to
text-related geometry. Renaming it makes it more obvious that these are
not related.

getatomprop: Add forward declaration

No functional changes, but for every other function we have a forward
declaration here. getatomprop should be no exception.

setmfact: Unify bounds for compile-time and runtime mfact

There are two places that mfact can be set:

- In the mfact global, which is defined at compile time and passed
  into m->mfact during monitor setup. No bounds checks are performed,
  but the comment alongside it says that valid values are [0.05..0.95]:

      static const float mfact = 0.55; /* factor of master area size [0.05..0.95] */

- By setmfact, which adjusts m->mfact at runtime. It also does some
  minimum and maximum bounds checks, allowing [0.1..0.9]. Values outside
  of that range are ignored, and mfact is not adjusted.

These different thresholds mean that one cannot setmfact 0.95 or 0.05,
despite the comment above that lists the legal range for mfact.

Clarify this by enforcing the same bounds in setmfact at runtime as
those listed for mfact at compile time.

config.mk: fix POSIX_C_SOURCE macro for feature test for snprintf()

The feature test was incorrect:
_POSIX_C_SOURCE=2

"The value 2 or greater additionally exposes definitions for POSIX.2-1992."
http://man7.org/linux/man-pages/man7/feature_test_macros.7.html

A higher value is needed (atleast 1995):
https://pubs.opengroup.org/onlinepubs/9699919799/functions/snprintf.html

FreeBSD feature test macro:
on
https://github.com/freebsd/freebsd/blob/master/include/stdio.h line 297

This was already fixed in dmenu.

This fixes a warning on FreeBSD, reported by Plasmoduck on IRC, thanks.

Prepare 6.2 release.

pledge: add rpath promise for the ugly Xft font fallback

Makefile: just show the compiler output

Don't be fancy and just show the actual output so debugging is simpler.

Do not strip at link stage

Building with debug symbols is worthless unless LDFLAGS are manually
adjusted as well.

Pledge on OpenBSD

config.def.h: ClkTagBar missing from comment

by Christopher Drelich <cd@cdrakka.com>

Patch was mangled on the ML, also adjusted the order to be the same as
the enum in dwm.c

Function declarations in correct order.

In dwm.c function declarations are in alphabetical order except for
updategeom(). There doesn't appear to be any reason for this, so this
patch corrects that, and now all function declarations are in
alphabetical order.

remove old TODO and BUGS entries

the bug in the dwm man page is an (ancient) Java issue.

Thanks David and quinq for the patches and feedback!

update README: remove mentioning the old dextra repo

Thanks Christopher Drelich <cd@cdrakka.com>

All functions in alphabetical order except for this one.

ColBorder has been moved to the enum with ColFg and ColBg.

dont NUL terminate _NET_WM_NAME

Reported by Kernc, thanks!

"This makes a particular program that uses libwnck [1] fail after:

    Wnck-WARNING **: Property _NET_WM_NAME contained invalid UTF-8

in this code [2] because the returned string contains a '\0' and the
documentation for g_utf8_validate() [3] explicitly states that when
string length is provided, no nul bytes are allowed."

It is not entirely clear it is incorrect, other WM's seem to not
NUL terminate it either though.

sync dmenu drw.{c,h} code: use Clr* (was Scm)

gettextprop: check result of XGetTextProperty (undefined behaviour for XFree)

Set class name on status bar

This is useful for configuring compositors to ignore the status bar
window.

simplify isfixed conditions

yet another cleanup

The previous patches introduced some unclean space-based indentation
patterns. This patch fixes them.

Don't restrict snap in mousemove

This also fixes a bug where client windows only switch to floating mode when the
mouse is dragged in one specific direction.

Button passthrough when client is not focused

Before this change it is not possible to press a button in a client on the first
click if the client is not yet focused. The first click on the button would
only focus the client and a second click on the button is needed to activate it.
This situation can occur when moving the mouse over a client (therefore focusing
it) and then moving the focus to another client with keyboard shortcuts.

After this commit the behavior is fixed and button presses on unfocused clients
are passed to the client correctly.

cleanup

- unify multi-line expression alignment style.
- unify multi-line function call alignment style.
- simplify client moving on monitor count decrease.
- clarify comment for focusin().
- remove old confusing comment about input focus fix in focusmon(). The
  explanation is already in the old commit message, so no need to keep it in the
  code.
- remove old comment describing even older state of the code in focus().
- unify comment style.
- break up some long lines.
- fix some typos and grammar.

applied Markus' tagset purge of alternative view on _NET_ACTIVE_WINDOW event

applied Ivan Delalande's NET_SUPPORTING_WM_CHECK patch for gtk3 compatibility

applied Ian Remmler's man page adjustment suggestions

applied Markus' decouple color-scheme patch

applied Markus' clarify status text padding patch

LICENSE: update people

dwm.1: add keybinding for spawning dmenu

die() on calloc failure

thanks Markus Teich and David!

die() consistency: always add newline

config.def.h: style improvement, use color Scheme enum

import new drw from libsl and minor fixes.

- better scaling for occupied tag squares.
- draw statusline first to omitt some complicated calculations.

Configure geometry before applying rules

Configuring geometry before applying rules makes it possible to have
more complex constraints in applyrules that depend on the initial window
dimensions and location.

fix fullscreen clients not resized on X display resolution change

patch provided by Bert Münnich <ber.t_AT_posteo.de>, thanks!

Shut up glibc about _BSD_SOURCE being deprecated

code-style consistency

unboolification

sort include + whitespace fix

separate program-specific c99 bool and X11

True, False are X11-specific (int), make sure to use c99 stdbool for
program-specific things.

Makefile: package all files with make dist

setfullscreen: don't process the property twice

Some clients try to set _NET_WM_STATE_FULLSCREEN even when the window is
already in fullscreen.
For example, c->oldstate was set two times in a raw and window would
then always be floating.
We must check that it's not the case before processing it.
(original patch modified with suggestion from Markus Teich
<markus.teich@stusta.mhn.de>)

sync updated drw code from dmenu

important:
- drw_rect: didn't use w and h, change the dwm code accordingly.
- drw_text: text is NULL is not allowed, use drw_rect().

dwm: use ecalloc, prevent theoretical overflow

cleanup, dont use c++ style comments

- signal: print error string.
- die: start message with lower-case (consistency).
- bump version to 2015.

dwm: cleanup: free schemes and cursors as array

config.h: use common default font, pass Xft font name to dmenu

dmenu uses Xft now (soon to be released).

config.mk: add $FREETYPELIBS and $FREETYPEINC, simpler to override (ports and *BSDs)

Add Xft and follback-fonts support to graphics lib

removed .hgtags, thanks Dimitris for spotting

applied Hiltjo's resize/move limitation

"Limit the amount of updates when resizing or moving a window in floating
mode to 60 times per second. This makes resizing and moving alot smoother
and by limiting it it also uses alot less resources on my machine.

