Qemu - the loop to initialize the system color map takes two seconds over here.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15149 a95241bf-73f2-0310-859d-f6bbb57e9c96
an app_server code. It's always up-to-date as it just reads the header
directly.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15148 a95241bf-73f2-0310-859d-f6bbb57e9c96
ServerApp now waits up to 3 seconds for windows before killing them - it now
waits on the death semaphore, and only kills a window if it didn't quit fast
enough.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15147 a95241bf-73f2-0310-859d-f6bbb57e9c96
The mouse filter needs to lock RootLayer, but the event dispatcher lock is held
during that time, too, so that the focus cannot change in the mean time.
On the other side, ServerWindow needs to lock the event dispatcher when adding
or removing a listener, or for AS_GET_MOUSE - but since it always helds the
root layer lock during message dispatching this easily resulted in a deadlock.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15146 a95241bf-73f2-0310-859d-f6bbb57e9c96
sends a SIGSEGV signal (and lets the debugger handle the rest).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15145 a95241bf-73f2-0310-859d-f6bbb57e9c96
application got interrupted, thread_at_kernel_exit() was called - but
that expected interrupts to be enabled for signal handling.
However, only exceptions 3 (breakpoint) and 99 (syscall) are trap
gates, and thus, only those actually had interrupt enabled at that
point.
If a KILL signal was pending when a hardware interrupt interrupted a
user space thread you were entering KDL before ("acquire_sem() called
with interrupts turned off").
Thanks to mouse interrupts I finally got this often enough to find
this...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15143 a95241bf-73f2-0310-859d-f6bbb57e9c96
to match the name of the application's port for that task.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15142 a95241bf-73f2-0310-859d-f6bbb57e9c96
* now with actual view layers inside the windows
* implemented the resize modes (from Adis code)
* windows have resize handles and more correctly
clip the views inside
* bringing windows to front or sending them behind all others
* one active window, the others are inactive
* with and without focus follows mouse mode
* bugs:
- the region marked dirty when views are
resized is not correct yet
* todo:
- move the dirty region from being managed by the
desktop to being managed in each window (and being
local too)
- scrolling
- hiding/showing of windows and views
I plan to extend this to fully simulate asynchronous
drawing from clients, to see any problems before
using this in the real server one day.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15132 a95241bf-73f2-0310-859d-f6bbb57e9c96
Layer no longer knows anything about its subclass WindowLayer.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15131 a95241bf-73f2-0310-859d-f6bbb57e9c96
was no valid "_view_token" in the mouse event.
* _SanitizeMessage() will now only add the "be:transit" field for
B_MOUSE_MOVED messages. It will also now only update fLastMouseMovedView
for B_MOUSE_MOVED messages, and not for all mouse messages.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15124 a95241bf-73f2-0310-859d-f6bbb57e9c96
a bit:
* listeners are now managed per target (ie. per messenger) (by the Target class)
* therefore, the fFocusTokens/fLastFocusTokens lists are no longer needed; every
Target knows its listeners already.
* this also fixes the obvious bug that the focus window's views would get both
keyboard and pointer events, no matter which of them they originally wanted.
* renamed event_target to event_listener (there was actually a mix up in naming
before - to the outside it was "listener", and internally, "target" was used)
* WinBorder::MouseMoved()/MouseUp() now also add the view token to the message;
the messages weren't received by the target before (unless the view used
tracking via BView::SetMouseEventMask()...) - maybe the client should only
update fLastMouseMoved on B_MOUSE_MOVED events, and direct B_MOUSE_UP/DOWN
(plus wheel changes) always to fLastMouseMoved...
* some cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15123 a95241bf-73f2-0310-859d-f6bbb57e9c96
client's token space.
* finding layers by token is now O(1) instead of O(n) (as they are looked
up in the token space).
* removed Layer::FindView() as it's no longer needed (or should be used).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15113 a95241bf-73f2-0310-859d-f6bbb57e9c96
NextChild() and PreviousChild() - the current WinBorder list is now rebuilt on every
change; this is not perfect, and only a temporary solution (but cleaner than the
previous one).
* Introduced Layer::PreviousLayer()/NextLayer() methods that return the previous resp.
the next sibling.
* Moved {show|hide}_winBorder() into {Show|Hide}WinBorder() and got rid of the former.
* Renamed Layer::fServerWin to fWindow.
* removed some unused stuff, minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15108 a95241bf-73f2-0310-859d-f6bbb57e9c96
