Add include of cairo.h when using librsvg

This should be compatible with both old and new librsvg versions.

Update version to 3.6

Add missing file so make distcheck passes

Merge branch 'master' of git://git.openbox.org/dana/openbox

Update the gnome-session support for version 3.9.90 in Ubuntu Trusty

This change is not backward compatible unfortunately. The format of
the gnome-session .session files has changed, such that every part
of the session is now listed on the RequiredComponents line.

Special case for unusual monitor setup

When no monitor is at (0,0), the screen rect containing all monitors
is not at the origin, but screen_apply_gravity_point assumes the passed
x/y are relative to the passed area, so adjust them.

Update CHANGELOG again.

Update Norwegian translation.

Clarify versioning comment instructions in configure.ac

Clean up translation files

Correct obvious and remove other fuzzy translations
Update line numbers
Fill out Language: header correctly (just the filename without the .po)

Add Galician translation.

Update Polish translation.

Move some more code around to make it shorter

Reduce length of theme.c a bit

Make theme.c indentation more locally consistent

Also removes some comments that is just the code with s/_/ /g

Remove wrong code from button mask loading code, bug 6065

The READ_MASK_COPY macro already uses these exact fallbacks, so these
lines just overwrote the custom bitmaps with the default fallback bitmaps.

Only add xqueue listener fd once, bug 6326

Allow selecting menu entries during menuHideDelay if you click again, bug 5501

Move final frame adjust in flashing to timeout func

This possibly avoids badness when the timeout destroy notify is removed
from the frame release func.

Make sure to reset all the GSource timer ids

Avoids warnings like 'Source ID 8382 was not found when attempting to
remove it'. In particular some removals were missing in menuframe.c
resulting in a warning being printed every time a submenu was opened.

Update my email address

Fix some compiler warnings for ShowMenu

Some fixes found by coverity for things used before a null check

In the second case, we actually check the wrong variable.

Initialize some stuff to 0 before taking their address

Move assignment to before break, hopefully this doesn't break stuff

Fix an if statement getting moved into a loop it shouldn't have

Found by coverity code analysis.

Use PANGO_UNDERLINE_SINGLE for rendering shortcuts

The pango docs say to use _LOW for accelerators, but that causes the line
to be drawn beneath the ink extents which for us may be outside the area
that gets displayed and we end up with no line at all. We also don't want
to extend every menu entry with enough space to fit this possible line.

Fix compile failure when X_HAVE_UTF8_STRING is not defined

Use GString for appending strings

It can take a very long time to print a _NET_WM_ICON property for firefox with the previous code, several minutes. Fixes bug #6042 and also implements the change for append_string.

Fix use after free in menu_entry_frame_free

Fix copy paste typo in swap_byte_order

Update CHANGELOG with changes so far

Make desktop actions noninteractive

Remove freshmeat related stuff from release script, the site is dead

Clip client rect to monitor bounds while picking a corner for resizing

This way you can actually do everything to it when it's offscreen,
not just resize the one edge that's poking out

Change some while loops to for loops in config.c

Allow specifying several mousebinds in one go, like for keybinds

Also fixes minor indent error

Make RaiseLower work for the dock as well

As a side effect, _NET_RESTACK_REQUEST now also allows specifying the dock as the sibling.

Add keyboard/rebindOnMappingNotify option

I have to switch to a qwerty layout to play steam games, and every time
I switch back and forth, we lose some random bindings so I also had to
reconfigure Openbox every time, now I don't.

Pass the GravityPoint as const* instead of by value

Let the menu placement code know if the position was user specified

This way, it can avoid running the code that moves it to a good position
relative to a mouse that isn't there

Move common gravity application to screen.c

Resolve gravity coords in menuframe so it can use the widths

Add <position> tag to ShowMenu action

Fix negative relative coordinate placement

double space

Fix thinko in if.c, *o in Stop can never be pointing at the right place.

Use a file static boolean instead.

Fix typo in variable name

Avoid initializing a VLA of 0 length

Declare variables at start of blocks in place.c

Break loop if is_true was set to false in If

Don't segfault when there's no target in If

Some fixes in french translation

Cache results of offset/shift functions so we don't do 6 function calls per pixel when reducing color depth

Fix gcc warnings in screen.c

Added missing break statement (Bug 6043)

Add a strict option to the ToggleShowDesktop action

When the strict option is used, normal windows are not able to show themselves
while showing the desktop.

Portuguese update for openbox

Allow FillToEdge to grow when all its edges are blocked.

This change points out how useless the return value from client_find_resize_directional()
was, so it also removes that.

When all edges are blocked, the FillToEdge action will try to grow all four edges agin
without blocking any of them on their current edge. This more closely matches the
behaviour of the GrowToEdge action.

Clean up the FillToEdge action implementation

This extend the client_find_resize_directional() method to support
two growing modes, and return a bool for whether it was able to grow/shrink.

The client_find_resize_directional() method now takes an enum instead
of a bool, with two growing modes. The old mode which always tries to
grow, the a new mode that will only grow if the client's edge is not
already at a grow stopping point (ie against the edge of another window).

Created a 'GrowToFill' action. (Bug 3356)

Use the BSEARCH() macro in overlap placement

Currently the code rolls its own binary search, but now that we have
a well-tested binary search implementation in obt/ we can make use
of that.

Fix bugs and add unit tests for BSEARCH()

Also adds running unit tests to the release/go script so we can be
sure that all tests pass before we make a release.

Fix nits for overlap placement center option

Add the old <center> option for the placement policy. (Bug 5946)

Original commit messages:
. Reformat to move closer to house style
. Add center on top of leat overlap place algo
. Add sentinel value to edge arrays
. Use a Size instead of a Rect for a centering field
. Fix off by one bug
. Need to declare dx and dy
. Pass length of edge array instead of recomputing
. Fix missing open-brace in config.c
. Address the more trivial subset of danakj comments
. Revert "Remove now-unused config_place_center option."
  This reverts commit 5e282dae08be3b900e0337efa0fae8f3ffa92cd7.
. Remove reliance on sentinel value when scanning edge arrays
. Avoid need to initialize Size structure by removing it :)
. Clean up field expansion code somewhat
. Compress code further by using a structure for common args
. Fix search for next grid point
. Squeeze it even more by not using Size at all

Fix gcc warnings

Fix a typo causing submenuShowDelay to be ignored (Bug 6019)

Add window type to If as well

This one required some extra helpers so is separate commit

Allow matching class, name and role in If action

Add ForEach action which is like If but runs on all clients

Also adds a Stop action that lets you stop running, in case you only
want to run actions on the first match.

Remove the else-on-newline rule from HACKING

Add queries to If actions

This allows the If action to run queries against a client other than
the target of the actions being run, for example to check state on
the focused window while performing actions on another window during
focus cycling.

The syntax looks like

<action name="If">
  <query target="default">
    <title>FooBar</title>
    <maximized>yes</maximized>
  </query>
  <query target="focus">
    <desktop>3</desktop>
  </query>
  <then>
   <action name="NextDesktop"/>
  </then>
</action>

The above checks the client window that the actions will run on to
verify that its title is "FooBar" and that it is maximized. If that
is true, it also checks that the currently focused client window is
on desktop 3. If that is true also, then it runs the NextDesktop
action.

The target="" option can be set to "default" which uses the client
window that the actions will run on, or it can be "focus" which uses
the client window that is currently focused.

The <query> tag is optional, and the conditions inside the query can
be placed directly inside the If <action> tag, as they were before
this change. In that case, a default <query> tag is assumed with
target="default" which matches the previous behaviour.

Multiple <query> tags can be present, and they must all be true in
order to run the actions in <then>. If any one is false, the actions
in <else> will be run instead.

Update HACKING for brace rules

The rules in HACKING for braces are kind of unclear, so add some
more rules and provide examples too.

Add a note to add the version to the bug tracker for new releases.

Also important to update IRC and keep hydrated.

Update po files for 3.5.2

Add CHANGELOG for 3.5.2

Update version in configure.ac to 3.5.2

No library interfaces were changed since 3.5.1.

Maximized undecored windows have incorrect top border. (Bug 5996)

If the window has no border since the theme has bwidth=0, or because the
window is truly undecorated via MWM hints, then we should not allocate space
for the frame of the window which will just be empty pixels.

window.inactive.label.text.color should default to white.

Currently it defaults to itself, which is NULL, and causes a crash.

The spec http://openbox.org/wiki/Help:Themes#window.inactive.label.text.color
claims that the default is white.

In 98b02c6b60bbde2a5db026b3ee3e6e6dc44d1a92 we had this property being read
twice, once with the NULL backup and once with white. It seems that we kept
the wrong one, so just switching this to white now.

Fix a write out of bounds in splitvertical gradients (Bug 3612)

If the gradient has height 1, then y1sz is 0. We don't want to use the
first color and move the data pointer, since this will move it past the
end of the array.

We use congfigure.ac not configure.am

Fix some compiler warnings

Void function should actually be declared void. Not-reached switch
condition should still return something.

Update line numbers in po files

Change "plain" to "exact" for If's <title>

Update CHANGELOG

Add matching which monitor the client is on in the If action (Bug 5426)

Use enum value for monitor in MoveToCenter

Let the If action match window titles with GRegex

Use a helper function to make if.c a bit shorter

Add activedesktop to If

Dock used showDelay for both hiding and showing (Bug 5811)

Allow specifying only one of width and height in per-app settings size

Fixes bug 5819.

Fix use after free in error message

Add some comments in MoveResizeTo code

Remove fuzzy and outdated translations, add plural expressions.

Clean up the .po files for this release. Plural expressions are needed in
all files that have plurals. And replace fuzzy (incorrect) translations
with empty ones instead.

Make nplurals=2 in Turkish translation since there are plurals translations

The plural should be a c expression that says if n is plural or not.

Remove po_PL from LINGUAS as that doesn't exist, should have be pl.po

Fix format specifications in Interlingua translation

3.5.1 changelog

New Polish translation was an update to existing one.

Move the po_PL.po file to pl.po as it should have been named.

Remove Imlib2 mention from po files

Remove deprecated and unneeded g_type_init call.

Remove unneeded obrender/button.h

Update version to 3.5.1.

obrender broke backward compatibility by changing the layout of the
theme structure.

obt broke backward compatibility by changing the method signature of
some keyboard methods.

Both libraries added some new methods as well.

Move shortcuts to latin characters in Greek translation

The _ keyboard shortcut in menus only works on latin characters.

Move shortcuts to latin characters in Hebrew translation

The _ keyboard shortcut in menus only works on latin characters.

Updated Italian translation

Updated Serbian translation

Updated Spanish translation

Updated Czech translation

Update Hungarian translation

Add Afrikaans translation

Add Romanian translation

Updated Turkish translation

Updated German translation

Updated Arabic translation

Add Belarusian translation

Add Interlingua translation

Add Hebrew translation

Updated Dutch translation

Add Polish translation

Add Greek translation

Sort po/LINGUAS alphabetically

Updated Lithuanian translation

Change the message about no icons in menus.

The message used to mention Imlib2, but now there are two libraries
that can be used to support icons. So make the message more generic
and show it only when both libraries are not present.

Add support for loading SVG icons using librsvg.

This adds a configure option --disable-librsvg, but defaults to
using the library if it is present during configure.

When enabled, Openbox will attempt to load svg image files using
the library, similar to how Imlib2 is used for other image
formats.

Since librsvg uses the libXml2 library, their errors end up in
the same global namespace as Openbox config file parsing. To
avoid this, we reset the libXml current error whenever we start
loading a file, and save the last error that occurred when we
are finished, by storing the error in the ObtXmlInst.

Add autopoint to README.GIT list of packages

Fix a typo that prevented setting client size for the height only (???)

Convert two // to /*

Constrain the moveresize popup to the monitor which has the client

Rename bounds to monitor in place_overlap and move a const

Fix least overlap to fall back to the chosen monitor, not (0,0).