same as before with dwm.c as well

updated copyright notice in LICENSE file

applied Lukas' focus suggestion at startup, thanks

applied improved version of Martti Kühne's dmenu/multi monitor approach from dwm, no dmenuspawn required

do not take our font declaration as default for st

applied Jochen's drw_text patch, thanks

applied Julian's enum approach,
	however renamed theme into scheme resp. Theme into ClrScheme

finished libsl/drw integration

include font argument for st by default

added st to SEE ALSO section

use st as default terminal from now on

shut up about deprecated Xlib functions

renamed draw into drw

continued with draw.c and draw.h implementation, now the integration begins

continued with draw.c abstraction, also started util.{h,c} implementation, that will be used by draw.c as well

removed DDC, all is Draw-dependent

continued, distinction of Draw and DDC is bad, needs to be merged

reverting the xkb dependency, I don't care if this function is deprecated, it seems it breaks other stuff instead.

basic draw.c structure

reverting to plain X11 fonts in order to implement draw.c default

removed obsolete bugs from BUGS file

compile fix

applied Neil Klopfstein's patch, slightly modified

starting with initial draw.h

incorporating Xft instead of cairo, cairo provides far too many options

keep 6.1 intact

reverted to old updategeom() after several complains, we need to optimize the old way

applied James Turner's XkbKeycodeToKeysym patch, thanks

drastically changed updategeom() handling, see comment in updategeom() for details

reversed Andreas Amann's fullscreen fix with the approach proposed by Gary Langshaw:
- idea is not supporting movemouse/resizemouse/togglefloating for fullscreen windows
- as fullscreen windows are broken anyways, they should only be adjusted by their own means

added kludge to fix some input focus misbehavior in gedit and anjuta, thanks Martti Kühne

applied Andreas Amanns' netwm_client_list patch, but with some indentation fixes

applied Andreas Amann's netwm_active_window patch, thx

applied Andreas Amann fullscreen fix, some minor modifications

unfocus on slow sloppy monitor focus

applied Eckehard Berns fix fix, sorry that this took so long

added 20h's clarification

applied Eckehard Berns barwin leak fix and his suggestion to deal with restack() -- the latter aspect needs further investigation.

config.mk cleanup

Added tag 6.0 for changeset ec4baab78314

bump version to 6.0

add sloppy monitor focus

new default colour scheme

calculate window/monitor intersection

honour fullscreen hint on map

testing Brians multiscreen issue fix

make ewmh dialog windows float

allow 0 nmaster

fix big-border corner case

added keyrelease remark to TODO

improve tile spacing

add nmaster binds to manpage

apply nmaster patch

apply resize hints in floating layout

hide clients to the left, not the right

applied Connors cleanup patch of Eckehards proposed fix of applyrules(), thanks everyone involved

applied Connors and Valentins patch to improve the unmapnotify handling of broken clients

applied Peter Hartlichs nice interim Xinerama and map fix patches, for debugging purposes I also added his transient test driver

applied anonymous code cleanup patch

added a marker for the flash fullscreen issue

changed sleep 20 into sleep 1 in example script as suggested by Snader_LB

applied Connors aesthitic buf fix in tile(), thanks

applied Peters magic float mode bugfix

applied lolilolicon's floating center patch, further investigation wrt his second issue reported needed

thanks to recursions on IRC for his remark

Added tag 5.9 for changeset dd74622a4785

enabled release flags

applied Peters wintoclient/wintomon optimisation in enternotify()

removed unneeded offset recalculation, thanks Jukka, let's see if this breaks some other client

undo the focus optimisation patch for 5.9

applied Peters two patches, please test hg tip, if nothing breaks this is 5.9

applied Garys patch in a slightly modified way

applied Nicolas Capit's patch as interim solution until the multiscreen support is reworked in 6.0

fixed scroll lock (thanks bogdan)

fix numlock (thanks mikhail)

making enternotify less focus hungry

applied Andreas Amann's patch from Oct 2010, thanks

applied Rudys barwidth patch, thanks Rudy

don't check for monocle layout in zoom() exit condition

Thanks for the pedantic review :)

fixing some minor issues, next week is dwm-5.9 release time

rm draw.c from 5.9 branch

fix typo

cleaner drawing

update draw.c

added draw.h, draw.c

applied Evil_Bobs cleanup patch

applied another minor change to the README

applied Thomas' README patch

applied Jack's mplayer fullscreen patch

missed declaration

minor fix of the NetActiveWindow multi-monitor flaw, slight rearrangement

applied Peter/Andreas NetActiveWindow patch in a slightly modified version

applied Brendan MacDonell's WM_TAKE_FOCUS patch, thanks Brendan!

applied another patch of Hiltjo, thanks

applied Hiltjos' BUGS patch from 23 Mar, sorry took a while :)

applied Hiltjo's multimon mouse-based resize fix

applied anonymous patch, I don't think the reversed array access has semantic side-effects

added a todo

applied Hiltjo's cleanup patch

applied Hiltjo's tiny cleanup fix

applied Hiltjo's tiny cleanup patch, thanks!

does this make a difference?

added TODO and bumped version to 5.9

applied Hiltjo Posthuma's line saver patch

applied Hiltjo Posthuma's double-XineramaQueryScreens() patch, thanks Hiltjo!

Added tag 5.8.2 for changeset a043f0800805

increased version

implemented better fullscreen handling, please test

Added tag 5.8.1 for changeset 1ed1e75c9c2e

removing Sylvain's patch because it breaks more than it fixes unfortunately, re-issuing a bugfix release 5.8.1

Added tag 5.8 for changeset 60ea8fed13ab

removed DEBUG code and added 2010 to the copyright list

applied Sylvain Laurent's EWMH fullscreen state patch, simplified his patch a bit

added the related bug reported by voltaic

pretty printing bugs a bit

selmon optimisation is needed

more debug output, experimental focus optimisation

quick typo fix

it's debug season for upcoming dwm 5.8, so only use this if you really want to run experimental code or debug code

another bug report

removed misleading comment, thanks Romain Bertrand

added nsz's last bug corner case

added BUGS, next version is 5.8

Added tag 5.7.2 for changeset 15761ac5e2f1

applied Ryan Zheng's patch and re-releasing 5.7.2

Added tag 5.7.1 for changeset 48c3f87c335d

fixed the layout symbol bug reported by Nibble (but using a different approach as he suggested)
- optimised drawbar() and arrange() handling for multiple monitors, arrange only arranges the relevant monitors when applicable, same with drawbar
- need to release 5.7.1

Added tag 5.7 for changeset 257403d4cd96

switching to release flags

implemented nn < n case, k-zed please recheck

use buffer instead of pointer for mon->ltsymbol

added missing scan

implemented different version of updategeom

added TODO to updategeom in order to implement a decent version of it soon

new experimental updategeom() additions that should avoid several problems with Xinerama, this is EXPERIMENTAL and might break something, the algorithms in use are quite complex and cumbersome, patches and comments welcome

implemented dynamic layout symbol stuff

applied Tony Lainson's config.def.h patch

update

hard-core destruction

renaming isdestroyed into destroyed

sync when a client is being killed

added isdestroyed flag to unmanage

always updategeom when root is configured

another small optimisation

also update title if the client is on an unfocused monitor

applied nsz's dwm.1 patch, also added wmname

applied Jukka's cosmetic patch

applied Jukka's sigchld patch

added merged patch of anydot and Neale

fixed nn declaration

Added tag 5.6.1 for changeset e47a47bd3ed4

applied nsz patch

updategeom fix for same geom screens

applied nsz' man page comment, thanks!

