3 - Atomicity. Currently a lot of the atomicity in Wayland relies on
4 how we batch up all requests in a protocol buffer and only flushes
5 in the "blockhandler" in the client. Consensus was that we need
6 something more reliable and explicit. The suggestion is that we
7 make surface.attach a synchronization point such that everything
8 before that is batched and applied atomically when the
9 surface.attach request comes in. For cases where we need atomicity
10 beyond a surface.attach, we can add an atomic grouping mechanism,
11 that can group together multiple surface.attach requests into a
12 bigger atomic change. To be researched a bit.
14 - Maybe try to make remote wayland actually happen, to see if there
15 is something in the protocol/architecture that makes it harder than
18 - Add timestamp to touch_cancel, add touch id to touch_cancel (?)
20 - The output protocol needs to send all the ugly timing details for the modes.
24 - mime-type guidelines for data_source (ie, both dnd and selection):
25 recommended types for text or images, types that a clipboard
26 manager must support, mime-types must be listed in preferred order
28 - we need a "no kb focus please" mechanism. Or should this be
29 implicit in a specific surface type?
33 - configure should provide dx_left, dx_right, dy_top, dy_bottom, or
34 dx, dy, width and height.
36 - move to workspace, keep on top, on all workspaces, minimize etc
37 requests for implementing client side window menu? or just make a
38 "show window menu" request to let the compositor display and manage
41 - window move and resize functionality for kb and touch.
43 - Protocol for specifying title bar rectangle (for moving
44 unresponsive apps). Rectangle for close button, so we can popup
45 force-close dialog if application doesn't respond to ping event
46 when user clicks there. We could use the region mechanism here
49 - popup placement protocol logic.
51 - subsurface mechanism. we need this for cases where we would use an
52 X subwindow for gl or video other different visual type.
56 - Don't wl_display_iterate in eglSwapBuffer, send an eventfd fd?
58 - Land Robert Braggs EGL extensions: frame age, swap with damage
60 - Make it possible to share buffers from compositor to clients.
61 Tricky part here is how to indicate to EGL on the server side that
62 it should make an EGLImage available to a client. We'll need a
63 "create a wl_buffer for this EGLImage for this client" kind of
66 - Protocol for arbitrating access to scanout buffers (physically
67 contiguous memory). When a client goes fullscreen (or ideally as
68 the compositor starts the animation that will make it fullscreen)
69 we send a "give up your scanout buffer" to the current fullscreen
70 client (if any) and when the client acks that we send a "try to
71 allocate a scanout buffer now" event to the fullscreen-to-be
78 - Needs a mechanism to pass buffers to client.
80 buffer = drm.create_buffer(); /* buffer with stuff in it */
82 cache.upload(buffer, x, y, width, height, int hash)
84 drm.buffer: id, name, stride etc /* event to announce cache buffer */
86 cache.image: hash, buffer, x, y, stride /* event to announce
87 * location in cache */
89 cache.reject: hash /* no upload for you! */
91 cache.retire: buffer /* cache has stopped using buffer, please
92 * reupload whatever you had in that buffer */
94 - A "please suspend" event from the compositor, to indicate to an
95 application that it's no longer visible/active. Or maybe discard
96 buffer, as in "wayland discarded your buffer, it's no longer
97 visible, you can stop updating it now.", reattach, as in "oh hey,
98 I'm about to show your buffer that I threw away, what was it
99 again?". for wayland system compositor vt switcing, for example,
100 to be able to throw away the surfaces in the session we're
101 switching away from. for minimized windows that we don't want live
102 thumb nails for. etc.
108 - draw window decorations in gtkwindow.c
110 - Details about pointer grabs. wayland doesn't have active grabs,
111 menus will behave subtly different. Under X, clicking a menu
112 open grabs the pointer and clicking outside the window pops down
113 the menu and swallows the click. without active grabs we can't
114 swallow the click. I'm sure there much more...
118 - Investigate DirectFB on Wayland (or is that Wayland on DirectFB?)
120 - SDL port, bnf has work in progress here:
121 http://cgit.freedesktop.org/~bnf/sdl-wayland/
123 - libva + eglimage + kms integration
128 - A wayland settings protocol to tell clients about themes (icons,
129 cursors, widget themes), fonts details (family, hinting
130 preferences) etc. Just send all settings at connect time, send
131 updates when a setting change. Getting a little close to gconf
132 here, but could be pretty simple:
134 interface "settings":
135 event int_value(string name, int value)
136 event string_value(string name, string value)
138 but maybe it's better to just require that clients get that from
139 somewhere else (gconf/dbus).
144 - AF_WAYLAND - A new socket type. Eliminate compositor context
145 switch by making kernel understand enough of wayland that it can
146 forward input events as wayland events and do page flipping in
147 response to surface_attach requests:
149 - ioctl(wayland_fd, "surface_attach to object 5 should do a kms page
152 - what about multiple crtcs? what about frame event for other
155 - forward these input devices to the client
157 - "scancode 124 pressed or released with scan codes 18,22 and 30
158 held down gives control back to userspace wayland.
160 - what about maintaining cursor position? what about pointer
161 acceleration? maybe this only works in "client cursor mode",
162 where wayland hides the cursor and only sends relative events?
163 Solves the composited cursor problem. How does X show its
166 - Probably not worth it.