We forgot to count the dock in LeastOverlap

Remove some evil tabs

Refactor the per-app parsing code into a helper function and reduce nesting by 2 levels

Make MoveResizeTo work on the dimensions of the frame, not the client

Allow application rules to control window size (Fix bug 4661)

Use the following in your per-app rules:

<size>
  <width>A</width>
  <height>B</height>
</size>

A and B can be integer values to specify a size in pixels. They can also be
percentages or fractions to be relative to the size of the monitor the window
is placed on.

Remove now-unused config_place_center option.

The new LeastOverlap placement doesn't use this option, and it kind of defeats
the purpose of the algorithm.

When an undecorated window is maximized give it a 1px border at the top (Fix bug 5506)

We want to keep some way for a user to re-decorate the window with the default
bindings. However with a large border width, the top border is much bigger than
it needs to be! So make it fixed at 1px large in this case.

Allow non-interactive focus cycling.

Adds an <interactive>bool</interactive> option to the NextWindow and
PreviousWindow actions. When it is false, the action is not interactive and will
immediately switch focus to whatever the next focus target is.

Removing the "interactive" flag from the focus_cycle() method, as it was unused
previously, and the new code does not make use of it either. In order to be
non-interactive it simply starts a focus_cycle then immediately ends it when the
action ends.

The "interactive" flag in focus_cycle() forced a linear cycling order which may
not be what you want, so the new method is preferrable anyhow.

Don't let autostart checking for background-setting tools end up in logs (Fix bug 5186)

Add a $pointer execute variable substitution and make $wip and $pid consistent (Fix bug 5758)

$wip and $pid would output values for the current target window if there was
one, but output nothing at all if there wasn't making the output difficult to
parse. Changed to make these always output a 0 if there is no target.

Added a $pointer variable that is replaced with the x and y coordiates of the
pointer, separated by a space.

Inspired by the work of Denis Kaganovich <mahatma@eu.by>.

Remove unused variable

Fix duplicate const warning

Fix some namings and xml formatting in rc.xml and rc.xsd

When a window is the focus target and it is removed from the focus order, reset the focus cycling (Fix bug 5410)

We were only doing this if a dialog was showing. But if it's the target, then
a dialog is not relevant, we should always recompute the focus cycling. This
avoids a crash when not using a dialog and closing a window during cycling.

Fix negative ResizeRelative values with terminal windows (Fix bug 5228)

Inspired by patches from Brian Mock <mock.brian@gmail.com>.

Get the window's title in the fake-managed case since we apply app rule matching (Fix bug 5277)

Print error message to stderr (Fix bug 5731)

As stdout is redirected to /dev/null by some DMs, we need to print error
messages to stderr.

Address a couple bugs in rc.xsd (Fix bug 5746)

Make .desktop files obey specs where possible (Fix bug 5737)

Allow windows created by execute actions to steal focus if the user isn't interacting with another window (Fix bug 5419).

When the execute action was run, we would say that the user had used the focused
at that time. Then when a new window popped up, we'd think the user was busy in
the current window and prevent the new one from steal focus.

Now the execute action does not update the "user interacted with the focused
window" timestamp anymore. So, if they aren't currently typing in some window
when they trigger an execute action, and the window appears, it will steal
focus.

Allow window matching based on the group leader's name and class (Fix bug 5721)

In Openbox 3.4 we used the group leader's WM_CLASS value when it was available.
This prevents windows in the group from overriding with a specific value,
however which is bad. More rationale can be found in
http://icculus.org/pipermail/openbox/2010-September/006790.html

Some applications (eg. Firefox) use command line flags to set the WM_CLASS
property on the group leader but do not change the property on the mapped
windows themselves. This makes matching these windows not possible in Openbox
3.5.

We resolve this by exposing the group's WM_CLASS values alongside the individual
window's values. We add _OB_APP_GROUP_NAME and _OB_APP_GROUP_CLASS properties
along with "groupname" and "classname" attributes for the rc.xml application
tag.

Allow loading of menu files outside of your XDG_CONFIG_HOME (Fix bug 5711)

When the given file name can not be found in your XDG_CONFIG_HOME, ie in
 ~/.config/openbox, then try the file name directly.

This means if you specify a menu file such as "/home/dana/helloworld.xml",
openbox will try, in order:
1) ~/.config/openbox/home/dana/helloworld.xml
2) /home/dana/helloworld.xml

And it will load the file you meant when it tries the second one.

Make LeastOverlap placment replace the behaviour of Smart and address style nits

LeastOverlap placement option (Fix bug 5385)

Adds a new placement algorithm that finds a place on the monitor that overlaps
the least amount of windows as possible.

Original patch by Ian Zimmerman <itz@buug.org>.
Port to Openbox 3.5 by David Vogt <dv@adfinis.c>.

Use the nearest monitor when the search query rect does not intersect any monitor (Fix bug 5500)

Previously we would try to find the primary monitor and use that when the search
was outside any monitor. However, if the primary monitor is chosen by the mouse
position and the mouse is not inside any monitor, we enter infinite recursion
trying to find the primary monitor.

The nearest monitor is a better metric anyhow, and this ensures
screen_find_monitor() is never recursive as it always returns a value without
depending on other screen.c methods.

Don't edge-warp with the cursor based on a monitor that does not contain the pointer (Fix bug 4992)

Pick the monitor most relevant to a rectangle more cleverly.

When monitors overlap (this happens with cloning), we were choosing a monitor
to associate with a window, for maximization for example, somewhat arbitrarily.

Now we have a more clever algorithm that considers the configured primary
monitor first, and that does not prefer monitors based on their sizes, but only
how much of the window is in the monitor, excluding parts that were claimed
by another monitor already.

Theme button-specific values should fallback to the button general values consistently (Fix bug 5443)

Prompt should not always use the currently focused button as its result (Fix bug 5518)

Update manpage to current version. (Fix bug 5444)

Don't strip leading whitespace from menu labels (Fix bug 4782)

Create some Openbox release scripts

release/bugs: Prints a list of bugs that are mentioned in git commits for a
  git revision, since previous release.
  - Very useful for updating the CHANGELOG file!

release/go: Tests a git revision for correct compilation, and prepares files
  for release.
  - Makes the tarball
  - Makes a GPG signature for the tarball
  - Tags the release
  - Spits out URLs to edit and gives the changelog for copy/paste.

release/email: Sends an email to the Openbox mailing list with the changelog
  and details about the release.  Call this with the same parameters used for
  running release/go once it is finished, and the files are uploaded, etc.
  - Also emails mikachu re freshmeat.net

Allow client destroy notify functions to be removed based on the data associated with them.

Fix prompt cleanup (was not going through the list correctly when freeing items)

Select the menu item under the mouse when the mouse is moved (Fix bug 5237)

Make ResizeRelative resize in terms of pixels for windows with resize increments (Fix bug 5228)

If the resize is less than the window's resize increment then we bump it up
to ensure a resize does at least happen.

Make the inactive osd text color fall back to the old osd property, and lastly inherit from the unfocused window color.

Allow an xml tag to be unregistered from the xml parser.

Properly build path names for obt_paths_try_exec

add menu_sort_entries() which sorts all entries in an ObMenu

this function sorts each group of entries that appear together between two
consecutive separators (or ends of the list)

Add another BSEARCH function that lets you search through an array of objects.

If you give it a value x, the macro lets you pull a value out of each object
that you want to compare to x.

skip empty strings when splitting paths in the environment

hide the menus before shutting them down on reconfigure

Fix for when the menu's update function returns FALSE and no menu is shown.

The ObMenuFrame was not being freed, and the child pointer was being left
pointing to the frame that was not visible.

Just put copyright starting year in README so we don't need to keep updating it. Make --version match README.

Generate a fake leave event on actions that move windows (used to generate another enter event?)

Link X into libobt (Fixes compile on debian).

Ignore enters on windows that are raised over focused window while being mapped.

Use the KeyCode to directly find the modifier mask. (Fix bug 5173)

We were finding the KeySym first, and then converting back to a modifier mask.
But KeySym on a key's release can differ from on its press, and we don't need
them to determine the modmask from the keycode.

[setxkbmap -option "grp:shifts_toggle"] turns Shift_L into XK_ISO_Prev_Group on
key release, and Shift_R into XK_ISO_Next_Group.

Don't make desktop switching interactive when no mods used. (Fix bug #5203)

Make mod state passing more consistent, and always give actions the full state
instead of stripping sometimes. (They ended up expecting it stripped always).

Fix crash on unexpected NET_WM_MOVERESIZE_CANCEL messages

gtk+ 3.4 apparently sends these randomly when you select text in a
GtkEntry. This also fixes bug #5460.

Update german translation

Fix bootstrapping, forgot to change WANT_AUTOMAKE to 1.11 in previous commit

Quiet build outbut on by default

Fix window width/height going less than 1 on some platforms (bug #5246)

ignore .rej files

Fix maximized windows and such after screen resizings

Hear screen resizes when they don't change the root window (cloning)

Check the return value for launching the startup command.

Copy/paste error?

Fixes for gnome-session 3.

We run by default with a panel (default gnome-panel) and then run Openbox
without any panel if one cannot be found.

- Adds a fallback session for if a panel (default gnome-panel) not found.
- Removes notifications as a requirement.  Seems to be a legacy thing judging
  from the ubuntu .session files.  Notifications are being provided by
  notify-osd on modern systems, and you can't check for its presence in
  gnome-session (gnome-classic.session always fails because it looks for it, so
  ubuntu falls back to gnome-fallback.session).

Remove unused variable

respect the active/mouse options for monitor placement, and use ObMonitorPlace for per-app too

Allow per-app settings to pick a monitor for a window without choosing a position in it.

Add "active" and "primary" options to the <monitor> placement option for per-app settings (bug #5180)

make undecorated windows place according to their undecorated state (bug #5179)

splits client_setup_decor_and_functions() into 3 functions.

1. add client_setup_default_decor_and_functions()
- called from client_get_all() to get the maximum decor/functions that will
  be available for use by the client.
2. add client_setup_decor_undecorated()
- sets up the client's undecorarted decor if the flag is set by the per-app
  settings or session state.
- we do this before setting up the frame so the frame reflects the window as it
  should be when getting placed.
3. client_setup_decor_and_functions()
- calls the above 2 to perform the same functions as before.
- added to client_apply_startup_state() so that we can ensure it was run fully
  at least once in the mapping process, since it is not called in
  client_get_all() anymore.

it would seem this cannot return NULL now

It will get the primary monitor if there is not monitor under the pointer. But
assert so it's clear something went wrong if this does happen. Note that there
was previously no check for the return value even though the comment claimed
there should be.

Link libobt into libobrender, as it uses things from libobt (bug #5150)

libobrender uses ObtPaths stuff, so it needs to link in libobt.

Propagate _NET_WM_WINDOW_OPACITY to the frame window (bug #5132)

Strip leading/trailing whitespace when parsing colors and other strings (bug #4937)

free undecorated border colors

Specific border width/color for undecorated active/inactive windows (bug #4889)

The attached patch adds three theme options with backward compatible defaults:

undecorated.border.width
  defaults to border.width

window.undecorated.active.border.color
  defaults to window.active.border.color

window.undecorated.inactive.border.color
  defaults to window.inactive.border.color

Fix crash when keyboard map changes and no keybindings exist.

Don't assume old tree is non-null and start rebinding it.

Change default doubleclick timeout to 500ms and keep track of where last click was

Bug #5152 - "mouse double-click time is too low by default - 200ms"

We only use the doubleclick in one place in the default configuration,
for doubleclicking titlebars to maximize windows, so any negative impact
of increasing the timeout should be minimal, especially with the
addition of requiring the two clicks to be in the same place.

Doubleclicks are hardcoded to occur within 8 pixels for now, it doesn't
seem worth it to add a config until someone complains. A possibility is
using the drag threshold, but some people have that set very low so it
could be hard to doubleclick then.