merge

removed Standard input hint

removed misleading comment

merge

it's all nsz's hard investigation effort, hail nsz! ;)

final style fixes

removed int cast in TAGMASK as suggested by nsz

some stylistic changes

removed obsolete BUG

mark broken clients as broken

fixing updatetitle

extended rule to apply monitors if set up accordingly

reverting to optimised compiler options, current state seems stable

extended man page to mention number of visible windows

added client number printing right of layout symbol

removed monsyms, useless

restricting number of mons by length of monsyms

introducing const where it might make some sense

no that's definately better

some minor changes

die if malloc sizeof(Monitor) fails

applied Mate's patch, added Mate to LICENSE

some cleanups

hotfix

changed focusmon/tagmon to work on prev/next instead (-1/+1), changed shortcuts to Mod1-, Mod1-. and Mod1-Shift-, Mod1-Shift-.

removed some empty lines

interactive resizals allow display dimensions again

more sensible

hmm, can't decide on the default monitor symbol set

update

introduced monitor symbols

minor fix

several bugfixes

some minor fix if xinerama is disabled, still some odd behavior in there

updated year ranges

fixed usage of sx, sy, sw, sh

don't draw monitor number, if there is just one monitor

fixed order of transient monitor applicaion and rule application

hotfix

several simplifications

play safe

applied a collection of Christof Musik patches, and fixed an issue reported by waistcoats on IRC

some cleanup handling for index based mon search

using mon in function names for consistency reasons

fixed several issues with focus handling via mouse, also added sending clients to the right monitor they belong to after mouse moves/resizals

fixed focusmon brokeness

tag fix

typo fix

assign selected tags of target monitor to client when tagmon is performed (less obtrusive imho)

added some TODOs, some other focus fixes

disabled XINULATOR code for now

made bar based monitor switching working

several other focus fixes, introduced unfocus()

some focus fixes

fixed title rendering on non-active screen

fixed getclient brokeness

some hotfix cleanup related to wild selmon-> destruction

some restack fixes, still odd behavior if n>1 on unmanaging clients

added some dual head simulator code

simplified left over re-assignment

fixed tagmon, re-using detach/detachstack and attach/attachstack

simplified ISVISBLE and nexttiled

moved clients/stack/sel to Monitor, that's a per monitor structure from now on

declare buf only if XINERAMA support is used

if compiled without XINERAMA support assume 1 monitor by default (n = 1)

some more refactoring, bar toggle fixes

hotfixes

changes monitor structure to be a list

we need a tagset per monitor, removed tagset declaration from config.h

some minor fixes regarding the new xinerama support

several fixes through ISVISIBLE change (takes Monitor into account)

experimental xinerama support, two new actions, Mod1-w/e and Mod1-Shift-w/e

introduction of Monitor for better Xinerama support, work in progress

removed MAXTAGLEN

next version will be 5.6, shortened some very long lines

Added tag 5.5 for changeset deaa276abac1

applied Marc Andre Tanners showhide patch, the removal of ntiled

applied Gottox' patches, and also removed usegrab

applied the patch I missed from Gottox, thanks mate!

applied Gottox' resizehints patch, thanks Gottox!

some NULL fixes

override_redirect is Bool

applied Gottox' adjustborder removal patch (thanks!)

applied nsz' remark

fix of typo, thx to Enno Gottox Boland

Added tag 5.4.1 for changeset 85a78d8afa0f

fixed an issue reported by Nibble, also fixed s/2008/2009/

Added tag 5.4 for changeset 8b7836a471f8

avoid reverting the border if it's not necessary (avoids some flashing on view())

spotted missing spaces

reverting some border patches

changed adjustborder to be a macro

applied yiyus applyrules() patch

implemented noborder for 1 client in the view

made status script example of .xinitrc more useful

ok, using signal.h for portability reasons

removed unnecessary closure of stdin reported by Frederic Chardon, thx

added sys/signal.h inclusion for BSD compliance

re-applied Neale's spawn patch, credited Neale in LICENSE

applied Neale Pickett's xprop status reading patch, updated README and dwm.1 accordingly

applied yiyus fix

fix

Added tag 5.3.1 for changeset 335301ed102f

integrated yiyus clearurgent refactoring

reverted spawn

Added tag 5.3 for changeset 4004d6116035

added Neale Pickett's spawn patch, thanks Neale

applied Gottox' comment fix (thanks)

several changes towards 5.3, XINERAMA is disabled by default, introduced usegrab Bool for grabbing the server during mouse-based resizals/movements (disabled by default), continued debugging tile() with resizehints == True and a lot of terminals supporting resizehints, still no optimal solution, I need to think about it, considering recursive algorithm for the space optimization

removed NOBORDER(broken and badly used), introduce WIDTH/HEIGHT macros for computing client size with border added, fixes bug announced by sergey_m

Added tag 5.2 for changeset e4bcaca8e6ef

using malloc() instead of calloc() and sticking to static initializer and struct assignment, not using *c = (Client){}; right now to avoid some ugly gcc warning, possibly a gcc bug since we are using -std=c99

applied XGetWMNormalHints fix

I prefer doing the check in showhide

applied Donald Chai's showhide patch in slightly modified ways

applied Martin Hurton's checkotherwm simplification

applied Peter Hartlichs aspect revert fix

applied Gottox' grabkey patch

applied yiyus tagbar patch

merged merge

introduced NOBORDER macro to hide the nasty - 2 * c->bw in various calculations, moved tagset to config.def.h

grabbuttons() and grabkeys() are now independent from prior numlockmask initialization

reverted some resize() changes, reverted setlocale removal

applied Peter Hartlich's patch regarding aspect calculation with slight modifications

removed artifact from wmii

changed grabkeys, removed initmodmap

applied Peter Hartlich's initmodmap patch

applied fix of toggletag by Jan Kaliszewski

Martin Hurtons typo fix

fixed error I did when applying Martin Hurton's drawtext patch

applied Martin Hurtons resizemouse patch

applied Martin Hurton's movemouse() patch

using None instead of 0 for trans

abc...

removed the i = textnw... as remarked by Martin Hurton

applied Martin Hurton's scan() patch with slight modifications

changed order of variables in drawtext

applied Martin Hurton's drawtext() patch

applied Martin Hurton's view() simplification, not checking arg

making a comment more explicit

removed setlocale() stuff, not necessary if Xmb in use

added a comment about FAQ regarding mfact meaning

added some comments regarding FAQ about s{x,y,w,h}, w{x,y,w,h}, b{y,h,lw}

initialize trans with 0

fix of XGetTransientForHint in manage()

applied cleanup from M. Hurton and S. Nagy

fixes using arg->i instead of arg->ui

applied Johannes Hofmann's patch, please test

made readin a config.h variable

Added tag 5.1 for changeset ce355cea9bb8

forcing fullscreen windows to bw=0, though most fullscreen apps are broken anyways

potential crash fix if xinerama behaves broken, though I doubt it

make hg tip compilable with default config

local use of xidx is useless, got rid of it, falling back to screen 0 if pointer query fails for whatever reason

got rid of compile time xidx configuration, querying mouse pointer instead

reverted uint redefinition

applied anydot's urgency hint patch, thanks!

