3 - generate pointer_focus on raise/lower, move windows, all kinds of
4 changes in surface stacking.
8 - dnd, figure out large object transfer: through wayland protocol or
9 pass an fd through the compositor to the other client and let them
12 - copy-n-paste, store data in server (only one mime-type available)
13 or do X style (content mime-type negotiation, but data goes away
16 - Discard buffer, as in "wayland discarded your buffer, it's no
17 longer visible, you can stop updating it now.", reattach, as in "oh
18 hey, I'm about to show your buffer that I threw away, what was it
19 again?". for wayland system compositor vt switcing, for example,
20 to be able to throw away the surfaces in the session we're
21 switching away from. for minimized windows that we don't want live
24 - Initial placement of surfaces. Guess we can do, 1)
25 surface-relative (menus), 2) pointer-relative (tooltips and
26 right-click menus) or 3) server-decides (all other top-levels).
28 - When a surface is the size of the screen and on top, we can set the
29 scanout buffer to that surface directly. Like compiz unredirect
30 top-level window feature. Except it won't have any protocol state
31 side-effects and the client that owns the surface won't know. We
32 lose control of updates. Should work well for X server root window
33 under wayland. Should be possible for yuv overlays as well.
35 - what about cursors then? maybe use hw cursors if the cursor
36 satisfies hw limitations (64x64, only one cursor), switch to
37 composited cursors if not.
39 - clients needs to allocate the surface to be suitable for
40 scanout, which they can do whenever they go fullscreen.
42 - multihead, screen geometry and crtc layout protocol, hotplug
44 - input device discovery, hotplug
46 - Advertise axes as part of the discovery, use something like
47 "org.wayland.input.x" to identify the axes.
49 - keyboard state, layout events at connect time and when it
50 changes, keyboard leds
56 - synaptics, 3-button emulation, scim
58 - Figure out if we need the batch/commit scheme and what to do
59 instead. Since dropping the "copy" request, we have a race between
60 copy from back to front and reporting damage. "copy" did this
61 atomically, but copy is a rendering operation (wayland doesn't do
62 rendering) and requires synchronization between server and client
63 before client can reuse backbuffer.
65 The race condition happens when a client copies new content into
66 its window and then, before the client reports the damage, the
67 compositor then does a partial repaint (triggered by another
68 client) that only pulls in part of the repainted area. It's only a
69 one-frame glitch, as the client will submit the damage and the
70 compositor will repaint the damaged area next frame. And ideally
71 clients should do all rendering as early in the frame as possible
74 - auth; We need to generate a random socket name and advertise that
75 on dbus along with a connection cookie. Something like a method
76 that returns the socket name and a connection cookie. The
77 connection cookie is just another random string that the client
78 must pass to the wayland server to become authenticated. The
79 Wayland server generates the cookie on demand when the dbus method
80 is called and expires it after 5s or so.
82 - or just pass the fd over dbus
84 - drm bo access control, authentication, flink_to
86 - Range protocol may not be sufficient... if a server cycles through
87 2^32 object IDs we don't have a way to handle wrapping. And since
88 we hand out a range of 256 IDs to each new clients, we're just
89 talking about 2^24 clients. That's 31 years with a new client
90 every minute... Maybe just use bigger ranges, then it's feasible
91 to track and garbage collect them when a client dies.
93 - Add protocol to let applications specify the effective/logical
94 surface rectangle, that is, the edge of the window, ignoring drop
95 shadows and other padding. The compositor needs this for snapping
96 and constraining window motion. Also, maybe communicate the opaque
97 region of the window (or just a conservative, simple estimate), to
98 let the compositor reduce overdraw.
100 - multi gpu, needs queue and seqno to wait on in requests
106 - eek, so much X legacy stuff there...
108 - draw window decorations in gtkwindow.c
110 - start from alexl's client-side-windows branch
112 - Details about pointer grabs. wayland doesn't have active grabs,
113 menus will behave subtly different. Under X, clicking a menu
114 open grabs the pointer and clicking outside the window pops down
115 the menu and swallows the click. without active grabs we can't
116 swallow the click. I'm sure there much more...
118 - Port Qt? There's already talk about this on the list.
122 - move most of the code from xf86-video-intel into a Xorg wayland
125 - don't ask KMS for available output and modes, use the info from
126 the wayland server. then stop mooching off of drmmode.c.
128 - map multiple wayland input devices to MPX in Xorg.
130 - rootless; avoid allocating and setting the front buffer, draw
131 window decorations in the X server (!), how to map input?
133 - gnome-shell as a wayland session compositor
135 - runs as a client of the wayland session compositor, uses
136 clutter+egl on wayland
138 - talks to an Xorg server as the compositing and window manager
139 for that server and renders the output to a wayland surface.
140 the Xorg server should be modified to take input from the system
141 compositor through gnome-shell, but not allocate a front buffer.
143 - make gnome-shell itself a nested wayland server and allow native
144 wayland clients to connect and can native wayland windows with
145 the windows from the X server.
147 - qemu as a wayland client; session surface as X case
149 - qemu has too simple acceleration, so a Wayland backend like the
150 SDL/VNC ones it has now is trivial.
152 - paravirt: forward wayland screen info as mmio, expose gem ioctls as mmio
154 - mapping vmem is tricky, should try to only use ioctl (pwrite+pread)
156 - not useful for Windows without a windows paravirt driver.
158 - two approaches: 1) do a toplevel qemu window, or 2) expose a
159 wayland server in the guest that forwards to the host wayland
160 server, ie a "remote" compositor, but with the gem buffers
161 shared. could do a wl_connection directly on mmio memory, with
162 head and tail pointers. use an alloc_head register to indicate
163 desired data to write, if it overwrites tail, block guest. just
164 a socket would be easier.
166 - moblin as a wayland compositor
168 - clutter as a wayland compositors