Parse all the actions in a single menu item

Don't remove openbox-gnome.session on make clean - it's not generated.

Make openbox-gnome-session compatible with gnome3 with gnome-session 3.0+ support.

Yay they did something nice and made it easier for other window managers to be used.  How unexpected and pleasant.

Don't show prompts inside the message handler to prevent recursion

Showing prompts causes messages to be created which causes the glib message
handler to abort().  Save the messages and show them when done all other
processing for the current event.

bug #5253 kde4 is out now, use kde4 menus by default.

Startup scripts should export "XDG_MENU_PREFIX=kde-4-" but they export
"XDG_MENU_PREFIX=kde-".

Make "no" a valid choice for the dialog option in cyclewindows.c.

It is the same as "none" for backward compatibility (it used to be a boolean).

Update copyright notice in --version output

Fix typo in configure.ac, noticed by hwoarang

Consistency with helper/modal windows and multiple desktops.

When focusing a window, bring its helper windows.  This happened before when
switching its desktop, but if you have 2 windows on different desktops and
switch between them, the helpers should come to both when they are being used.

When a new helper/modal window maps on a different desktop and you are using
the application, then the helper/modal window is moved to the current desktop
to be with the application you are using.

Fix focus stealing for changing desktops/Use timestamp for user input events

Change the logic for when to allow stealing focus across desktops.
- It was possible to call event_time_after() with a CurrentTime in the old code.
- It would disallow a user requested change which is crazy.
- It would change desktops on you when a new window appeared but this is
  generally not desirable.

event_source_time() is supposed to give the time which the user made things
happen. we leave it at 0 for user-input events right now which means stuff like
changing desktop doesn't save any timestamp at all.  we should use the
timestamp from x for user-generated events.

Fix a memleak introduced by 0dd91ba0e2382a8f6177f501f8a0b26e0daf26a8, some cleanups

be a bit more explicit about what is being copied from one texture to another

Separate theme options for osd prompt buttons.

I managed to keep backwards compatibility without really cluttering
the code so here is the patch
http://bugzilla.icculus.org/show_bug.cgi?id=4874 too.

Please keep in mind that this is my first piece of code for openbox
and that I'm not a die hard openbox user (yet), not to tell the patch
was not exhaustively tested. Anyway I think it's pretty much in a good
shape but any criticism will be welcome.

Basically the patch add the following theme options for controlling
buttons in osd prompts:

%%%% colors
%
% for the text inside the button
osd.button.unpressed.text.color
osd.button.pressed.text.color
osd.button.focused.text.color
%
% for the line art around the button
%  (if you don't wan't the box just make box.color = bg.color)
osd.button.pressed.box.color
osd.button.focused.box.color

%%%% textures
%
osd.button.unpressed.bg
osd.button.pressed.bg
osd.button.focused.bg

The buttons can be in three states:

 unpressed: neither clicked nor selected
 focused: selected but not clicked
 pressed: clicked (and of course selected)

I discarded the previous distinction between press and pfocus as in
fact it was only a formal distinction, in that both appearances
mimicked each other in every sense. It think that it was just
inherited from the way titlebar buttons are managed so I decided to
simplify it a bit.

All the options default in a way that preserves backwards compatibility:

osd.button.unpressed.text.color  -> osd.active.label.text.color
osd.button.pressed.text.color -> osd.active.label.text.color
osd.button.focused.text.color -> osd.active.label.text.color

osd.button.pressed.box.color ->  window.active.button.pressed.image.color
osd.button.focused.box.color -> window.active.button.hover.image.color

osd.button.unpressed.bg -> window.active.button.unpressed.bg
osd.button.pressed.bg -> window.active.button.pressed.bg
osd.button.focused.bg -> window.active.button.hover.bg

Notice that a good deal of locs where added to theme.c but in
compensation prompt.c is pretty much simpler now because the
appearances and textures are created while loading the theme.

Merge branch 'm4/master'

Revert these two commits temporarily to make the buttons merge easier

1e427a3358deeadf8abc326e714ee201fddc43d6
2288da0ae3b44c987687e3deeb7864fcb7b23bb9

Update portuguese translation

Update estonian translation

Add Plural-Forms header to fi.po and remove fuzzy mark.

Update polish translation.

add --disable options for x extensions

so we can test compiling without them present

Update Brazilian Portuguese translation

add the 3.4.10+ changes to the CHANGELOG

Add 3.5.0 changes

Update other po files

Update swedish translation

Fix some translation string markings

Allow setting icons for submenus

Fix menu placement to avoid dead xinerama areas, possibly break other stuff

Use screen_find_monitor(area) instead of screen_find_monitor_point(
topleft corner) in order to find a better monitor when the menu isn't
opening with the mouse cursor in the top left corner.

I made screen_find_monitor return the primary screen when it failed to
find a monitor containing the rect, instead of the total area, no idea
what behaviour this will change but I doubt it will be worse.

Link libraries directly to compile on some systems (debian).

If program (A) is linked to library (B) which is linked to library (C),
then A must also link C directly in order to use things from it directly.

The environments are passed in from the openbox-session script so use them.

Fix bug #4877 (Some harmless code quirks involving booleans)

be a bit more explicit about what is being copied from one texture to another

Separate theme options for osd prompt buttons.

I managed to keep backwards compatibility without really cluttering
the code so here is the patch
http://bugzilla.icculus.org/show_bug.cgi?id=4874 too.

Please keep in mind that this is my first piece of code for openbox
and that I'm not a die hard openbox user (yet), not to tell the patch
was not exhaustively tested. Anyway I think it's pretty much in a good
shape but any criticism will be welcome.

Basically the patch add the following theme options for controlling
buttons in osd prompts:

%%%% colors
%
% for the text inside the button
osd.button.unpressed.text.color
osd.button.pressed.text.color
osd.button.focused.text.color
%
% for the line art around the button
%  (if you don't wan't the box just make box.color = bg.color)
osd.button.pressed.box.color
osd.button.focused.box.color

%%%% textures
%
osd.button.unpressed.bg
osd.button.pressed.bg
osd.button.focused.bg

The buttons can be in three states:

 unpressed: neither clicked nor selected
 focused: selected but not clicked
 pressed: clicked (and of course selected)

I discarded the previous distinction between press and pfocus as in
fact it was only a formal distinction, in that both appearances
mimicked each other in every sense. It think that it was just
inherited from the way titlebar buttons are managed so I decided to
simplify it a bit.

All the options default in a way that preserves backwards compatibility:

osd.button.unpressed.text.color  -> osd.active.label.text.color
osd.button.pressed.text.color -> osd.active.label.text.color
osd.button.focused.text.color -> osd.active.label.text.color

osd.button.pressed.box.color ->  window.active.button.pressed.image.color
osd.button.focused.box.color -> window.active.button.hover.image.color

osd.button.unpressed.bg -> window.active.button.unpressed.bg
osd.button.pressed.bg -> window.active.button.pressed.bg
osd.button.focused.bg -> window.active.button.hover.bg

Notice that a good deal of locs where added to theme.c but in
compensation prompt.c is pretty much simpler now because the
appearances and textures are created while loading the theme.

move Imlib to the option deps in the git readme

Update note on dependencies wrt autopoint

Update README.GIT to the packages in current ubuntu distribution and include imlib2

Comment fix

bump version to 3.5.0 and library versions

Fix Java apps always appearing on desktop 0.

Fake managing a window doesn't read a requested desktop, but ended up placing
  a NET_WM_DESKTOP hint on the window (with value 0).  Fake managing doesn't
  need to set the DESKTOP hint since the window is not actually being managed,
  so remove it from the codepath.

don't re-run the startup command on reconfigure

fix global autostart installation, and the pointer to it so it is run

Resizing removes maximization now, so don't only snap "grow to edge" to screen edges.

This caused a serious annoyance when shrinking a maximized window, it would
shrink to the other end of the monitor, effectively reducing it to its minimum
size.

Fix compiling without sync extension present.

Make waiting_for_sync always present, but it is set to 0 (not waiting) if we aren't using sync.

Fix regression: hitting a key accel in a menu didn't move focus to the submenu

Show the submenu and move focus there with the enter key when the submenu isn't shown yet.

don't free the XIC if it wasn't created.  this appears to cause a crash

cap the client (and frame) window's dimensions at the range of an unsigned short (bug #4596)

extra \n in debug output

Move focus to a window when the mouse enters a window during a pointer grab, if using "focus under mouse" (Fixes bug #4617)

when focus=yes in the per-app settings, then the focus request for the new window is from a user.  also if allow_other_desktops is false, then always disallow other desktops (Fixes bug #4752)

Make NET_ACTIVE messages always treated as from the user.  Loosen up focus stealing for user-requested focusing.

Seems panels such as xfce's and gnome's still treat their activation requests
  as being from an application when a user has requested it.

Make the focus stealing code more lenient for user-requested focusings
  (_NET_ACTIVE).  But treat new windows as not user-requested unless they
  gave a launch time.

When activating a window, if another window would be the one to actually get
  focused, then activate that instead (avoid clicking a window in the panel and
  nothing happens).

don't use -Wextra unless --enable-super-warnings

setenv and unsetenv dont exist in Solaris 9.  (Fixes bug #4663)

Seeing as how you already use glib, I just substituted g_setenv and g_unsetenv
as appropriate, and it now works for me.

Some versions of X, have the Shape extension, but apparently not "ShapeInput" (Fixes bug #4662)

the sawfish window manager has ifdefs for this sort of situation.
I followed suit, and #ifdef'd it, and it now works for me.
patch attached.

Slight changes to the patch from danakj@orodu.net for readability

make the default drag threshold 1.  it is smooth and feels and looks nice.

(also macos has it and i liked it.)

XML_PARSE_XINCLUDE flag seems to do nothing. Use xmlXIncludeProcessFlags() instead

if no launch time is provided for an application, make one up.

if the window is related to other existing windows
  and one of those windows was the last used
    then we will give it a launch time equal to the last user time,
    which will end up giving the window focus probably.
  else
    the window is related to other windows, but you are not working in them?
    seems suspicious, so we will give it a launch time of NOW - STEAL_INTERVAL,
    so it will be given focus only if we didn't use something else during the
    steal interval.
else
  the window is all on its own, so we can't judge it.  give it a launch time
  equal to the last user time, so it will probably take focus.

this way running things from a terminal will give them focus, but popups
without a launch time shouldn't steal focus so easily.

fix segfault for placing windows without a group

add primary monitor stuff to rc.xsd

Fix small leak in If action option parsing

Big changes to placement across multiple monitors.

Add a Primary option for which monitor to place new windows on.  Make "Active" the default instead of "Any", which is just totally crazy.

When a window is being placed in the FOREGROUND, use a monitor chosen in
  the following order:
  1. same monitor as parent
  2. primary monitor if placement=PRIMARY
     active monitor if placement=ACTIVE
     pointer monitor if placement=MOUSE
  3. primary monitor
  4. other monitors where the window has group members on the same desktop
  5. other monitors where the window has group members on other desktops
  6. other monitors

When a window is being placed in the BACKGROUND, use a monitor chosen in the
  following order:
  1. same monitor as parent
  2. other monitors where the window has group members on the same desktop
   2a. primary monitor in this set
   2b. other monitors in this set
  3. other monitors where the window has group members on other desktops
   3a. primary monitor in this set
   3b. other monitors in this set
  4. other monitors
   4a. primary monitor in this set
   4b. other monitors in this set

Decide to focus the new window before placing it, so we know if it will be
  placed in the foreground or background.

Always choose a single monitor, then place on it, rather than possibly moving
  to a "backup" monitor.  Unpredictable monitor placement is horrible.

add function screen_compare_desktops()

tells if two windows' current desktops are considered logically on the same
  desktop (taking "all desktops" into account)
if a window is on "all desktops" it is considered to be on the current desktop
  only - windows can only be in one place at a time.

don't steal focus if someone used a window within 1 second ago

(instead of 0.5 seconds)

Fixing commit 066a15d7aa64eaa0.  Don't ever use CurrentTime as event_curtime.