renamed eprint die

simplified detach()

removed aux* stuff from Client

removed useless comment

minor fix to view()

applied Frederik Ternerot's grabbuttons patch with slight modifications

locale update

removed useless characters

applied James Turner's dwm.1 patch, thanks James!

fix of monocle

another merge

does this fix anything?

applied Gottox' ClkTagBar patch

made arrange again like it was once

fix

use sel instead of seeking the list

resize should apply if !banned

applied Gottox patch

untested monocle

non-zero

Added tag 5.0 for changeset 06eb9644e2da

branch merge

tiled layout resizehints should be respected by default

restored y-coordinate fixing of client windows

updated man page regarding Mod1-m

minor fixes towards 5.0

removed explicit warp in movemouse

removed scroll-wheel based focussing on window title clicks

fix of swapped focusstack mouse buttons

removed the ButtonPressMask for root windows as well

removed root window click handling

removed font and color definitions

minor fix

fixed tag click handling, however ClkRootWin doesn't work for me

update

some bugfixes of the patch application yesterday

integrated yiyus mouse.diff (though the bar click handling is slightly broken, I'm to tired to debug it now, yiyus could you please?)

made Xinerama screen index customizable

removed trailing spaces reported by Soleen

crash fix

made lt visible again in config.def.h

added nsz' patch

added SHCMD support for pipe-based commands due the new spawn() versio

using foo layout during cleanup (suggested by Gottox), add Arg->v handling for togglelayout() suggested by anydot

lt will point to a foo-layout during cleanup now (Gottox' suggestion), and togglelayout respects Arg->v

integrated Peter Hartlich's patch, removed const char *c from union, simplified togglelayout

applied anydot's patchset.diff

final version -- Gottox verified it using the test driver

revert of introduced problem

applied nsz' textnw patch thank you

updated the for-loop with Gottox' proposal

minor change

applied nsz's patches (many thanks!)

small change to drawtext

fixed the tile() issue with xpdf

using anydot's memcpy-approach in drawtext, however it still looks awkward to me

Gottox' drawtext simplification

applied noviewprev.diff, fix.diff and unusedflags.diff

fix

lt->arrange in mfact

isfixed implies isfloating

applied nibbles fixes, slightly modified

applied yiyus tip patch from tue

core dump fix in cleanup

updated the initial comment to reflect the use of bit arrays

slight changes

removed bx and bw, unnecessary

renamed domax into ismax

simplified tile()

removed Layout->updategeom, unnecessary

removed TEXTW

applied yiyus domax patch with slight modifications

replaced isvisible with a macro

removed emallocz

s/int/uint/ in config.h

removed debug output, sanitized tag limit check

setmfact argument was wrong

Key.mod is uint, Client.[old]bw is int

s/unsigned long/ulong/

s/nextunfloating/nexttiled/, changed zoom() behavior

s/unsigned int/uint/

s/void */const void */

applied Gottox bitmask + void *arg patch

some minor fixes

simplification

reverted dist target in Makefile

merged tile.c again into dwm.c

take bar into account

be more polite to clients which like to appear outside the window area, but still on the screen

fixed comment

fix

make it easier for the user, if Xinerama support is given, always use the screen 0 as window area/bar area, everything else can be used for floating clients

improving space usage if master is left of stack (default)

only snap within window area

s/DEFGEOM/CUSTOMGEOM/

update

updatetilegeom should be fine for setmfact

recent changes, introduced togglebar, changed some defines into variable declarations where possible

new stuff

removed the <M> togglelayout call

s/tilegeom/updatetilegeom/

removed monocle for now

removed the exact focus mechanism of next/prev window

moved all tile()-related stuff into tile.c which is included from config.def.h, the default dwm is now nearly independent from the arrange() algorithm in use

removed tileh, renamed tilev into tile again, removed counttiles, tilemaster

removed Layout->isfloating

removed reapply() -- unnecessary

renamed setlayout into togglelayout

removed temporary wildcard handling

added wild-card handling for tags

applied nsz's another style patch

applied nsz's style.diff patch

just added a comment

minor fixes

having monocle on Mod1-m

hotfix

removed Geom stuff, introduced updategeom() again, still view is somewhat broken?

cleaned up config.def.h

implemented exact focus next, if arg != NULL to focus{next,prev}

applied the proposal by nsz

applied JUCE patch

simplification of view() as proposed by anydot

applied yiyus tagset patch

applied yiyus fgeom patch

applied dfenze drawtext simplifications

this is the correct way

copyright notice is in LICENSE

applied Ph's seltags-simplification with some modifications

applied Ph's MIN/MAX patch, nice work!

applied Ph's patch regarding geom and lt initialization

applied Gottox' applyrules() fix

geometry fix proposed by Jukka

applied applyrules-fix by Jukka, thank you Jukka!

fix

applied the proposed monocle patch from the mailinglist

next version is 5.0

Added tag 4.9 for changeset 22c669b2dd36

yet another cleanup

some cleanup changes

uncommented dual layout in preparation of dwm 4.9

applied Peter Hartlich's simplification patch of setmfact and his revival of MFACT, appliead Janness Hofmann's simplification of grabbuttons() -- thanks guys!

aspects hints seem broken for fullscreen apps

removed uneccessary line

bugfix

setmfact should not have any effect if in floating layout

minor fix

revival of mfact and setmfact

setlayout and setgeom are now togglable again

geom indicator and layout indicator is only displayed if there are several geoms/layouts

hotfix of idxoftag

blw/bgw calculation bugfix

updated configurenotify

minor bugfix in applyrules

added sample of {grow,shrink}master to config.def.h

renamed c->border into c->bw, fixed monocle to subtract c->bw from each h/w value

geoms are now drawed in the status bar

removed the string-based setgeom approach, introduced a new Geom type instead and a helper macro

made the string-based setgeom working

some experimental state DO NOT USE THIS, I plan to have a nicer interface to change geometries

some changes towards 4.9

Added tag 4.8 for changeset 607015ddb091

removed the comment again

some polishing in tileh/tilev

minor bugfix

added a new comment

some cleanup

updated my geoms function

update

set layout already here

well I use a different default tile

added updatebarpos()

removed all defines of geoms, implemented setgeoms() instead, added config.anselm.h to show how I'd like to see that people edit their geoms

some polishing

updated dwm(1)

added bx, by, bw, wx, wy, ww, wh, mx, my, mw, mh, mox, moy, mow, moh, tx, ty, tw, th, wx, wy, ww, wh ad variables

changed config.def.h

new stuff

implemented setlayout in the way proposed on the ml, split tile() into two functions, a third will follow soon

allow for vstack

added some comments what the TODOs are for 4.8

changed config.def.h

integrated the new -x -y -w toggles of dmenu into my setup

removed some more useless clunk

implemented the stuff as I discussed on dwm@

fixed urgent hint handling

renamed maximise to monocle again.

