3 - The message format has to include information about number of fds
4 in the message so we can skip a message correctly. Or we should
5 just give up on trying to recover from unknown messages.
7 - generate pointer_focus (and drag focus) on raise/lower, move
8 windows, all kinds of changes in surface stacking.
10 - make a client side circular buffer of pending ping requests with
11 callbacks and data. if buffer fills up, just iterate until an
12 entry becomes available. wl_display_ping(dpy, func, data), basically.
13 func is called when the reply comes in for the ping request.
17 - dnd, figure out large object transfer: through wayland protocol or
18 pass an fd through the compositor to the other client and let them
23 How to handle drop decline (accept with type=NULL)
25 - Targets must send a NULL type in accept if they don't accept a
26 drop at the drag_focus/drag_motion position. Root window will
27 send a NULL type or x-wayland/root-something type if the source
30 How do we animate the drag icon back to the drag origin in case of
33 How to handle surfaces from clients that don't know about dnd or
34 don't care? Maybe the dnd object should have a
35 dnd.register_surface() method so clients can opt-in the surfaces
36 that will participate in dnd. Or just assume client is not
37 participating until we receive an accept request.
39 - Pointer image issue:
41 - A touch input device doesn't have a pointer; indicate that
46 - copy-n-paste, store data in server (only one mime-type available)
47 or do X style (content mime-type negotiation, but data goes away
50 - Discard buffer, as in "wayland discarded your buffer, it's no
51 longer visible, you can stop updating it now.", reattach, as in "oh
52 hey, I'm about to show your buffer that I threw away, what was it
53 again?". for wayland system compositor vt switcing, for example,
54 to be able to throw away the surfaces in the session we're
55 switching away from. for minimized windows that we don't want live
58 - Initial placement of surfaces. Guess we can do, 1)
59 surface-relative (menus), 2) pointer-relative (tooltips and
60 right-click menus) or 3) server-decides (all other top-levels).
62 - When a surface is the size of the screen and on top, we can set the
63 scanout buffer to that surface directly. Like compiz unredirect
64 top-level window feature. Except it won't have any protocol state
65 side-effects and the client that owns the surface won't know. We
66 lose control of updates. Should work well for X server root window
67 under wayland. Should be possible for yuv overlays as well.
69 - what about cursors then? maybe use hw cursors if the cursor
70 satisfies hw limitations (64x64, only one cursor), switch to
71 composited cursors if not.
73 - clients needs to allocate the surface to be suitable for
74 scanout, which they can do whenever they go fullscreen.
76 - multihead, screen geometry and crtc layout protocol, hotplug
78 - input device discovery, hotplug
80 - Advertise axes as part of the discovery, use something like
81 "org.wayland.input.x" to identify the axes.
83 - keyboard state, layout events at connect time and when it
84 changes, keyboard leds
90 - synaptics, 3-button emulation, scim
92 - Figure out if we need the batch/commit scheme and what to do
93 instead. Since dropping the "copy" request, we have a race between
94 copy from back to front and reporting damage. "copy" did this
95 atomically, but copy is a rendering operation (wayland doesn't do
96 rendering) and requires synchronization between server and client
97 before client can reuse backbuffer.
99 The race condition happens when a client copies new content into
100 its window and then, before the client reports the damage, the
101 compositor then does a partial repaint (triggered by another
102 client) that only pulls in part of the repainted area. It's only a
103 one-frame glitch, as the client will submit the damage and the
104 compositor will repaint the damaged area next frame. And ideally
105 clients should do all rendering as early in the frame as possible
108 - auth; We need to generate a random socket name and advertise that
109 on dbus along with a connection cookie. Something like a method
110 that returns the socket name and a connection cookie. The
111 connection cookie is just another random string that the client
112 must pass to the wayland server to become authenticated. The
113 Wayland server generates the cookie on demand when the dbus method
114 is called and expires it after 5s or so.
116 - or just pass the fd over dbus
118 - drm bo access control, authentication, flink_to
120 - Range protocol may not be sufficient... if a server cycles through
121 2^32 object IDs we don't have a way to handle wrapping. And since
122 we hand out a range of 256 IDs to each new clients, we're just
123 talking about 2^24 clients. That's 31 years with a new client
124 every minute... Maybe just use bigger ranges, then it's feasible
125 to track and garbage collect them when a client dies.
127 - Add protocol to let applications specify the effective/logical
128 surface rectangle, that is, the edge of the window, ignoring drop
129 shadows and other padding. The compositor needs this for snapping
130 and constraining window motion. Also, maybe communicate the opaque
131 region of the window (or just a conservative, simple estimate), to
132 let the compositor reduce overdraw.
134 - multi gpu, needs queue and seqno to wait on in requests
140 - eek, so much X legacy stuff there...
142 - draw window decorations in gtkwindow.c
144 - start from alexl's client-side-windows branch
146 - Details about pointer grabs. wayland doesn't have active grabs,
147 menus will behave subtly different. Under X, clicking a menu
148 open grabs the pointer and clicking outside the window pops down
149 the menu and swallows the click. without active grabs we can't
150 swallow the click. I'm sure there much more...
152 - Port Qt? There's already talk about this on the list.
156 - move most of the code from xf86-video-intel into a Xorg wayland
159 - don't ask KMS for available output and modes, use the info from
160 the wayland server. then stop mooching off of drmmode.c.
162 - map multiple wayland input devices to MPX in Xorg.
164 - rootless; avoid allocating and setting the front buffer, draw
165 window decorations in the X server (!), how to map input?
167 - gnome-shell as a wayland session compositor
169 - runs as a client of the wayland session compositor, uses
170 clutter+egl on wayland
172 - talks to an Xorg server as the compositing and window manager
173 for that server and renders the output to a wayland surface.
174 the Xorg server should be modified to take input from the system
175 compositor through gnome-shell, but not allocate a front buffer.
177 - make gnome-shell itself a nested wayland server and allow native
178 wayland clients to connect and can native wayland windows with
179 the windows from the X server.
181 - qemu as a wayland client; session surface as X case
183 - qemu has too simple acceleration, so a Wayland backend like the
184 SDL/VNC ones it has now is trivial.
186 - paravirt: forward wayland screen info as mmio, expose gem ioctls as mmio
188 - mapping vmem is tricky, should try to only use ioctl (pwrite+pread)
190 - not useful for Windows without a windows paravirt driver.
192 - two approaches: 1) do a toplevel qemu window, or 2) expose a
193 wayland server in the guest that forwards to the host wayland
194 server, ie a "remote" compositor, but with the gem buffers
195 shared. could do a wl_connection directly on mmio memory, with
196 head and tail pointers. use an alloc_head register to indicate
197 desired data to write, if it overwrites tail, block guest. just
198 a socket would be easier.
200 - moblin as a wayland compositor
202 - clutter as a wayland compositors