Force event_curtime to be some valid timestamp.

Allow XIncludes in xml documents we read.

Add the XInclude namespace in the default config file (named "xi")

use top_builddir, not builddir variable in Makefile.  seems builddir doesn't always exist/anymore

add some more warning flags for debug builds

-Wno-write-strings so that "foo" is treated as a const char* not a char*.
-Wextra for more fun

Fix icons.c test for 64-bit platforms

update the last user interaction timestamp during a move/resize of the focused window

update the user-interaction timestamp when running a (non-interactive) action on the focused window
update it at the end of running actions instead of multiple times (once for
 each action)

don't steal focus from a window if it was used very recently, when someone uses _NET_ACTIVE request.

mem leak in config.c

Big rework of image.c and the image cache system.

Added a lot of comments, simplified call graphs.
Added full (not second-class) support for images coming from named sources (files, icon themes).
RrImage holds an RrImageSet.  RrImageSet holds a bunch of RrImagePic, which are different sizes of a logical image.
RrImageSet objects can be merged if it is discovered they (will) share an RrImagePic.  The RrImage objects are updated to use the new merged RrImageSet.

Use >= instead of > when comparing timestamps, or we will ignore some events

add some tests for obt pieces (binary search, .desktop parsking, linkbase, and inotify watch)

oops, crash if searching for something smaller than * in the array

less redundant calls to screen_update_areas() and client_move_onscreen()

Ignore MotionNotify on menus, it reset the submenudelay timer

It also caused the slightest cursor movement to activate the entry under
the cursor when navigating via the keyboard.

Use g_list_find instead of weird bouncing loops

Parse multiple keys in one keybind

allow the res_class in WM_CLASS to mark a window as a dock app, if it is set to the string "DockApp"

Don't hog the user during the hideDelay

Handle STRING and COMPOUND_TEXT type text properties.

STRING should be latin1 text (plus TAB and LF)
COMPOUND_TEXT should be encoded in the current locale.

Revert "add an inotify watcher for directories (doesnt work without inotify yet)"

This reverts commit c36b89ba12eae18d3011c8516906c21e9abb89dc.

fix make distcheck

glib can't handle -Wcast-qual

add wmhints test that toggles motif decor hints on a mapped window

Fix a double free when you have a mousebind without a button attribute

Introduced in f307a3feabedd9bcadeaafd0fa8e1b1a60736eb2

Fix a copypasto that broke interactive keybinds with meta and hyper keys

Revert "add XFlush to g_timeout callbacks"

This reverts commit aa3685d16af4a565eecdc39047ee8d140ef5cd99.

Mika's commit using a GSource to handle X Events fixes this problem, so this
commit is not needed

Fix 3.4 compatibility for SendToDesktop

Correct typo occured -> occurred in function name

Add signal handling with the GMainLoop

Provided through a very simplistic interface in obt, found in the
 obt/signal.[ch] files

add stackabove test, that lets you move a client window just above a specified window in the stacking order

80 cols

Continue parsing contexts when an unsupported one is encountered

Make the dock a context and add actions LowerDock and RaiseDock

Move usage of iconh into the branch that assigns it

settings can't be NULL here, or we'd have already crashed anyway

Use a gsource instead of gio channels for watching x events

Fixes not getting some focusin events sometimes.

obt returns libxml2 structures directly, so anyone linking against it better also link to libxml2

add XFlush to g_timeout callbacks

GMainLoop doesn't cause an flush to occur after handling a timeout.

Fix the EXTENTS reported on window frames.

Change the order things are done on map, so we apply startup state without
  any states set first, thus getting all the functions possible for the window
  type etc.  then change the order states are applied, as some remove the
  ability to apply others (ie fullscreen comes last).
Add an oldsize to ObFrame that remembers the size of the frame last reported to
  the world through the EXTENTS property.  If you frame_adjust_area(FAKE) then
  frame_adjust_area(NONFAKE), the EXTENTS would not be updated since the
  oldsize wasn't remembered across fake updates.
Make the extentsrequest test ask about windows with states set also.

keep stacking_list_tail a pointer to the tail of the stacking list

allow a user to set "undecorated" on a window without decorations

if not based on window type, the app may decide to have decorations again
sometime.  this was not always legal but chromium has made it so.

add an inotify watcher for directories (doesnt work without inotify yet)

Use GMainLoop instead of ObtMainLoop

save the list of mime types listed as supported by an application

parse app Categories better.  parse a list of strings into a gchar**

parse Categories tag and save them as an array of GQuarks

improved .desktop parsing.

properly (and quickly) check for existence of required keys
figure out what an app can open from its exec key
validate the %fields in an app's exec key

free the user's group ids

free the path dirs

parse Path, Terminal, and StartupNotify

parse OnlyShowIn/NotShowIn

allow new windows on other desktops to get focused if specified in a per-app rule

you can not client_active() with user=false and desktop=true and activate things on other desktops, by the general focus steal restrictions, if you wanted to do that sorta thing

Allow sending windows on other desktops to the current desktop.

Add some more conditions to the If action

It can now match on window titles, virtual desktop (number, "current",
"other"),
demands_attention, and undecoratedness.

don't need a force_reply when changing the decor for a window

it will send a configurenotify if the client's rootwindow position changed

client_update_protocols() doesnt change anything that would affect a client's decor or functions

allows a window to unmax by resizing.

makes "clever" use of the client's pre_max_area values to allow this to happen
without flashing, and preserving them in case the user cancels the resize
after it has become unmaximized.

set the frame's initial size to something invalid so the extents hint will be
set for sure the first time.

set the interactive action to NULL before doing its cleanup phase

if the cleanup ends up cancelling the "current interactive action", it would
cause the end function to be called again, which sucks.  for instance if the
cleanup function uses the Focus action. this prevents the unwanted loop
into the interactive action cleanup.

fix focus when starting up, especially when replacing another instance of openbox

adds a function event_reset_time() that forces event_time() to look for a new (future) timestamp

remove a debug print in the keyboard obt stuff

avoid new gcc warning (casting ** to const** is not valid for good reason)

merge some contexts together in the default rc.xml mouse bindings

allow multiple contexts separated by space in a mouse binding

example: context="Top Left Right Bottom"

turn off focuslast and on undermouse in mouse focus config example

Remove some unneeded stdlib.h includes

Add support for using relative expressions in move and resize actions

Have MoveResizeTo use config_parse_gravity_coord instead of duplicating it locally
Allow MoveResizeTo positions and sizes and per app positions to be relative to screen size
Rename to config_parse_relative_number so it can be used for sizes too
Add relative numbers to width/height in MoveResizeTo
Add relative numbers to MoveRelative
Add relative numbers to ResizeRelative, these are for the client size, not screen size

new enum values need to be handled in switches

(the new enum values were added to count the number of entries in the enums)

fix typo in enum

add _NUM_TYPES to obrender enums, to allow enumeration of the enum values

show empty properties in obxprop

parse number values write in a hex string in obxprop

advance while parsing a hex input in obxprop

add cleanup functions to the client list menus that removes everything from them when they are not being shown (saves iterating thru the list uselessly when a client is unmanaged)

add cleanup function for menus that fires when the menu is no longer visible

don't try go to selected submenu when there is nothing selected

use clicks in prompts

add a test to set a mouse cursor on an InputOnly subwindow

fix XSYNCronization of resizes (especially with two monitors).

1) a ConfigureNotify must always follow a Sync notification
2) determine the final size of the window with its position else struts are not applied properly for (partly) maximized windows
3) set the sync counter to a value when managing a new window

make sure clients stay on screen when the root window changes size, and maximized/fullscreen windows must be on some monitor or it doesn't make much sense (and asserts)

Remove two unused variables

Remains from 64adc0eeba598cb1469b2140777fba30e3053f0e

found memory corruption in the obt xqueue and paths

make an event queue for X events.  the queue's min size is 16 XEvents (~3k)

don't put reparent events back on the stack.  the window is no longer being managed anyways.  and it messes up the order of events in the queue

putting events on the event queue does not send them back to the server.  it was just going to get ignored the next time around cuz the window wouldnt be in our list of clients

don't use XPutBackEvent and mess up timestamp order in the event queue

add asserts to make sure we don't add things to stacking list that are not managed

don't change stacking stuff on windows when moving them during the unmanage phase.

event_curtime is replaced (publicly) by event_time()

This function never returns CurrentTime, which is nice, cuz using CurrentTime for XSetFocus always sucks.

If the current XEvent did not have a timestamp, then event_time() will find one.  It finds the first timestamp available in the X event queue, meaning the earliest timestamp >= the current (nontimestamped) event. All future events should have a timestamp >= event_time(), so using this in XSetFocus() should not mess up any future calls we make to it.

This change seems to work well, as it appears to fix bug #3648.

When determining the current timestamp, try get something a lil more accurate

Get the first timestamp from the event queue, rather than (potentially) the
  last.
also treat it as the actual event_curtime, meaning it is used when focusing a
  newly mapped window etc.

when client removes its transient_for hint, don't keep the window as transient for the group when it shouldn't be (fixes bug #4586)

allow warping the mouse pointer when switching desktops by bumping into the edge of the monitor with a window

based on a patch by Nathaniel Gephart <computinchuck@gmail.com>

show the focus cycle popup's icons in linear order when cycling in linear order (bug #2319)

add an option to next/previous window to only include hilited/flashing/urgent windows (feature req/bug #4341)

Link libobrender, not openbox, to imlib2

add some comments for binary search

stat() can give an error, handle that

use const Rect* not Rect const*

add BSEARCH() macros in obt/bsearch.h that to a binary search on an array, use it when searching for a group in obt_paths_try_exec()

parse some of the .desktop stuff into an ObtLink structure

localized names still don't work.
most of the app-specific stuff isn't done yet.
categories aren't handled yet (to only show in/not show in some category)

add obt_paths_try_exec() that will see if an executable exists

save the LC_MESSAGES locale category

if a launch time isnt there (like for client_activate) assume the window launched before the last desktop switch, not after. ie don't change desktops if you're not sure.

make keypad enter do the same thing as return

figure out what type the value should be for each known .desktop key, and watch for invalid keys

Add a new ObtLink type, which is going to be a generalization of a .desktop entry.