fixed urgency hint, though Xinerama integration is still ongoing

next on TODO

renamed MAXLEN into MAXTAGLEN (backward compliance)

renamed monocle into maxmise, documented the keybindings in dwm(1)

monocle goes mainstream

fixed applyrules bug

removed View cruft, now back to the roots

made tag names snappier

fixed some issues nsz reported in IRC log

well, AIM_XINERAMA should not be enabled

made the basics of the tagging concept working -- if people want dynamic tags, that's even possible with this concept, the vtags[] array needs to be modified during runtime for this -- the new code is quite experimental, ugly and needs polishing

removed initags -- we autoselect the first tag in each view instead

a small fix to buttonpress

some more changes towards a better dwm

	proceeded, though we still miss a real Tag struct

disabled AIM_XINERAMA

simplified dwm

certain fixes, though still a lot of the mutex stuff missing

pushing my changes of tonight upstream (hg tip is NOW very UNSTABLE -- but those changes are necessary to get a decent multihead support) -- I renamed Monitor into View, to reflect in a better way the dwm terminology of the past

fixed Gottox' mail address

well typo fix

added Gottox to Copyright holders after all his contributions, applied his last patch

some other fixes, resize contains a minor bug

crash fix

made arrange() Monitor-dependent as well, fixed the movemouse/resizemouse issues

minor

applied Gottox' monitor.diff patch (thanks btw)

made restack, drawbar also Monitor-related only

changed arrange functions to contain the Monitor as first argument

applied Jukka's zoom-patch

resize handles offscreen issues

applied some necessary changes

removed Monitor->dc, unnecessary

removed Monitor->root, since we do not support classical multihead

some drawbar() polishing, and certain related fixes

urgency hook handling needs also to invert the square if present

Chris pointed me to the fact, that the window geoms calculation can be done once for each monitor, so I applied this remark

fixed missing else branch

applied dme's patch to prevent changing prevtags if nothing actually changed

applied dwm-4.8-snaptileds.diff

implemented urgent hint handling (with multihead support)

got initial Xinerama support working, though there is a lot work todo

some more polishing/consistency changes

proceeded with multihead/Xinerama support

removed maximize, there will be monocle soon

fixed wrong tagging stuff

added dmenu_run to config.h

root window and screen have to be in Monitor struct, we want side-by-side Xinerama and Multihead support (only Xinerama when enabled and present)

merged Christof Musik's Xinerama support patches, though this needs some polishing!

simplified Mod-m

implemented reapply for re-applying the tagging rules during runtime, Mod-r

next version will be 4.8

Added tag 4.7 for changeset d6d3085307d8

fixed focus steeling bug done by clients like opera

removed support for the NetSupportingWmCheck stuff, netbeans, argouml and others also don't work with compiz, so it is Suns problem to provide a fix

yeah compiz didn't did the trick, but identifying dwm as LookingGlass (LG3D)

foo

hack that adds NET_SUPPORTING_WM_CHECK handling, dwm identifies itself as compiz, hence I believe this might workaround the JDK 1.6+ XToolkit bug

applied Ritesh's patch to stext handling with some minor modifications

initialize prevtags in setup, now users can setup which seltags should be selectedin config.h

Using a new tags definition (const char [][MAXTAGLEN] - thanks go to Szabolcs!

moved LENGTH to dwm.c, moved prevtags to dwm.c

we check variable == value, and not the other way - the other way is for beginner programmers.

removed a misleading comment about client title windows, which don't exist anymore

full names in -v output of dwm

doing it in a shorter way

removed fgets usage, increment offset until a line is read, dwm will drop all lines read in one call, except the first!!! one (previously it preferred the last) - but the current approach is simplier and works better for general purpose in conjunction with the offset handling

simplified

made error handling more proper

replaced low-level stext reading with an fgets call

revival of RESIZEHINTS

sanders patch for b2 toggle

replaced Nmacros with LENGTH(x) macro

replaced ISTILE with domwfact/dozoom bools, removed nrules, nlayouts and ltidx, added NRULES, NLAYOUTS and Layout *layout as alternatives, removed isarrange(), checking against layout->arrange instead.

some sanity changes

going toward 4.7

Added tag 4.6 for changeset bcd7e18e196a

fixed a comment

also consider width for tile fallback enforcing

removed RESIZEHINTS and enhanced tile for fixed or aspect-ratio'ed clients

some cleanup, removed ntags variable, defined NTAGS macro, simplified tag(), view() and idxoftag(), fixed some NULL comparisions

parenthized use of ISTILE macro in dwm.c

thx to Toni Lainson

fixed two comments

removed dwm.h, just include C-files in config.h if you extend dwm, that's simplier and most flexible than all other possibilities

small cosmetic fix

uncommented DEBUG CFLGAS/LDFLAGS

just making dwm.h saner

applied Eric Mertens patch to mainstream dwm, however this needs testing

cleaned up dwm.c/dwm.h somewhat that it allows easier integration of patches

small comment

recreated dwm.h

added Mod1-Tab description to dwm.1

added antoszka's viewprev patch with some minor modifications, restored Client->tags as Bool *, however kept the static initialization of ntags and seltags (prevtags) - this seems to be the best compromise

applied Brandon MacDone's static initialization patch for seltags and Client->tags

removing NULL-terminating **tags definition in config.h

removed two spaces reported by Soleen

reverted Peters patch to tile, I will discuss the reasons at dwm@

s/xterm/uxterm/

added hint for downloading dextra

fixed font definition

fixed man page

fixed an issue in Peter's patch (it is no good idea to restack() all clients on enternotify()

applied Peter Hartlich's border collapse patch

improved tile() for the RESIZEHINTS == True case, now more space is consumed by the clients (esp. if those clients use increment handling heavily)

fixed colors

fixed inclusion order

updated with my favorite colorscheme

applied colors depend from lavishs proposal

switching to white normal bg, renaming tag 9 into www, for static use in conjunction with ff

btw 4.6 will be the next release

renamed config.h into config.def.h, config.h will be created if not present, this seems less annoying after all

introduced ISTILE, which can be easily extended with other layout functions to allow reuse of setmwfact() and zoom()

setmwfact and zoom check isarrange(floating) now, not !isarrange(tile) - this makes it easier to play well with bstack and nmtile patches

Added tag 4.5 for changeset 2acc60d6dfe2

cosmetic fix

fixed a comment

applied Peters patch, applied yiyus hint to initfont

applied Peter Hartlich's togglemax patch to allow toggling tiled clients to maximum

made all stuff non-static - so you can choose wether to use dwm the static or the extern way when extending it

there might be envs which need Xlib.h

Mod1-Button2 on a floating but not-fixed client will make it tiled again

now tiled windows can be resized/moved, their floating state will be toggled implicitely

some more rearrangements

ordered all functions alphabetically

corrected a misleading comment

macros which have been defined in config.h can only be used at function level, however you can nest code into config.h now for implementing a different layout (just for example), eg. #include "supertile.c"

ordered function forward definitions

backporting my intro-comment of old dwm.h

removed grabkeys, not necessary

micromizing dwm step 1

new colorscheme (16-bit compliant)

moved bar-related stuff to bar.c (merged draw.c into that)

renamed drawstatus into drawbar

renamed config.default.h into config.h

dist target only needs to add config.default.h

removed config.arg.h, only 1 config.h necessary

small changes of the colors

other color

using light colorscheme, preparing merge of config.arg.h with config.default.h

I work with enabled RESIZEHINTS, simply because I force myself to continue the st development ;)

introduced new define RESIZEHINTS, which allows to enable/disable size hint handling in tiled resizals

