I originally thought I would do this inside WindowServer, but let's try to
make it as a standalone app that communicates with WindowServer instead.
That will allow us to use LibGUI. :^)
- If the GInputBox has a parent and the parent is a GWindow, center the
input box window within the parent window. This looks quite nice.
- Stop processing events in a nested event loop immediately after it's
been asked to quit.
- Fix GWidget::parent_widget() behavior for non-widget parents.
This patch adds a simple GMessageBox that can run in a nested event loop.
Here's how you use it:
GMessageBox box("Message text here", "Message window title");
int result = box.exec();
The next step is to make the WindowServer respect the modality flag of
these windows and prevent interaction with other windows in the same
process until the modal window has been closed.
I'm gonna want to have nested event loops sooner or later, so let's not
pollute GEventLoop with things that are meant to work globally.
This patch also changes key events to pass around their modifiers as a
bitfield all the way around the system instead of breaking them up.
This patch adds a GShortcut class. Each GAction can have a GShortcut which
will cause the event loop to listen for that key combination app-globally
and activate the event in case it's pressed.
The shortcut will also be displayed when the action is added to a menu.
Use this to hook up Alt+Up with the "open parent directory" action in the
FileManager app. :^)
This way GWindow doesn't need to do synchronous IPC to fetch the appropriate
size for the window's backing store. This is mostly only relevant during
live resize.
Windows now learn when the mouse cursor leaves or enters them.
Use this to implement GWidget::{enter,leave}_event() and use that
to implement the CoolBar button effect. :^)
These functions don't exit immediately, but rather on the next iteration
of the event loop.
Since exit() is already used by the standard library, let's call it quit()
instead. That way, saying exit() means the same thing here as anywhere else.
This is a monster patch that required changing a whole bunch of things.
There are performance and stability issues all over the place, but it works.
Pretty cool, I have to admit :^)
This is really cool! :^)
Apps currently refuse to start if the WindowServer isn't listening on the
socket in /wsportal. This makes sense, but I guess it would also be nice
to have some sort of "wait for server on startup" mode.
This has performance issues, and I'll work on those, but this stuff seems
to actually work and I'm very happy with that.
This is factored a bit stupidly. It would be nicer to just have the
read() in drain_events_from_server() be blocking, but the fd is opened
with O_NONBLOCK right now.
This makes everything run real snappy once again. :^)
In order to move the WindowServer to userspace, I have to eliminate its
dependence on system call facilities. The communication channel with each
client needs to be message-based in both directions.
GMenu now has an "on_item_activation" callback that fires whenever one
of its items are activated. The menu item identifier is used to distinguish
between items.
Use this to implement font switching in Terminal. :^)
To facilitate listening for action on arbitrary file descriptors,
I've added a GNotifier class. It's quite simple but very useful:
GNotifier notifier(fd, GNotifier::Read);
notifier.on_ready_to_read = [this] (GNotifier& fd) {
// read from fd or whatever else you like :^)
};
The callback will get invoked by GEventLoop when select() says we
have something to read on the fd.
Clicking the button generates a WindowCloseRequest event which the client app
then has to deal with. The default behavior for GWindow is to close() itself.
I also added a flag, GWindow::should_exit_event_loop_on_close() which does
what it sounds like it does.
This patch exposed some bugs in GWindow and GWidget teardown.
GObjects can now register a timer with the GEventLoop. This will eventually
cause GTimerEvents to be dispatched to the GObject.
This needed a few supporting changes in the kernel:
- The PIT now ticks 1000 times/sec.
- select() now supports an arbitrary timeout.
- gettimeofday() now returns something in the tv_usec field.
With these changes, the clock window in guitest2 finally ticks on its own.
Instead of clients painting whenever they feel like it, we now ask that they
paint in response to a paint message.
After finishing painting, clients notify the WindowServer about the rect(s)
they painted into and then flush eventually happens, etc.
This stuff leaves us with a lot of badly named things. Need to fix that.