Move the .desktop file parsing details into obt/ddparse.c, which generates
  some hash tables holding the data of the file.
Add a new obt/link.c which will build an ObtLink from a parsed .desktop file,
  and may support other ways to create these links in the unforseeable future

ignore the rendertest executable

ignore generated autostart script

save key values based on their value, but we dont know what values they have yet

add an empty function to parse the Desktop Entry group in a .desktop file

parse key/value pairs from the .desktop file and save them in a hashtable

close the .desktop file after parsing it

free the path after done with it

g_warning adds a \n

watch for groups in .desktop files and remember the current group. validate the names, and make sure no duplicates and "Desktop Entry" comes first

avoid ascii control characters in strings

add beginning of .desktop file parsing, grabs a line of text from the input file at a time and strips out any \0 characters

make enum values match numbers in the spec, and make the .desktop file reader function take a list of paths to look for the file in

install the obt/ddfile.h header

add function to give the list of autostart directories

add some functions for parsing a value in a .desktop file

allow non-normal windows to go to 0,0.  and remove a debugging g_print

Add menu seperator style to Clearlooks

The default themerc is quite old and does not include the slightly newer
option to theme the menu seperator, here is a patch to fix that and some
very trivial tweaks to overall theme to better match the metacity/gtk
version.

add a function client_is_oldfullscreen() that determines if a window is an "oldschool fullscreen window" or not.  use this throughout.  and allow oldschool fullscreen windows to move to 0,0 even with a strut there.

when an app tries to move to 0,0 and there is a strut there (java swing!), move it to inside the strut instead (unless it is an oldschool fullscreen app)

missing a g_slice_free

libexecdir is provided by automake, and gnome-settings-daemon is not always in /usr/libexecdir, use the dir from automake

Type is a mandatory field of a .desktop, so add it to the old gnome-interfacing-desktop file

Show client menu on right click anywhere on a window frame (left and right were excluded)

Added shade-icon to client-menu

make array large enough

get the modifier mask for a modifier key that is pressed/released more accurately, using the table we built already for other purposes

save the state of the modifiers only when starting an interactive action

move the startup command launching into a helper function

revamp the autostart functionality, so autostart apps launch after openbox.

add /usr/libexec/openbox-autostart which runs the global/user scripts and
  launches /usr/libexec/openbox-xdg-autostart (which runs .desktop autostart
  stuff)

make openbox-session call openbox with --startup option, to make openbox launch
  /usr/libexec/openbox-autostart *after* it is initialized.

add /etc/xdg/openbox/environment and ~/.config/openbox/environment files to let
  user change environment variables (locale) for their Openbox session, which
  openbox-session loads before running openbox.

don't use a variable as a format string

Add skeleton for start of dot-desktop-file parser

lower the default submenu show delay

a couple more g_free() to g_slice_free()'s

more using g_slice_new() instead of g_new()

missed a g_free() that should be g_slice_free()

dont trust the _NET_ACTIVE_WINDOW timestamp. (fixes bug #4519)

if we decide to focus the window, do not use their provided timestamp.
chromium gives a very old timestamp, which means we think we're focusing the
window but it never actually gets focus, leading to inconsistent behaviour by
openbox.  use the timestamp for making decisions about passing focus if you
want (we dont right now), but not for the XSetInputFocus call.

Remove mismerge-added files

include imlib2 as a dep of obrender in the pkg-config file if it's using it

make the obt .pc file do -lobt, not -lobrender

remove a bunch of g_new()/g_free() by returning Rect const*'s from screen area functions when possible

use g_slice_new() instead of g_new() part 4

use g_slice_new() instead of g_new() part 3

use g_slice_new() instead of g_new() part 2

using g_slice_new() instead of g_new() part 1

have obt refresh the xlib keyboard stuff on mappingnotify events

destroy XIC's that belong to an XIM before closing the XIM, else Xlib will seggie!

Check node->name for null

Seems to happen sometimes with a well placed xml comment.

dont print warnings when reading keys which dont generate text

make control keys work in menus/dialogs/etc with the new obt code, using XLookup stuff

initialize the action shutdown function to NULL

pass an input context to the interactive input function

make menu input work with the new obt keyboard stuff.

create an Input Context for use during keyboard grabs, and use this to
  translate KeyPress events properly.
make the menu respond to KeyPress events, since these are the only ones that
  can be translated properly, but still execute things from KeyRelease events

make the menu use the new obt stuff when reading keyboard input so it can handle input from the non-base group and composed input

make obt translate keypress events to a valid utf8 character (note this makes openbox not compile)

don't strip the state for keyboard events

this allows translation of keys not in the base keyboard layout

add interface in obt to create an Input Context for a window

create an X Input Method in obt for the app to use for all input

get xlib to reload the keyboard on MappingNotify events

XKB header comes from obt

check the xkb version before using it

make the execute action not segfault when using a prompt (bug #4543)

also save the client when using a prompt so it can be used if you choose "yes" from the prompt

add an optional shutdown function which actions can register

i learnt what xkb does with the state.

use the state from keyrelease events directly, rather than query the state (which is not as accurate!)

the xkb state (as opposed to the normally sent compat state) contains extra info like the keyboard group, the pointer buttons, etc.  so we can just strip that stuff out. (See section 2.2.2 of the XKB proto document)

dont assert when there is no valid key layout present

Use correct ac variable in version.h.in

don't link libobt into libobrender

it makes "warning relinking" weird things happen when doing make install.
we already list it as dependency in the pkgconfig file anyway

fixing up obxprop manpage

Merge branch 'backport' into work

Merge branch 'backport' into work

Init all fonts to NULL, so if they arent in the config they use the default font

Add a hook to unset the client in the open menu frame instead of closing it

Also nukes menu_frame_hide_all_client() as it is now unused.

Don't close the menu when you held control even for execute

Don't alt-tab to skip_taskbar iconic windows if they have parents

adjust who shows up in the alt-tab list using SKIP_TASKBAR (Fixes bug #4503)

allow window types != NORMAL to not show up in the alt-tab list
but don't allow DIALOG types to hide from the list
also any window which is DEMANDS_ATTENTION should appear in the list regardless
  of its SKIP_TASKBAR status

don't use client_move() to end a move

client_move() uses the client's current width/height, whereas all along we
  were using the width/height the client had at the start, so this causes
  client_configure() to behave differently at the end than during the move

allow multiple escaped _'s in a menu label, and allow a real _ to come later in the label (Fixes bug #4355).

make edge detection for maximized windows only look at the edges of screens

catch the SIGTTOU and SIGTTIN signals when an app we launch tries to use its stdin/out (but we don't have one either).  Fixes bug #4072.

if we get an event on a subwindow of root that we don't know about, but it is related to a press we received on the root window, then process the event (Fixes bug #3702)

Remember the maximized state of a window when it goes fullscreen and try restore it after

If a window places itself at (0,0) and there are struts there, assume it is a bug.  Also allow oldschool fullscreen windows that cover all monitors on a multihead setup to work properly

Added a test 'oldfullscreen' that makes a oldschool fullscreen window

allow app rules to match windows by their title when mapping

and save the title in the _OB_APP_TITLE property

if openbox isn't running, show messages on the console rather than just losing them (like when another WM is running)

Merge branch 'backport' into work

Conflicts:

	openbox/frame.c

remove the session.c file from actions, as it is no longer used

NULL ic when we failed to load an image

Merge branch 'backport' into work

Conflicts:

	configure.ac
	data/rc.xml
	openbox/client.c
	openbox/event.c
	openbox/focus_cycle.c
	openbox/focus_cycle_popup.c
	openbox/openbox.c
	openbox/prop.c
	openbox/prop.h
	openbox/screen.c
	parser/parse.c
	version.h.in

Fix ordering so fallback for inactive osd to label works

The themeing options (sic) for inactive osd labels added in
cfdbe23e0d9b72b4c220a10ddf60a4e981967b63 had the fallback to inactive
label before that was read in, so fallback failed to pick up some
things.

allow application rules to place transient/dialog/splash windows, and allow application rules to match only on the role or type if you wish.

Rework the code provided by Kadlcik Libor for loading/showing icons in Openbox menus.

This changes how the imagecache works, you can load an image into it directly,
or you can load it by name (then it will load it from a file on disk).

NOTE: The name part is incomplete, as it needs to use the freedesktop.org icon
spec to search for the right file.  Also to resize it should look for another
icon on disk with the same name but different size (icon themes).

No \n on ob_debug in this branch.

Some style changes, less ifdefs.

This patch implements support for icons in user-defined menus into Openbox

Image loading is done using the Imlib2 library.
I chose Imlib2 because it's pretty fast, it's easy to use, supports many file
formats (tested xpm, gif, jpeg, png) and doesn't introduce too much bloat (it
depends :)).

I ported the patch to 3.4.7-pre3 and added some enhancements. Caching is much
better now, and icons can be disabled at compile time using --disable-imlib2
option.

What's new?
Syntax of configuration files (namely rc.xml and menu.xml) has been changed
slightly to allow users to associate icons to menu entries. This is done by
specifying path to icon file in the new "icon" attribute in "<item>" element,
e.g:
<item label="Vim" icon="/usr/share/pixmaps/vim-32.xpm">
  <action name="Execute"><execute>x-terminal-emulator -T Vim -e
vim</execute></action>
</item>

If user doesn't want to display any icons in his user-defined menus, he/she can
disable icons in rc.xml, inside "<menu>" section:
<menu>
 ...
  <showIcons>no</showIcons>
 ...
</menu>
Default value is "yes".
(New boolean variable "config_menu_user_show_icons" has been added to source
code.)

An icon is loaded (using menu_item_attach_icon()) when a new entry of menu is
created. Fortunately, I haven't notice any performance problems because of this
:).

Make focus_valid_target() take a desktop argument.

So you can find valid targets on desktops other than the currently visible one

use the valid_focus_target() method for choosing what to show in the client list menus

allow alt-tab etc to move focus to skip_taskbar windows when they are iconic

use obt_xml_* to navigate the xml file (also XML tags are case sensitive)

ignore some added tests

Don't handle input events on prompts if they should be used for a binding/menu instead.

Also, be more careful about making the prompt buttons look pressed, don't make them pressed from a motion notify event if they didnt first handle the press.

Merge branch 'backport' into work

Conflicts:

	Makefile.am
	openbox/actions/focus.c
	openbox/config.c
	openbox/event.c
	openbox/menuframe.c

Set the XDG_MENU_PREFIX to "kde-" when running an Openbox/KDE session.

reflect the change that parse.c/h was renamed to xml.c/h in the file header comments

when holding Shift for the client list menus, bring the selected client to the current desktop (Request #4284)

check more headers before including them

rename the obt_parse library to obt_xml (since it is very xml specific)

add obt_parse_save_file() method to the obt parse library

Set up work for making a 3.5 prerelease.

Set version stuff to 3.5.0-rc1.
Copy the CHANGELOG from 3.4-working.
Rename the obt-4.0 and obrender-4.0 pkgconfig stuff to obt-3.5 and obrender-3.5
Rename the "render" directory to "obrender" so that the public headers can be
  installed in <obrender/*>

Merge branch 'backport' into work

make a config option for the inactive osd font

Add themeing options for inactive osd labels, osd.inactive.label.*

(rename the current osd label options from osd.label to osd.active.label, similar to the window.(in)?active.label options.

Add a focus option, unfocusOnLeave that removes focus from a window when the pointer leaves it

This uses the same delay to unfocus as is used for focusing on enter

Merge branch 'backport' into work

Conflicts:

	openbox/screen.c

Show the desktop popup as long as the interactive desktop switch is active

Merge branch 'backport' into work

Make the relative desktop switching actions interactive

The desktop popup can now be shown "permanently" in which case it stays until
you tell it to hide.

Allow the interactive cancel function to be optional

Add some directional focus bindings to the default config file.

Make it possible for an action name to choose whether it is interactive or not based on its options.

This way we can use the same name with options for an interactive action and a
non-interactive action.

Shorten the names of the ObActionsInteractive* functions to ObActionsI*

Add a ObActionsIPreFunc that is called for interactive actions
before the interactivity (key/mouse grab) is started.

Add a ObActionsIPostFunc that is called for interactive actions
after the interactiviti (key/mouse grab) has ended.

Merge branch 'backport' into work

Conflicts:

	obt/keyboard.c
	obt/keyboard.h
	openbox/event.c
	openbox/menuframe.c
	openbox/moveresize.c
	openbox/openbox.c
	openbox/screen.c

Merge branch 'backport' into work

Merge branch 'backport' into work

Conflicts:
	openbox/config.c
	openbox/event.c
	openbox/prop.c
	openbox/prop.h
	openbox/screen.c
	openbox/screen.h

Merge branch 'backport' into work

Conflicts:

	openbox/actions/all.h
	openbox/actions/session.c
	openbox/client.c
	openbox/event.c
	openbox/grab.c

strip leading/trailing whitespace off stuff when reading it from the configs

Merge branch 'backport' into work

Conflicts:
	openbox/popup.c

Fix an unsigned/signed warning.

Pointers are long-sized on x86_64.

Fix spacing from 8f37a97a4cc29d35f6a632ef2f71ac760e301ee5

Add a test that grows a window by changing its normal hints.

Revert "Revert "make windows stay on screen when..."