14px fonts drives me nuts

Added tag 4.4.1 for changeset 7c117df5d202

prepared 4.4.1 bugfix and minor feature enhancement release

Added tag 4.4 for changeset 408014d21261

removed unnecessary include, prepared dwm-4.4

setlayout should perform strcmp's if arg != NULL, because Layout is local to screen.o

reverted release CFLAGs

removed the _DWM_PROPERTIES handling, reverted ban/unban to XMoveWindow(), and changed argument of setlayout to layout[N].symbol check

renamed char prop[] into buf[]

fixed misappearance of iconified windows on SIGKILL

moved updatebarpos to screen

added screen.c, removed layout.c and tag.c

prepared merging layout.c and tag.c into screen.c

small fix

hmm I doubt the usefulness of storing this information...

cleaned up settags-handling

applied Gottox mwfact patch

replaced static Layout *lt with static unsigned int sellayout... (will be adapted later when _DWM_CONFIG is serialized as root window property)

renamed seltag into seltags

applied Jukka Salmi's setmwfact patch

made plural arrays

fixed the issue observed by various people, that clients appeared on empty tags

made tag/view/toggle{tag,view} work on pointer to tags-array, there was the need to define Key key[] not static to do this. split focusclient into focusnext/prev, fixed config.*.h's

applied Jukka's patch

fixed _DWM_CONFIG persistation, fixed the client disapperance bug during restarts

fififi

fix of resize (thanks Sander for the hint!)

fixed a typo

small bugfix

tags should be persistent now during X server run

small bugfix

made tile simplier

made Layout a static struct in layout.c, added some convenience getters in layout.c, now lt->arrange accesses are not possible anymore, arrange() is the super-arrange function which sets up all layouts

made resize more precise

moved floating to layout.c, kept tile.c outside

renamed MASTER into MWFACT, master into mwfact, and incmaster into addtomwfact

fixed a comment

separated layout-specific stuff into separate .h and .c files which are included in config.h resp. config.mk - this allows writing layouts for dwm without any need to patch existing code

let not overlap the borders

removed VRATIO, NMASTER, inc*(), renamed HRATIO into MASTER, see mailinglist for details

slight change of my color scheme

removed spow(x, 0); calls, I did them for consistency's sake, but it should be rather obvious how the scaling works anyways

applied Jukka's dwm.1 patch

changed shortcuts as described on the mailinglist, added sanity checks for ratios during tile (fallback to wah if clients would get too small), documented that new stuff/shortcuts in dwm(1)

small fix of static function order

I introduced {H,V}RATIO and inc{h,v,}ratio() functions - the default behaves like in dwm-4.3, config.arg.h shows how I prefer the ratio being handled (for the future I plan to change const char *arg into ..., and renaming Client into Win.)

removed a C++ style comment

implemented ratio tile as described on the mailinglist

changed the fix of yesterday, the resize should only be necessary at manage() time, not on any unban()

foo

set client state to iconic on maprequest

changed ban/unban implementation to not move the windows anymore, but map/unmap them instead - PLEASE TEST THIS

removed shiftview(), if you scroll the views in the tag area, you can also use Button1 instead
removed focusclient([-]1) bound to scroll wheel - if you slightly move the mouse you will focus the first client anways which is in the way

applied Jeroen's {clean,spell}.diff patches, thanks Jeroen!

applied Jeroen Schot's shiftview patch

applied dwm-windowscrolling patch

Added tag 4.3 for changeset e0ec0d5d8b1e

restoring tip to be a working dwm again (switching FONT to terminus in config.arg.h)

please recheck this patch with Mathematica from Wolfram Research - lemme know if it fixes the geyish blob window bug

removed crappy createnotify handler

applied restack patch of anydot, with slight changes

I change the style of init somewhat, init as early as possible.

uncommented creatnotify

added an creatnotify event handler

applied anudots [un]ban repair patch

making variable declarations in dwm.h extern, this seems to be more ansi compliant

replaced BORDERPX with sel->border in togglemax(), in other places this is not possible.

Added tag 4.2 for changeset c13cb8c6b7a5

referred to LICENSE file

Jukka also belongs to Copyright holders after all he has contributed and done for dwm, updated -v as well

applied Jukkas patch

changed -v output (now also anydot and nsz are contained in this output)

added nsz to copyright holders as well, because he did a lot recently

added anydot to Copyright holders, because he contributed a lot recently

applied anydot's 3 minor patches, thank you anydot

applied Sanders patch

applied anydots proposal to let togglefloating restore fixed windows

fix if n < nmaster of remainer calculation for master windows

rev 900, cool

fo

calculating the remainder for master and stack area correctly

fixed issue reported by Christian Garbs

removed the only one and superfloues strncmp

foo

reverted last change after Sander pointed out the original decision

applied anydots remark of togglefloating()

applied Szabolcs proposal for zoom() as well

applied Szabolcs proposal to simplify setlayout()

Added tag 4.1 for changeset 00f4180df72b

no that change breaks fullscreen apps

raise barwin in restack, that's the most elegant solution I think

simplification

raising the barwin has no effect becasue of restack

fix

barwindow raising

using BarTop as fallback if BARPOS is set to BarOff as default for toggling

removed strip, added -s to LDFLAGS

added new Mod1-b functionality to dwm(1)

another fix, call lt->arrange() in togglebar only

fixed bpos init

made bar togglalble

thanks to Jukka

added the GTK Save-As bug report to BUGS section of dwm(1)

applied dfenze cleanups, fixed some comments in dwm.h

small fix of fix

fixed a potential security flaw

applied Maarten Maathuis recenter-patch for floating clients only requesting new width and height exceeding the screen space

s/remain/remainder/

fixing remaining space calculation

thanks to Juka to check -Wall ;)

next version is 4.1

applied patch of Paul Liu to allow onthefly resizing due to xrandr changes

Added tag 4.0 for changeset 018c38468422

fixed the border issue for mplayer, ff is definately broken when using F11 (fullscreen mode)

touch border

set border at manage time

using pixelcarnage-monospace (proggyclean), because this is better to the eyes

changed border handling

changing order of c->border restorage

I used 2006 in other places as well

yet another fix of copyright compactisition

making Copyright notices more compact

next version will be 4.0, so don't expect it within the next days

make also transients floating when we do not know the main window

Added tag 3.9 for changeset 55478328b242

next version will be 3.9, but don't expect it this week

add an additional check in resize() to prevent a crash of dwm

Added tag 3.8 for changeset 2ea201354cf0

we don't need to set the font all the time

some changes to updatesizehints, I don't change the aspect ratio algorithm now - I can't think, it is a mess

removed sendevent

Escape -s in dwm.1

Added tag 3.7 for changeset baee494346e5

prepared 3.7, ready to rambo

removed an unnecessary newline

removed superfluous externs (except for tags, because tags is defined in the source)

replacing Mod1-i with Mod1-Shift-j, Mod1-d with Mod1-Shift-k

oops, small bugfix in my config

default masterwidth should also be at 600

changed order if h/l

well I agree to several people claiming h/j/k/l is the better default than Tab/S-Tab/g/s for focus and master resizing

renamed untiled into floating, keeping tiled instead of tiling (afaik tiled sounds more correct) - English speakers convinced me

hahaha, untiled and non-untiled sounded really cumbersome

made Fnt an anonymous inner struct

made Fnt an anonymous inner struct

nah, I don't want 640 as MASTERWIDTH

using MASTERWIDTH=640 and 32px steps by default (incmasterw())

simplified focusclient()

renamed versatile into untiled

merged focus{prev.next} into focusclient(1/-1)

made dwm.h more tidy (thx Jukka for the zoom() hint)

oops

small bugfix

fixed some issues due to the Arg->const char * transition

restoring default keybindings as Sander complained

replaced Arg union with const char *arg, seems cleaner to me, even if we need atoi() in some places

re-added xterm to config.default.h

removed button4/5-bindings for incnmaster on mode label - that is misleading

fixed order of key bindings described in dwm.1

status needs update even in togglemax() - since we got an indicator for this

reusing drawsquare for client title, empty square before title means versatile window, filled square before title means versatile maximized window.

