For some reason I have yet to divine we don't get an initial unmap
event due to the reparent of the energyxt dialog windows and the
"ignore that first unmap" flag does not get cleared because it doesn't
happen. later when the dialog is withdrawn {9unmapped) this is ignored
then when you close the dialog... thus keeping it around as far as e
is concerned.
so to fix this - add a small timeout to clean this flag after a
show/map.
@fix
assume it went to 0 size if removed and already a csd frame window
which is what chomium does going fullscreen - i didnt see this as i
used chromium with system titlebars not its own.
@fix
shaped input doesnt affect or make custom copies of pixel data, thus
precise inside can't work (it can only work if the pixel data is local
- thus textur efrom pixmap is not going to produce valid data thus...
you get the drift). this fixes some odd focus/event things with some
windows that use shaped input only
@fix
just because an item is not the selected one does not mean it should
be disabled. that is not how radios work. you choose 1 of n menu
items... only items you should NOT select (are not available for
selection) should be disabled.
@fix
Add an event handler for E_CLIENT_FULLSCREEN. Remove the client,
destroying the popup and icon region.
This resolves the issue where a client fullscreen request would
leave a rogue icon region and popup if switching to full screen
with some applications.
@fix T8996
it's not right that windows stay shaded, iconified or stacked below
when you use a binding to switch focus like "focus prev" to cycle just
with a plain key. this fixes that
@fix
we get a spurious mouse out if your window is small then told to go
fullscreen which then causes in ponter focus a unfocus event which
causes e to restore window to its non-fulscreen mode which then may
cause a mouse in again if mouse is positioned right which causes a "go
fullscreen now again" and so on... fix this and ignore that mouse out
right after going fullscreen.
@fix
as per icccm - client should remove WM_STATE when withdrawing... and
qt relies on WM_STATE to know if it re-show a window - the property
it itself refused to remove...
e didn't set shape rects for input overflow areas for the resize
handles in theme to work on top of other clients and client areas.
this should make that all work now and make the resize handle area
bigger than it actually looks.
@fix
so - some people have issues if we open devices. why... i don't know,
but add an option to toggle this and be conservative and have it off
by default
@fix
didnt remove the lists with del callbacks that accessed the cfdata
struct to set lisrts to null on del before cfdata was freed...
callback hell. yay.
@fix
Carsten Haitzler
Alastair Poole
Marcel Hollerbach
Christopher Michael