Just kidding.  This doesn't revert d81d54c7a977a034199f24fc8fa7433e634aaf91 but
it redoes the work that patch reverted in a better, less intrusive, way.

Revert all commits for the Hooks feature to move it to a topic branch.

This reverts commit fd2f617be7ff57fcb187daa737f66d243544c8cd.
This reverts commit a47d0a53652a96ca1df96fc9268757df1431ae55.
This reverts commit a1908e076736e1a618bd305c5963dbfecaa30497.
This reverts commit dc2e6f6bf7143a56de360a393b33906735e63625.
This reverts commit 985e7dadf9a3ebf4bd265d955c3198e96405e5d2.
This reverts commit db781556d63d1a50bd1b1b4b6b5423ef703bf2c7.
This reverts commit feec8f663f0a11546c2da87575fecc8a88d97ca1.

Make clang happier

Add asserts to default: in switch statements
Store pointed to variables locally so it knows they don't change
Remove some dead assignments
Mark ob_exit_with_error as noreturn
Use "%s", msg instead of just msg to printf style functions
Use the c_pfocus variable

Merge branch 'backport' into work

Merge branch 'backport' into work

Add all the action names used in 3.4 so configs don't break

Merge branch 'backport' into work

Conflicts:
	openbox/actions/desktop.c
	openbox/client.c
	openbox/event.c
	openbox/extensions.c
	openbox/popup.c
	openbox/screen.c
	parser/parse.c

Fix a typo in debug message

Add support for 24bit

Close openbox.log when restarting

Merge branch 'backport' into work

Conflicts:
	openbox/config.c

Merge branch 'backport' into work

Conflicts:
	openbox/event.c
	openbox/frame.c

Merge branch 'backport' into work

Another problem pointed out by clang

Using format specifiers without arguments in printf is still bad.

openbox/debug.c:105:18: warning: format string is not a string literal (potentially insecure)
    fprintf(out, log_domain);
                 ^~~~~~~~~~
openbox/debug.c:107:18: warning: format string is not a string literal (potentially insecure)
    fprintf(out, level);
                 ^~~~~
openbox/debug.c:109:18: warning: format string is not a string literal (potentially insecure)
    fprintf(out, message);
                 ^~~~~~~

Merge branch 'backport' into work

Conflicts:
	openbox/event.c

Merge branch 'backport' into work

Merge branch 'backport' into work

Conflicts:
	openbox/client.c
	openbox/prop.c
	openbox/prop.h

No \n on ob_debug in this branch.

The code to replace $wid/$pid died on $foo.

It failed to increment a variable after finding a $ but no pid or wid,
which led to a fun infinite loop. Also don't check if c[1] is w if we
already established it was p.

Revert "make windows stay on screen when they..."

This reverts commit 73adc6480d59fc5dccf9fe19ed5234c9582b328c.
I know I had a good reason for reverting this commit, I don't know what
that reason was though.

Merge branch 'backport' into work

Conflicts:

	openbox/config.c
	openbox/keyboard.c
	openbox/moveresize.c

Don't need to explicitly check for xft.

make windows stay on screen when they are resized due to their normal hints changing

remove some verbose debug prints

Make alt-tab icon size configurable.

fix a bug from when $window changed to $wid in the execute action

Merge branch 'backport' into work

Conflicts:

	obt/keyboard.c
	openbox/actions/cyclewindows.c
	openbox/actions/directionalwindows.c
	openbox/frame.c
	openbox/openbox.c

Fix a typo in --help output (master only).

Merge branch 'backport' into work

Conflicts:

	openbox/client.c

add some debug prints in the session code

make --debug-focus and --debug-session only show their appropriate messages so you can filter easier

Merge branch 'backport' into work

Conflicts:

	openbox/session.c

Merge branch 'backport' into work

Merge branch 'backport' into work

Conflicts:

	openbox/actions/execute.c
	openbox/event.c
	openbox/openbox.c
	openbox/openbox.h

use g_debug for debug messages

show prompts when printing error messages

Merge branch 'backport' into work

Merge branch 'backport' into work

Conflicts:

	openbox/actions/desktop.c
	openbox/actions/if.c

missing a header

Merge branch 'backport' into work

Conflicts:

	openbox/actions/all.h
	openbox/event.c
	openbox/prompt.c

queue hooks so that they don't run as each thing happens, instead run them at the end.  keep the openbox code paths fast.

make it $pid and $wid, it's a little faster this way and they look the same so why not

Merge branch 'backport' into work

move some code around in the execute action

instead of passing things to the environment, substitute $pid and $window inline in the Execute command string

pass the targeted window's PID (if available) and WINDOW_ID to the environment of the Execute action

always send the exact client to the hook's actions that the hook fired for

Get hooks working

change some of the hooks, and add all the hooks to the code so that they run

Add a hook system.  They hooks don't run yet but they parse from the config file.

Merge branch 'backport' into work

Conflicts:

	openbox/client.c
	openbox/screen.c

render depends on obt, so make/install it first (fixes making deb package from it)

Merge branch 'backport' into work

Conflicts:

	Makefile.am

Simplify client_set_state.

set the right version in the new render and obt .pc files

obrender uses obt now, there is no obparser anymore

Merge branch 'backport' into work

Conflicts:

	openbox/actions/execute.c

Merge branch 'backport' into work

Merge branch 'backport' into work

Conflicts:

	openbox/client.c
	openbox/event.c
	openbox/event.h

Merge branch 'backport' into work

Merge branch 'backport' into work

Merge branch 'backport' into work

Conflicts:

	openbox/menuframe.c
	openbox/prompt.c
	openbox/prop.c
	openbox/prop.h

Merge branch 'backport' into work

Conflicts:

	openbox/openbox.c
	openbox/session.c

Merge branch 'backport' into work

Conflicts:

	openbox/grab.c
	openbox/keyboard.c

() -> (void) also in .h files, and fix a place that gave arguments when it shouldn't have.

() -> (void)

Merge branch 'backport' into work

Conflicts:

	openbox/frame.c
	openbox/menuframe.c

remove prop.c/h, as they were reintroduced in a bad merge (0dc7eca4cdfff6425e19a0bace0f9ae8834d04e8)

Merge branch 'backport' into work

Merge branch 'backport' into work

Merge branch 'backport' into work

obprompts do not have a stacking layer, as they don't appear directly in the stacking list

draw a box inside the key-focused button in an ObPrompt.  also fix rendering the buttons when they are parent-relative.

fix min-size calculations for lineart textures

make ObPrompts resize and redraw correctly when reconfiguring and changing themes with them open

put the char member at the bottom of the struct so that it doesnt mess up alignment for the other members (assuming the compiler doesn't pad the struct appropriately anyways)

Merge branch 'backport' into work

Diffing against the old work branch where most of the changes
in backport were cherry-picked from indicates this should be
alright. (0de9097017d4d1991388a35e380a57dc1135b431)

don't reference free'd memory, and don't leak memory if the image cache is not saving any resized pictures

use g_message not g_print so it can go to the log

pre-calc the sum of a picture added to an RrImage rather than calculating it every time

add a bunch of comments for images and image caches.  and make the number of resized pictures saved in an image tunable per-icon cache.

add an override_redirect InputOnly window utility.

The intent was the following:
 This is mostly a copy of override.c, it lets you map a small window
 (1x1 at 0,0 by default) that openbox will treat like the root window,
 useful for scrolling to change desktops over fullscreen and maximized
 windows.

Since then openbox has learned to ignore other's override redirect windows
but I think it might still be useful to have this util.

when resizing images, pick a source image with the same aspect ratio if possible

Introducing the icon cache.

If an icon is the same as one in the cache, then it uses that one.
icons of different sizes (from the same client) are linked together into
one, and resizes of icons are cached and linked to all the various sizes.
so you only need one icon in memory for all your terminals now. ya!

don't ping everything all the time. yay.  ping when you close, and if it doesn't reply then show the kill prompt.   also show a more correct prompt for windows connecting from non-local machines - ask to disconnect them from the X server.

Merge branch 'backport'

make the text area in alttab list mode the same hight as the icon _with hilite_

commit d9d580e0382 broke kdesktop again, so we have to fix it at an even finer level.  make a pending ReplayPointer happen before moving/showing/hiding a window in an action

Fix a minor issue with lastdesktop right after startup.

Wrap the focus action in actions_client_move.

When alt-tabbing to iconified audacious and the playlist or eq window pops up under
the cursor, they "stole" focus from the main window.

Improve menu behaviour when clicking on grandparents.

Some time ago the menus were changed to not trigger on mouseover on a grandparent or higher
so that you wouldn't lose a carefully navigated menu structure. Clicking on an item opens
it though. It used to be on ButtonRelease but now it is on ButtonPress. Additionally, a Press
now causes grandchildren to be closed when clicking on already selected parts of the hierarchy.

Merge branch 'wip/alttab'

tell what signal is going to be sent to the client when killing through the dialog

use the hover and press button colors (this was a bug)

when drawing and measuring strings in "flow" mode, don't ellipsize so that lines can wrap

don't kill our own ObPrompt windows with kill actions

if a prompt is already showing and you try show it again, then make it active.  in the "kill this?" prompt use the window's original title without any of the openbox-appended-ness

prompt to kill windows when they are not responding

let you specify return codes for the different buttons in a prompt, and specify a callback function for the prompt for when it is closed

position the buttons in the bottom right instead of centering them

key input works for ObPrompt windows now

make the prompt buttons respond to button presses. keyboard input code is there too but not working yet.

properly place the msg texture

Make ObPrompt windows get managed as clients, and make them able to reconfigure as well.

give prompts a border, and fix how they are laid out. and make them use the multi-line text capabilities in render for the message

allow font rendering to use multiple lines

you can create dialog windows called "prompts" which have a message and some buttons!  they don't do anything interesting yet.

make the internal window type more easily inheritable

Merge branch 'backport' into work

Conflicts:

	openbox/config.c

Remove the //.

split the version.h into render/ and obt/ copies, with the appropriate version info

make the xerror output more clear, to say when errors are being properly ignored

unused var

missing include

Merge branch 'backport' into work

Conflicts:

	openbox/client.c
	openbox/event.c
	openbox/mouse.c
	openbox/openbox.c
	openbox/prop.c
	openbox/prop.h
	openbox/screen.c
	parser/parse.c
	parser/parse.h

don't crash when using <raise> and alt-tab and there are no windows to cycle between

remove the generated obt's pc file

install the right headers for obt/obrender

Merge branch 'backport' into work

generalize the window managing process into window_manage, which handles dock apps and the like

unlink the log file when opening it so you get a new file

Fix a memleak. (unref ObtPaths)

make ob_debug use the g_log system, and make openbox log messages to..a log file ! this will help everyone since most people do not run openbox in a terminal..