dwm draws a small caret before the client title if it's a versatile client

nah grouped keybindings by context

fixed exit condition in togglemax()

fixing missing extern declars in dwm.h for {de,at}tach()

several changes, made togglemax extern and separated it from zoom() - moved zoom() and togglemax() into layout.c, changed void (*func)(Arg *) into void (*func)(Arg), changed default keybindings of focusnext/focusprev and incmasterw to h/j/k/l accordingly, made keys in config*h appear alphabetically (special keys first), renamed resizemaster into incmasterw, renamed MASTER into MASTERWIDTH

just ignore the FD_ISSET check in main.c of xfd, just call XPending (which does the same afair)

optimizing background color of terminals

using 333 as background, hey proggyclean is really good

switching to 13pt proggyclean (this fits better my 1920x1200 resolution than 12pt terminus which is too small or 14pt terminus which is too fat)

using smaller font

using a green selborder

reverting to bg 555

s/555/357/ for SELCOLOR

switching to urxvtcd for the moment (the uxterm flicker makes me headaches, st is really highest prio now)

Added tag 3.6.1 for changeset 20ec6976cee1

creating 3.6.1

reverted accidental removal of sx, sy

removed some global vars which should be static instead

moved focus{next,prev} and nexttiled from client.c to layout.c (because those are not client-specific), moved toggleversatile() from layout.c to client.c (because those are client-specific)

cleaned up and commented the config.*.h

Added tag 3.6 for changeset 75b1b25fe0d7e29400baf30568153f668324928b

fixed quoting and a comment

renamed setfont to initfont, getcolor to initcolor

fixed order

added draw.c again (except getcolor and setfont which are helpers in main.c)

separating drawsquare from drawtext, made drawtext extern

rechecked with OpenBSD

split screen.c into layout.c and tag.c (because the view is an implicit mixture of both)

using eprint instead of fputs

replaced togglelayout with setlayout

changed some odering in config*.h

renamed Client->versatile and Rule->versatile into Client->isversatile resp. Rule->isversatile

renames swim[ming] into versatile

renamed floating into swimming (this does not clash with C naming conventions and fits better the fish symbol) - also in man page

removed debug compile flags

introduced Layout struct

some more code polishing

forgot an extern declaration

bugfix of transient handling

removed procevent, more refactoring

draw.c is useless (belongs to main.c now)

renamed view.c into screen.c

renamed manage.c to view.c

renamed Rule members, renamed RReg into Regexps

merged tag.c, view.c and tile.c to manage.c

some more refactoring

added some new convenience functions

some more refactoring

introduced tile.c, some refactoring of functions

simplified configurerequest to a bare minimum, removed wrong ban() calls

fixed configurerequest according to Jukkas complains

fixed configurerequest according to the problem Jukka reported

don't resize master if not in tiled mode

removed useless space

well, resize should be called in dofloat anyways ;)

removed ugly ban(), extended resize() that it only resets the size if necessary, added border_width commit to manage()

removed getnext/getprev, redundant

added ban() which takes care than a banned window is not banned again... (this reduces the overall ConfigureNotify's to clients)

Added tag 3.5 for changeset 63ad05e7f9e1

simplifying regexps in config samples

made for/if/else constructs more consistent, some code polishing

made configurerequest more tidy

I didn't knew of c->isfixed, that should fix Jukkas issue with gkrellm ;)

making it more sure that transient checks will work in any case

stupid me, one needs to set c->{x,y,w,h} in configurerequest obviously ;)

simplified configurerequest

yet another consistency fix of dwm

I don't see any reason why we should select for input on override-redirect windows?

bugfix of bugfix (thx to Sander)

fixed a bug

during writing 2wm.1 I noticed some bugs in dwm.1

removing all EnterNotifys after resize

added pointer warp on drop in resize

pneubeck pointed out an obvious thing, that a second p = stext + strlen(stext) - 1 is unnecessary

treating acroread and mplayer floating by default

adapting John Grahor's patch to dwm-3.5

fixed copyright notice in Makefile

fixed order of Bool declarations

added TOPBAR option for making the bar appear what the user likes

next version is 3.5

making the  bar appear at bottom

Added tag 3.4 for changeset 9ede7b2d2450

removed debug flags

replaced getproto with a saner function, now old-school artifacts of WM times in the early 90s completely disappeared, no punned pointer warning anymore ;)

yet another fix

simplification

implemented aspect ratio support of windows

made some changes more concistent

got rid of LD (inspired by JGs patch to wmii)

applied apm's patch proposal, getting rid of XDrawLines

Added tag 3.3 for changeset 0f91934037b0

applied Sander's drop_bh patch

handling WM_STATE seems to make DnD in gtk/qt apps working, well let's handle this in dwm as well

regarding to http://plan9.bell-labs.com/sources/contrib/rsc/man.ps the BUGS section should appear after SEE ALSO section.

renamed CAVEATS into BUGS

Added tag 3.2.2 for changeset d3876aa79292

prepared yet another hotfix release

Added tag 3.2.1 for changeset f2cabc83a18f

hotfix release 3.2.1

applied offscreen appearance hotfix

Added tag 3.2 for changeset 4ce65f61f01b

renamed activescreen into selscreen

implemented Sanders remarks

small changes

removed a blank line

this version should also work with cornercases (like unmanage during !issel, etc.)

