Jump to content

KWin/Screen Edges: Difference between revisions

From KDE Community Wiki
Nalvarez (talk | contribs)
m formatting, mark some class names as monospaced
Mgraesslin (talk | contribs)
Documentation of the new implementation
 
(5 intermediate revisions by 2 users not shown)
Line 1: Line 1:
The screen edge utility is called "Electric Borders" in KWin's Workspace class. There are through the enum <tt>ElectricBorder</tt> (kwinglobals.h) eight borders defined: one for each side and one for each corner.
This provides a new protocol intended to be used by auto-hiding panels to make use of the centralized screen edges. To use it a Client can set an X11 property of type '''_KDE_NET_WM_SCREEN_EDGE_SHOW''' to KWin.


The electric borders are windows kept above all other windows. KWin watches for <tt>EnterNotify</tt> and <tt>ClientMessage</tt> XEvents on the electric border windows in the global event handler <tt>Workspace::workspaceEvent</tt> (events.cpp:228) which calls <tt>Workspace::electricBorderEvent</tt> (workspace.cpp:2621). EnterNotify is used to test whether the mouse entered a border, ClientMessage is used for the D&D case, though this is mostly broken (e.g. using D&D to Present Windows does not work/crashes the application you drag from).
As value it takes (CARDINAL/32, 1 value):


ElectricBorders need to be reserved and unreserved. This can be done from a KWin Effect (effect.cpp:962), though it does not have any information about which KWin effect reserves the border. Several effects could reserve the border. Some electric borders are reserved cause of some settings to e.g. lock screen or activate dashboard.
* 0: top edge
* 1: right edge
* 2: bottom edge
* 3: left edge
 
KWin will hide the Client (hide because unmap or minimize would break it) and create an Edge. If that Edge gets triggered the Client is shown again and the property gets deleted. If the Client doesn't border the specified screen edge the Client gets shown immediately so that we never end in a situation that we cannot unhide the auto-hidden panel again. The exact process is described in the documentation of ScreenEdges. The Client can request to be shown again by deleting the property.


Most of the electric border related code is in workspace.cpp about line 2325. Show/hideElectricBorderWindowOutline() are not related to electric borders, but to Quick Tiling. There is also some code in layers.cpp. In Workspace::propagateClients the WIds of the electric borders are added to the stacking order and there is Workspace::raiseElectricBorderWindows which is needed to re raise the windows if a fullscreen effect window is opened above.
If KWin gets restarted the state is read from the property and it is tried to create the edge as described.
   
As this is a KWin specific extension we need to discuss what it means for Clients using this feature with other WMs: it does nothing. As the Client gets hidden by KWin and not by the Client, it just doesn't get hidden if the WM doesn't provide the feature. In case of an auto-hiding panel this seems like a good solution given that we don't want to hide it if we cannot unhide it. Of course there's the option for the Client to provide that feature itself and if that's wanted we would need to announce the feature in the _NET_SUPPORTED atom. At the moment that doesn't sound like being needed as Plasma doesn't want to provide an own implementation.


Notifying effects, switching desktop and so on is done from Workspace::checkElectricBorder()
The implementation comes with a small test application showing how the feature is intended to be used.
 
== Refactoring ==
All electric border related code should be moved out of Workspace into an own class, e.g. ElectricBorder or more generic ScreenEdge. (Un)Reserving is probably not needed any more, all edges need to be reserved to inform other applications about triggered edges. Desktop switching while moving can probably be dropped from screen edge handling and should be merged with Quick Tiling which does not use electric borders.
 
== Seigo's Proposal ==
 
* Setting a ElectricBorder XAtom on a window would cause KWin to register a watch on the requests area for that window. When the window clears the atom or the window is deleted, then the area is unregistered. These would be called "window reservations" (as opposed to "effects" or "internal" reservations).
* The ElectricBorder XAtom should consist of a screen #, a screen location (specific edge or corner), optional offset and length (except for corners, for which this is meaningless)
* KWin would then manage a collection of reservations for entire or partial sides; as long as there is at least one reservation for a given side, that border is activated. When the pointer moves into the area (or that area is otherwise triggered, e.g. by a keyboard shortcut?), then all reservations that span the entire area or which contain the location due to their offset/length will get notified.
* Two glows (separated by color?) would be offered on mouse proximity. One for windows with edge triggers, and one for effects. This glow would become part of the KWin effects stack and use an input window.
* Window reservations should not be triggered even on full screen applications; in fact, it may be worthwhile to _not_ raise the input only windows over full screen apps if only window reservations are made for that edge.
* When a window reservation is triggered, KWin should set an XAtom on the window notifying it of such. Windows would remain responsible for their own actions, however, as KWin can not know what specific action to take or if the application needs to do some preparation before taking action (e.g. loading content in a window prior to displaying it).

Latest revision as of 13:19, 26 February 2014

This provides a new protocol intended to be used by auto-hiding panels to make use of the centralized screen edges. To use it a Client can set an X11 property of type _KDE_NET_WM_SCREEN_EDGE_SHOW to KWin.

As value it takes (CARDINAL/32, 1 value):

  • 0: top edge
  • 1: right edge
  • 2: bottom edge
  • 3: left edge

KWin will hide the Client (hide because unmap or minimize would break it) and create an Edge. If that Edge gets triggered the Client is shown again and the property gets deleted. If the Client doesn't border the specified screen edge the Client gets shown immediately so that we never end in a situation that we cannot unhide the auto-hidden panel again. The exact process is described in the documentation of ScreenEdges. The Client can request to be shown again by deleting the property.

If KWin gets restarted the state is read from the property and it is tried to create the edge as described.

As this is a KWin specific extension we need to discuss what it means for Clients using this feature with other WMs: it does nothing. As the Client gets hidden by KWin and not by the Client, it just doesn't get hidden if the WM doesn't provide the feature. In case of an auto-hiding panel this seems like a good solution given that we don't want to hide it if we cannot unhide it. Of course there's the option for the Client to provide that feature itself and if that's wanted we would need to announce the feature in the _NET_SUPPORTED atom. At the moment that doesn't sound like being needed as Plasma doesn't want to provide an own implementation.

The implementation comes with a small test application showing how the feature is intended to be used.