Merge branch 'backport' into work

Conflicts:

	openbox/client.c
	openbox/config.c
	openbox/event.c
	openbox/extensions.c
	openbox/focus_cycle_indicator.c
	openbox/focus_cycle_popup.c
	openbox/menuframe.c
	openbox/moveresize.c
	openbox/openbox.c
	openbox/screen.c
	openbox/stacking.c
	openbox/startupnotify.c

don't deiconify windows on reconfigure if they cant be iconified directly.  stop managing windows in reverse order on restart it messes up the dock among other things

don't check for the root window in the window hash tables. waste of cpu

be explicit than automake 1.9 is needed

Fix a typo that broke transient window hints.

increment the library version 1 so that it differs from the backport one, as they are actually different, and overwriting it sucks for obconf

don't assert when you window_find(0)/dockapp_find(0), just return 0

Use the # directive to simplify defines.

put openbox saved sessions in CACHE_HOME not DATA_HOME.  more correct! i think.

simplify window.h macros, and rename the window types to ObWindowClass for consistent name style

make an obt_root() macro

xerror.c is gone

its OBT_VERSION not OB_VERSION anymore

add support for XDG_CACHE_HOME to the obt paths

make menu event handling work more like how other events are handled, less special-casey

use -ggdb not -g3, as g3 breaks things

make the window.h interface more consistent with the rest of openbox, hide the window_map behind some functions, and make obdockapps not obwindows anymore

make the obt library branch compile again with all the changes merged in from backport

let you make an xevent listener for all windows

make the xevent stuff work for events from extensions (these go beyond LASTEvent)

add the xevent callback system to libobt in xevent.[ch]

make the mainloop not depend on an X display, and make it uses the obt_display automatically

initialize the lists to NULL (empty)

move the xdg path stuff into obt/paths.[ch], and make render and openbox use it

small api change for parsing

was using random memory for parsing pipe menus with the new parse api

fix constness

s/xmlChildrenNode/children/

add obt_parse_tree_from_root and use it, cuz it's nice

update openbox to use the current parser interface in libobt

there is also some random bug fixes for other libobt stuff in here.

add parsing to libobt with a new API! (it is not completely changed/fixedup yet)

change openbox to use the libotk's keyboard.[ch] stuff instead of the old modkeys.[ch] and some of translate.[ch]

add keyboard.[ch] to libobt. these are for tracking modifier keys, and translating to/from keycodes/keysyms/modmasks (also include a new obt/internal.h for putting functions that are not public)

add functions for setting locale strings (on top of existing functions for setting utf8 strings). they are untested though! wonder if they work?

ignore the new 4.0 pkgconfig files

make openbox use the obt prop.c/h stuff

holy search and replace batman

s/ob_display/obt_display/ and remove ob_display

add prop.[ch] to the libobt, but they are not used yet. add a global obt_display which obt can use, and the application.

setup versioning for libotk that is separate from libobrender (and libobparser for now), and bump library versions up to 4.0.

also make the render versioning more explicit that it's for the render library not all of openbox

move the code for tracking what extensions are available on the display into the libobt

add G_BEGIN/END_DECLS to the mainloop header

move the xerror handling into the libobt

Move the main loop out into the libobt

more precise includes

split the header up for obt

look mom, I made an openbox toolkit

Merge branch 'backport'

Merge branch 'backport'

Merge branch 'backport'

Add complementing ShrinkToEdge action.

Merge branch 'backport'

use clicks for mouse wheel, not press.

remove old unused variables

smarter mouse wheel default actions

bump the library version

Update the new strings in the french translation

add the per-app position stuff to the changelog

change the <execute> to <command> in the mouse focus example

add the "command" option to the xsd

use <command> for the execute actions in menu.xml

add an 'all' option for the monitor in MoveResizeTo

for MoveResizeTo make <width>current</width> and <height>current</height> and <monitor>current</monitor> work like <x> and <y>

update line numbers for the po files

for MoveResizeTo, make <x>current</x> the same as not specifying <x/> at all

update changelog for last few commits

fix a crash with the focused client closing during the delay of a submenu being shown (and add an assert for this too)

don't focus new windows when the user is in a menu or move/resizing a window

kill moveresize cleanly if focus is moving (it uses a keyboard grab)

Merge branch 'master' of git://orodu.net/openbox

Update spanish translation, and make translations a single changelog entry

Update portuguese translation

remove mysterious xkb code...

be explicit about the event

cleanup

more updates for the git instruction

Merge branch 'master' of git://mikachu.ath.cx/openbox

update changelog for translation updates

Updated Arabic translation

Updated Czech translation

update for the new autopoint

don't resist in orthoganal directions when key resizing windows with aspect ratios

update changelog with recent changes

don't move parents and children over when inserting desktops - cuz then they all get moved a bunch of times, just move the top direct parent

remove a debug print

let you resize aspect ratio windows in north and south directions, and make them resist properly when resizing

the resistance resizing has been changed from "lock corner" to "the direction being resized" which makes so much more sense and doesn't need all this conversion back and forth

don't crash from very short windows

Merge commit 'mika/master'

add new bindings to the changelog

xml values for per app settings should be case insensitive
change strcasecmp to g_ascii_strcasecmp

update italian translation

changelog for 3.4.3

don't resist when resizing terminals with the keyboard

fix growtoedge so it doesn't move terminal windows around

missing header

decorate actions were backwards

make the master branch's version very unique and fun

use keypress for navigating the menu but use keyrelease for running things

use keyrelease to fire menu actions not key press

when keypress is used, the keyrelease gets sent to whatever window is focused.  if it's a stupid program - like most any program - it will update its user_time, preventing the new window from being focused

ignore .libs and .deps dirs

bump version

don't restack windows when they request it with an invalid sibling

take care when sibling is specified as an unmanaged window

Grab the server when mapping a client and frame, to avoid race conditions

add TAGS to .gitignore

remove trailing whitespace

don't make moving terminals jump at the start

Require libstartup-notification 0.8

Don't cancel interactive action when another one is run which shares the same run function.  This is how actions can cooperate!

update the default rc.xml with the new actions syntax and the rc.xsd

Merge branch 'mika'

a number for placing windows, to keep them on screen, but let you place them off-screen

1. when a window is not on any monitor, don't just ignore it, stick it on the first monitor arbitrarily (rather than the closest monitor - room for improvement?)
2. when placing a window with per-app settings, don't rudely force it back on screen fully.

don't use freed memory as the coordinates to place windows

was incorrectly using the width for calculations instead of the height in client_find_onscreen

Update italian translation from Davide Truffa

previous should not mean the same as last

no splitting _() strings

Merge branch 'master' of git://orodu.net/openbox

80 cols everywhere

80 cols in dock.c

make .gitignore more sane

update polish translation by Piotr Drąg <raven@pmail.pl>

ok i take that back

i know c...honest.

add hardcoded key binds back

fix directionalcycle/target

the menu.c was renamed

move dockautohide back to ToggleDockAutoHide

moev showdesktop back to ToggleShowDesktop

move menu action back to ShowMenu

move directionalcycling actions into the same .c file. make interactive actions all kill delayed mouse focus which would kill the action.

make cyclewindows action into nextwindow and previouswindow again

rename switchtodesktop to gotodesktop.  make stuff compile as well.

move desktop and directionaldesktop actions to SwitchToDesktop and SendToDesktop, which take a <to> parameter saying where to go. (last, left, right, next, 5, etc)

update hardcoded bindings to previous changes

fix up the sendtolayer actions

split layer action back into togglealwaysontop, togglealwaysonbottom, and sendtolayer (which takes <layer> as an argument)

redo the maximize actions into maximize/unmaximize/togglemaximize with a <direction> argument

let you specify up/down as well as top/bottom for resizerelative action - backwards compat

let you specify <command> or <execute> for execute actions - backwards compat

split decorations into decorate/undecorate/toggledecorations

rename fullscreen back to togglefullscreen

don't move focus if the window appears or disappears

make omnipresent back into toggleomnipresent

adding comments

split shade back into shade/unshade/toggleshade

use dock_get_area instead of accessing dock struct directly

split out edge detecting code a bit so it is easy to add the dock area and do that too

change dock to use a Rect area instead of four ints

removing composite.h

fix building without startup notification

Signed-off-by: Dana Jansens <danakj@orodu.net>

renaming the readme for git-ish

Signed-off-by: Dana Jansens <danakj@orodu.net>

update dutch translation and mark translations with desktop strings as 3.4.3, the rest as 3.4.2

only add group transients when you have a group

Signed-off-by: Dana Jansens <danakj@orodu.net>

Remove composte.c

Signed-off-by: Dana Jansens <danakj@orodu.net>

add the .gitignore file and update the po files

Signed-off-by: Dana Jansens <danakj@orodu.net>

adding trunk
Signed-off-by: Dana Jansens <danakj@orodu.net>

clear the fuzzy strings and mark translations without the desktop strings as 3.4.2

update dutch translation

fix grow to edge stuff. also fix key-resizing terminals.

fix the remaining reversed actions_client_move

only end the interactive action if it's the action you just ran

*^^*

cancel the interactive action only if runing another one. i think this will work?

oops

fix if action's "focused"

cycleinwdows defaults to forward

clever code by mika-kun

finished rev 7756. renaming moveto->moveresizeto

renaming movetofromedge->movetoedge and moveto->moveresizeto

let moveto action resize

you can use -'s for opposite sides of the screen in moveto

missing header

add focused check to if action

fix focus cycle indicator positioning

create actions from string in a way that won't crash

add default bindings for the side edges

no more crash in splitvertical. i think i changed the size of the middle gradient too. it's always 1px now unless the texture is very small, then it can be 0. if theres a discrepency, the bottom color gets the extra pixel.

don't include debian dir in releases

backwardsness

child/parents don't get included in tests for occluding/occluded windows, because they always will be above/below eachother!

comment

fix for weird group transient crap.

  2 group transients
+ 1 direct transient of one of the above
--
the other group transient would be a child of the direct transient.

it's friggin annoying to look for this. i had to destroy my super clever but
impossible to maintain client_update_transient_tree in order to do it, and add
redundant execution, where children update their transient trees when a parent
does.

merge r7699 from 3.4-working, (let per-app settings override program specified position but not user specified)

remove debug print

user-sepcified margins

mssing header

merge r7694 from 3.4-working (note you can use negative positions for starting at the opposite edge)

let you specify shortcuts in your menus. change & to be _ for shortcut selection

checking the wrong extension! also 2 seconds maybe is better.

resize fixings

or a full second?

only wait 1/2 a second for sync reply

fancy resizing

move keyboard moving into its own function

grow to edge works again, using the new edge detection code

hm..

cancel interactive actions when anothr action runs properly

don't focus helper windows that map unless there are no other valid targets in the group (don't move focus from a group window to a new helper window).
this fixes gucharmap.

comment

move to edge works and the code can be used for resize now too, yay

rename <actions> to <finalactions> for the action i didn' know existed too

rename <actions> to <finalactions>

rename <execute> to <command>

desktop strings for swedish

add desktop editing strings to japanese

updated japanese translation

don't fallback to shaded windows

compliance over functionlity

rewrote the movetoedge code so it works with both types of edges (to edge and from edge)

merge r7664 from 3.4

cleanup

debug print

remove the highlight action. its not useful without event bindings

add a highlight action

don't focus docks when you click on them

typo

add startupnotify for obconf in the default menu

add shortcuts to add/remove desktops in the client list menus.

also make it so &-specified shortcuts are always underlined

warp desktops when you hit the edge of the screen while moving a window

missing header

move logic for add/remove desktop into screen.c

action.h not needed here

remove the old action files

put back default hardcoded mouse bindings. remove default key bindigns

add the "if" action

remove excessive debug print

add the growtoedge action

add moveto/fromedge actions

add the layer action

add dockautohide action

add the adddesktop and removedesktop actions

let you send windows with the desktop and directionaldesktop actions

always reconfigure the window to its new position even if it was fullscreened/maxed, because the pre-max/fs area may not match its actual current area, and this can lead it things not being updated

add resizerelative action

let you use "last" in desktop action

add directionaldesktop action

add the desktop action

rename showmenu to menu

add the decorations action

add the resize action

add the directionaltargetwindow action

add the directionalcyclewindows action

add omnipresent action

add the kill/close/cyclewindows actions

nuthin

add max horz/vert actions. add toggle/on/off ability to all the maximize actions.
add shade action.

add moverelative action

add the moveto action

add horiz/vertical options to the maximize action

add maximize action

add the fullscreen action

add the iconify action

add the unfocus action

add lower and raiselower actions

add raise action

remove the activate action, make the focus action "activate" but not raise or unshade

add the move action

add breakchroot action

add the activate action. it will replace the focus action, as it can just focus without raising now (or without unshading)

add the cyclewindows action

add restart action

add exit and reconfigure actions

add the showdesktop action

add the showmenu action

added the debug action

the execute action works again.

default hardcoded bindings are temporarily offline

well.. it compiles..