I think this is the best solution of multihead support

darker border

this variant is known to work, but focus() is ugly - we need in general a better way to handle multihead, this issel-stuff looks awkward (maybe it might be a good idea to set sel to NULL but to introduce a Client *revert which is set if a screen is unfocused, have to think about it further).

applied Sanders all5.patch (thanks for your weekend session, Sander!)

leavenotify also don't needs the check

deciding for focus(NULL);

replaced XSetBorder.../focus()

yet another multihead fix by Christof Musik

I prefer BORDERPX=1

personally I prefer 2px borders

applied a modified version of Christof Musik's multihead patch (though this is not sure if it works in all cases, have to wait for an ACK by Christof)

moved BORDERPX to config.*.h

Added tag 3.1 for changeset e1c8bef05e6e

removed erealloc (not used)

removed unnecessary bx, by, bw variables

s/444/555/g - enlightened selected background

removed firefox rule from default.h

applied new default colors

applied sanders patch of my config

Jukka, thanks for being faster than me!

changed restack, to fix undrawed tag indicators

removed drawclient and drawall (they performed useless operations/consumed useless cpu cycles)

implemented new color scheme accordingly to Sanders proposal

removed mode label stuff

removed client title bar

added missing space

nah the last change feels bad

changed incnmaster shortcuts in config.arg.h

Added tag 3.0 for changeset 76b58d21ea98

removing to allow nmaster=0

this looks better now ;)

switching to arabic tags

small fix of initial numlockmask value

allowing nmaster=0 (I think that's a straight idea)

added Button{4,5} description to dwm.1

added comment to %u in config.default.h, added Button{4.5} support on mode label

reorganized

implemented nmaster appearance in mode label (using %u)

Added tag 2.9 for changeset 3a5910fac3cc

I received Sander's complain just in time ;)

changed status fg

fofofo

extended default tags to 1-9

small grammar fix of dwm.1

changed a term in dwm.1

using shorter tags, using mon instead of cert

roman numerals suck badly, reverting

switching to roman tags, I need more tags

I inc/decrease on the fly, feels better

allowing swap() for first master client

prevent pop() if first sel == nexttiled(clients)

removed swap(), implemented pop for everything

preserving c1->snext and c2->snext in swap() as well.

fixing some minor issues

removed the term clients from dwm.1

adding some prevention that master clients get smaller than bh

changed shortcuts to M-i and M-d

allowing zoom within master area as well

changed plus key to MODKEY-Shift-plus (US-layout restrictions ;)

I prefer two master windows by default

added MODKEY-{plus,minus} shortcuts (increasing/decreasing master clients)

experimental version which allows master clients being increased/decreased

fixed comment of drawclient()

switching bakc to my previous color favorit

renamed drawtitle into drawclient

using more thinkpad compliant colors

changed arg's color scheme

corrections

next version will contain updated copyright notice

Added tag 2.8 for changeset 107719a9ce3bd0c79f9f1f626596eb338a276561

fixed cleanup, using clients instead of sel

changed comment of updatesizehints

renamed updatesize into updatesizehints (thx to Sander for this hint)

small simplification to dotile() (thx to Ross for this hint)

migrated arg's config.h to new dmenu command line options

added Solaris hints

Added tag 2.7 for changeset 21951c0dfbae5af68ed77821a4d87253ee91803f

applied Sanders patch

fixed man page, status bar displays EOF instead of "broken pipe"

added the java odyssee to CAVEATS section

removed Client->grav

removed gravitate for the moment

disabling configure() during resize

made gravitate effectless, waiting for complains ;)

fixed diagnostic error message

removed the hardcoded fixed fallback, it is useless in non-Latin1 environments

we don't use the term clients anymore, windows is the better term

Added tag 2.6 for changeset 5308dd22b6ee8e3218c81d9e7e4125f235bb5778

found compromise for fonts

nah, people should define 'fixed' to be compatible with their locale

using a UTF-8 capable fixed font

switching to uxterm again, I get kretze from this urxvt crap

with this patch everything works fine for me

don't use Xlocale crap, let's use locale.h instead

using the portable Xmb+UTF-8 way of life, will see if this works well...

improved the memory leak prevention

making terminus more explicit a regular font (otherwise olique font might be used in some setups like mine).

also setting LC_CTYPE only

enforcing using fontsets even if they are incomplete for some encodings

Added tag 2.5.1 for changeset c7f84f23ec5aef29988dcdc4ec22a7352ee8f58e

hotfix of a serious crashing bug

Added tag 2.5 for changeset dcbbfabc8ecc5f33a6cc950584de87da1a368045

applied Manuel's regex patch

made squares equally sized, and 1px bigger

if client is focused, the emptysquare don't needs to be drawed

and the winner is

sander1 proposal

dood

darkening the fg colors somewhat

less obtrusive indicator (only a top line)

final2 indicator

removing the -1 on final indicators, looks cleaner

final indicators

changed things like I described in last mail

improved tag indicator

removed viewall signature from dwm.h

next version will be 2.5

removed viewall(), replaced with view(-1); added tag(-1) to tag a client with all tags (new key combo MODKEY-Shift-0)

Added tag 2.4 for changeset 32b246925086910d63147483160281a91a47479f

extended dwm.1 with last change

togglefloat should only work in dotile mode (thanks to Sander for this hint)

applied Alex Elide's tricky patch, thanks!

returning to old Key struct

applied Szabolcs Nagy's patch (thank you!)

applied patch by Jukka

changing Key.func into Key.func[NFUNCS], this allows sequences execution of functions per keypress (avoids implementing useless masterfunctions which call atomic ones)

added man page entry

added togglefloat to hg tip (i consider this useful for some cases), using MODKEY-Shift-space as shortcut

applied Jukka's patch preventing some cornercases and making the EOF error message correct

applied yet another proposal of Manuel

applied Jukka's sizeof K&R compliance patch, applied Manuels' last-line printage proposal for stdin reading.

applied Jukka's stdinread patch

next version will be 2.4

applied Manuels patch (thanks to Manuel!)

small fix of man page

Added tag 2.3 for changeset 719b37b37b0df829d7cf017ac70e353088fe5849

using the term 'indicated' instead of 'higlighted'

updated man page with the bottom right corner indicator

fififif

added a similiar patch to Daves solution to indicate if there are clients with a specific tag

returning to old bar colorization behavior, like sander proposed for consistency reasons

applied Gottox patches

using iso10646 explicitely in arg's config.h

Added tag 2.2 for changeset 7e92f58754ae6edb3225f26d754bd89c1ff458cf

nah reverting to my prev style, that's really the best

applying aluminium style for arg's config.h

applied Gottox' windowarea patch

using a more blue-ish color...

making the selected color more lightning

next release will be 2.2

applied Jukkas patch

Added tag 2.1 for changeset a2c465098a3b972bbed00feda9804b6aae1e9531

renamed resizecol into resizemaster

applied Gottox patch to simplify the resizing of col, instead of resizing the current area, it only resizes the master area in the future (seems more predictable)

Added tag 2.0 for changeset 12deea36603da407e3f32640048846a3bd74a9ec

make sure that changing sx has no impact on snapping

in a 1920x1200 setup 40 pixels of snap value are much better than 20

fixed stupid bug of snap-to-screen

changing snap priority

removed useless abs() calls

added screen-border snapping in floating mode, feels quite well

and another fix

applied sanders try2 patch

applied sanders max size fix

stupid urxvt needs bg to highlight selections in a sane way, though that makes sense to some extend

reverting to original

applied sander's config.*h nitpick patch

some other change

apply small fix to prevent apps like mplayer wandering when toggling fullscreen

forgot to use -tr, which actually prevents the ugly flicker (using xsetroot -solid black as root window pixmap to make this work nicely)

using MASTER 600 again, it is definately better, and using urxvtc for the moment (it doesn't flickers on refreshes, but this is not because of Marc Lehmann, it is because of the original rxvt code)

changing MASTER in config.arg.h from 600 to 550 per thousand

my new 1920x1200 Z61p arrived, now I can use terminus in a sane way...

moved MOUSEMASK into event.c (not used in other places)

now being at v2.0

using lsx instead of Jukka's shell construct

hotfix

Added tag 1.9 for changeset a5567a0d30112822db2627a04a2e7aa3b6c38148