add code for interactive actions

add interactive action functions. some other changes to stuff that wasnt going to work

make mouse use the new action stuff

dont build the old action stuff.
config uses the new action stuff.
add actions_parse_string

maybe the new actions framework is kinda there now

some first structural stuff for new actions

first draft of beginning of new action code

ok the thumbnails in alt-tab were fun and slow so remove them

don't activate on raise

except without bugs

look, thumbnails in alt-tab. its such crap though. oh well, just for fun.

clearer code paths

always have the offscreen buffer's shape matched to the window's

don't call framerender directly

only enable automatic composite redirection for the client windows, not the frames. maybe this will avoid screen artifacts with the crap server support

i am starting to loathe the composite extension, and xorg in general

80 cols

i forgot to add these

get the offscreen pixmaps for all windows (including their frames)

backwards macro

check for damage and render extensions if going to use composite. require them for it. also include the cflags/libs. and only run the xcomposite stuff so far when the server supports it

redirect window contents to offscreen pixmaps with composite so they are available for use.

add the composite variables

check for composite on the server

check for the composite library

don't resize windows when they are user size _or_ user positioned

add ukrainian translation

fade out disabled buttonsx

don't set the default icon on child windows, their parents will have an icon

let you lower a window without lowering its children, or its siblings

fix move to center on xinerama

always show the client on desktop change, only use donthide to not hide it

raise windows smartly when merging desktops

add center option to placement section

little thing

dont fallback to undermouse

improve the add/remove desktop actions a bunch.
make the lastdesktop action not assert if you remove a desktop (i think it could have)

only send configurenotify for user-action configures when it moves

add force_reply back to client_configure(). client_reconfigure is nicer this time around. eat enter events in client_configure when user is FALSE.

fix focus flicker on sendto+follow

remove debug print

don't raise windows when moving them between desktops in some cases

fix moved comment

don't force reconfigure in a number of places, also eat enter events when reconfiguring. yay

add AddDesktopLast RemoveDesktopLast AddDesktopCurrent RemoveDesktopCurrent

let desklets snap to eachother

don't snap to below+skiptaskbar (desklets)

kill the side borders on the handle and titlebar when max_horz

let the theme specify all splitgradient colors

don't alloc/free colors every time splitvertical is drawn

that got commited by accident somewhere

no un-needed rendering

don't show vertical resize context/arrow when the window is shaded

fix the optimize

optimize

wrong vriables

add struts to the list. reboringify --debug-xinerama

fix some more crashing. and make the max area selection more correct to avoid dead areas

initalize static stuff.
make --debug-xinerama more intersting..

wrong variable

backwards args

fix another crash

fix 1 crash.

xinerama support like crazy for struts and everything else too. this probably crashes but its been way too long without a commit.

set workarea correctly

dont unshade when clicking the titlebar.
lower when rolling up, raise when rolling down with mouse wheel

ignore the dock in placing as a last resort

use all possible struts when maximizing to fill the screen

fix how the dock is placed, its struts, and set a window type hint on it

bugfix

add the dock struts to everything

some typos

plug memory leaks from the new strut code

very cool struts. partial struts actually are partial struts now. possibly way broken with xinerama?

comment

ratios are floats

use the same decision code to focus new windows as for focus cycling or focus fallback. yay consistency.

include padding in the menu header min width

only xungrabkeyboard when we actually have a passive grab on it

respect max/min sizes when fullscreen or maximized

add debug print

stop ignoring enters on window hide, and explain why

like unmap, don't  let enter events move focus when a window iconifies unless underMouse is on

dont use enter events from new windows appearing to focus them, unless underMouse is on

i dont think people shade windows with the intention of having them hidden by stuff. also bug fix.

try place under the mouse when theres no free space? maybe thats dumb

basically this is smart placement but its better now..

new window placement, but i was a lil tipsy when i started this so maybe its not so good. smart placement was just as useless anyways though, so yeah. and this code is much cleaner. on to stealing fvwm code now.

redundant if'ing

when undecorated and keepborder is on, show only the outside border not the clientborder

instead of falling back to any normal window. fallback to the same windows you can focus cycle to.

update changelog

bump version. fix library versions

put new things at the bottom of the struct

let the shadow be 0-256

don't limit the highlight at all

fix a debug print

make bevel strengths theme-selectable

move focus consistantly when changing desktops with different mouse-focus options - while still being not annoying with omnipresent windows

cache pipe menus without crashing!

make um.. the resizerelative action work right and also not move the window when resizing it and it cant get any smaller/bigger

cache pipe menus until the menus close

remove debug prints

allow noninteractive directional focus.

make keybinds noninteractive when the last key in the chain has no modifiers (i think this is what the code was supposed to do but failed.)

don't reset key chains during an interactive action. reset them after an interactive action finishes (honoring chroot)
so you can <C-t><C-Up>Stuff<//> and it'll do stuff as you hit C-Up, you dont have to keep hitting C-t

frigging gtk frig.

i think that other emacs fix was "the right one" and this one where we don't always send configurenotifies can go away. that should fix the xemacs stall thing.

copy/paste :x

fix resizerelative for right/bottom edges

remove SCIM by default. its caused people problems

delay re-rendering the frame while it is being iconify-animated until the animation finishes (avoid it flashing to unfocused)

use the osd theme stuff for the dock instead of active titlebar

update changelog

make client_under_pointer only count windows on the visible desktop. its called halfway thru showing/hiding windows on a desktop change when focusLast is off

let you use +-+-x in the positions for per-app settings, like x geometry

update changelog for 3.4.1 so far

update onyx themes

let you focus cycle to helper windows if there are only dialog windows around in its group (only normal type windows count as siblings for this case)

only configure the window after it sets WM_NORMAL_HINTS if it is outside of its new bounds

mismerged the 3.4 xsd stuff into trunk.

add shadows on inactive windows

wrongly indented line

revert change to theme.. *^^*

fix when 1or2px splitvertical

no crashing in splitvertical

middle gradient in splitvertical

small tweak for nonfocused windows? i dont think they are perfect yet.

better buttons for inactive windows - no hover and bright lightup pressed

add onyx themes to install. make non-active stuff grey

add Onyx themes!?!

add a hover for pressed

yay. toggled textures and colors for all states.

revert the change to frame.c

missed a spot

back xml themes out of trunk

bump version

stop processing x events or actions when loop->run becomes false

don't reconfigure twice

remove {} just cuz

fix dirfocus when no dialog

why is this function declared twice

let you lower siblings below eachother

better.

c comments

i dont think that line is valid, if it ever did run

put the directional focus dialog on the active monitor

make popup dialogs appear on the active monitor with xinerama

forgot to change the version string in the header

updated japanese translation, ui visible strings only

update vietnamese translation and rename to just vi

add debug prints

if Key_R is set to multiple mods, then it'll use them all now. but Key_L still takes precidence if they are different

rather than making you hit both Key_L and Key_R if they are bound to different things, make Key_L take precidence. But if only Key_R is bound then use that.  If Key_L is bound to multiple masks it'll use them all, but that won't work for Key_R right now.

get rid of client->transient_for. just client->parents and client->transient_for_group now.

transient_for is now only used to check for OB_TRAN_GROUP

more cleanup

more cleanup

nice code cleanup that's been needed for a long time. add parents list to client, which you can iterate instead of going thru the group.

add keyboard accels

add vietnamese translation

fallback under the mouse when config_focus_under_mouse

one more..

better group_full func

check if they have a group

make active-disabled shadow offset-y be proper

better code for the last commit. let transients get focus when a group member is focused too.

let transients take focus when anything in the transient tree _or the group_ is focused

remove some debug prints

fix parentrel bevels

add modifier changes

ignore control_l/r shift_l/r and caps_lock bindings. Control, Shift and NumLock are hardcoded to the modmasks for them.

fix for capslock bound to other modifiers. i think this is okay?
strip LockMask instead of what caps_lock is bound to.

combine the parent and child textures in fun ways when a parentrelative texture with a bevel is placed right overtop of another texture with a bevel

typo from code i removed

don't force clients on screen when restarting

add some padding inside the text box

make parentrelative raised labels look nice in popups

revert that, center the popups for move/resize and keychains

center text in popups when the popu is sized to the text width

add comments

use the label only for popups without an icon, if the label is not parentrelative

add osd theme stuff

border, not borderwidth.
add osd.border stuff to themetoxml

add osd border stuff

add osd theme elements

and interlacing!

reset interface age, cuz we added something

allow users of obrender to change the ellipsize mode of text textures

update changelog

3.4.0

dont place transients based on non-normal parents

dont use non-normal windows to pick a monitor for transients

allow focus to fallback to omnipresent windows, except during desktop switching which is when they are vile

don't copy the desktop from group parents when they are omnipresent

don't change the focus order when a client tries to focus but is not visible. client_activate will -make- it visible anyways.

fixes focus fallback's loop totally breaking (stopping on 32bit and looping forever on 64bit) when it encountered a non-visible window

update changelog

dont let windows place over menu or toolbars if they don't have a parent

don't place windows over docks. but let them place over menus and toolbars. i guess not utility windows though?

unused variables

placement improvements?

when placing transients don't place them relative to iconified parents.
try place avoiding max/fullscreen windows first, then ignore them entirely if that fails.

when placing transients in the stacking order, only use relatives that are on screen or share the desktop or whatever, not iconic ones

update my theme

fix stacking_add_nonintrusive for when only 2 windows, and one is moving below the focused window (fullscreen losing focus)

actually allow raised, but default to flat for parentrelative

allow effects on parentrelative textures, ie bevels and interlaces and borders

add --restart translation

when a bool field is empty return false, not signal 11.. :|

when an integer field is empty return 0, not signal 11

add --restart string to czech translation

revert a committed typo

when undermouse is off, ignore enter events from windows lowering/raising themsvelves (or from a pager restacking them)

update italiano translation

unused variables

update german translation

add 3.4 changelog stuff

much simpler and works just as well! client's reference position doesnt change just because the decor sizes do

remove fuzzy strings

remove the 3 badly out-of-date translations from LINGUAS (they wont be in releases for now)
removing hr.po Croatian
removing it.po Italian
removing ja.po Japanese

utf8

restart norway

fix a missing &

don't use the title separator color for shaded windows

make focuslast default to true

add underMouse option

support serials wrapping around

super amazing enter event skipping.

also mucho fixed for <underMouse>

add <underMouse> focus option

return if the window was restacked

missing theme->

free the new title separator colors

revert 7283 until we find out what version they are using

always ungrab the keyboard/mouse with current time because i dont know what else we can do with buggy x.org

there add new separator colors complete

add new separator colors

consistancy

remove the title separator width thing.
add a separate color for the title separator

use the new window.title.separator.width

add tswidth to theme.h, but its always == fbwidth

do the menu_can_hide thing properly. was such a hack before? and the variable was getting set from all kinds of events like motion notify. totally not reliable. now it works all nice.

rename activedisabled in xml to active-disabled

rename old xrbd activedisabled to active.disabled

better fix for focus going to the frame window when iconifying with RevertToParent. Treat the focus in like a focus out.

When focus moves to a window that doesn't exist, need to set the focus_client to NULL otherwise the client that lost focus won't know and won't update its decor

catch focus out events on the frame window - this happens with revert to parent when a window is iconified. we get focusin on the frame without a focus out - which then makes us fallback. but catch the resulting focusout on the frame to make the window lose its focused decor

when finding edges and we're omnipresent only use windows on screen

change which windows are used for finding edges. its the same as window resistance really now. and works if the window isnt on the current desktop

smrts.

compare windows for restacking based on their desktops and if they are iconified. dont compare to irrelavent ones. also if the sibling isn't valid then don't use it.

only do restack checks on windows that are visible

revert 7246 and simplify action_raiselower a bit, turns out 7246 wasn't needed for that after all

formatting

have stacking_restack_request return a bool that says if it did something with TopIf/BottomIf (other cases always return true)