3 - We need rotation information in the output (multiples of 90
4 degress) and we'll need a way for a client to communicate that it
5 has rendered it's buffer according to the output rotation. The
6 goal is to be able to pageflip directly to the client buffer, and
7 for that we need the client to render accordingly and the
8 compositor needs to know that it did.
10 - Atomicity. Currently a lot of the atomicity in Wayland relies on
11 how we batch up all requests in a protocol buffer and only flushes
12 in the "blockhandler" in the client. Consensus was that we need
13 something more reliable and explicit. The suggestion is that we
14 make surface.attach a synchronization point such that everything
15 before that is batched and applied atomically when the
16 surface.attach request comes in. For cases where we need atomicity
17 beyond a surface.attach, we can add an atomic grouping mechanism,
18 that can group together multiple surface.attach requests into a
19 bigger atomic change. To be researched a bit.
21 - We should make pointer sprites regular surfaces. Something like
22 input_device.set_sprite(surface). This also make client side
23 animated cursors simple/possible, since we now get a frame event
24 that can drive the animation.
26 - Maybe try to make remote wayland actually happen, to see if there
27 is something in the protocol/architecute that makes it harder than
30 - Key events need a bit more work/thinking/redesign. As it is we need
31 a mechanism to change and synchronize keymaps, repeat rates. But
32 as we started talking, we decided that we needed to go back and
33 research what other systems do instead of just essentially copying
34 the X model. Sending out unicode events in addition to keycode
35 events has a lot of benefits (OSK can send out unicode events
36 instead of fake keycode events, apps become much simpler...) Move
37 keymap handling and repeat to the server? Needs more research.
39 - Pointer axis events need modifiers (ctrl-scroll eg), but we either
40 need to send the modifier state with each axis/scroll event or send
41 keys down on pointer_focus and subsequent key events... or just key
42 events for modifier keys... or for the non-repeating subset?
44 - Input protocol restructuring: break up events into wl_pointer
45 (enter/leave/motion/button/axis events, set_pointer_surface
46 request), wl_keyboard (enter/leave/key events... what
47 else... unicode event, set_map request? pending kb work), and
48 wl_touch (down/up/motion/cancel events) interfaces. Rename
49 wl_input_device to wl_seat. wl_seat has zero or one of each, and
50 will announce this at bind time. Raw devices are also tied to a
51 wl_seat, but we may not do that for 1.0, we just need to make sure
52 wl_seat has a forward compatible way to announce them.
54 - Add timestamp to touch_cancel, add touch id to touch_cancel (?)
56 - The output protocol needs to send all the ugly timing details for the modes.
60 - clipboard manager interface? what's needed? just notification
61 that the selection has gone away. should the clipboard manager be
62 able to take over the selection "seamlessly", ie, with the same
63 timestamp etc? Doesn't seem like we need that, the clipboard will
64 have to set a new data_source anyway, with the subset of mimetypes
65 it offers (the clipboad manager may only offer a subset of the
66 types offered by the original data_source)
68 - mime-type guidelines for data_source (ie, both dnd and selection):
69 recommended types for text or images, types that a clipboard
70 manager must support, mime-types must be listed in preferred order
72 - we need a "no kb focus please" mechanism. Or should this be
73 implicit in a specific surface type?
77 - configure should provide dx_left, dx_right, dy_top, dy_bottom, or
78 dx, dy, width and height.
80 - move to workspace, keep on top, on all workspaces, minimize etc
81 requests for implementing client side window menu? or just make a
82 "show window menu" request to let the compositor display and manage
85 - window move and resize functionality for kb and touch.
87 - dnd loose ends: self-dnd: initiate dnd with a null data-source,
88 compositor will not offer to other clients, client has to know
89 internally what's offered and how to transfer data. no fd passing.
91 - Protocol for specifying title bar rectangle (for moving
92 unresponsive apps). Rectangle for close button, so we can popup
93 force-close dialog if application doesn't respond to ping event
94 when user clicks there. We could use the region mechanism here
97 - popup placement protocol logic.
99 - subsurface mechanism. we need this for cases where we would use an
100 X subwindow for gl or video other different visual type.
104 - Don't wl_display_iterate in eglSwapBuffer, send an eventfd fd?
106 - Land Robert Braggs EGL extensions: frame age, swap with damage
108 - Make it possible to share buffers from compositor to clients.
109 Tricky part here is how to indicate to EGL on the server side that
110 it should make an EGLImage available to a client. We'll need a
111 "create a wl_buffer for this EGLImage for this client" kind of
114 - Protocol for arbitrating access to scanout buffers (physically
115 contiguous memory). When a client goes fullscreen (or ideally as
116 the compositor starts the animation that will make it fullscreen)
117 we send a "give up your scanout buffer" to the current fullscreen
118 client (if any) and when the client acks that we send a "try to
119 allocate a scanout buffer now" event to the fullscreen-to-be
126 - Needs a mechanism to pass buffers to client.
128 buffer = drm.create_buffer(); /* buffer with stuff in it */
130 cache.upload(buffer, x, y, width, height, int hash)
132 drm.buffer: id, name, stride etc /* event to announce cache buffer */
134 cache.image: hash, buffer, x, y, stride /* event to announce
135 * location in cache */
137 cache.reject: hash /* no upload for you! */
139 cache.retire: buffer /* cache has stopped using buffer, please
140 * reupload whatever you had in that buffer */
142 - A "please suspend" event from the compositor, to indicate to an
143 application that it's no longer visible/active. Or maybe discard
144 buffer, as in "wayland discarded your buffer, it's no longer
145 visible, you can stop updating it now.", reattach, as in "oh hey,
146 I'm about to show your buffer that I threw away, what was it
147 again?". for wayland system compositor vt switcing, for example,
148 to be able to throw away the surfaces in the session we're
149 switching away from. for minimized windows that we don't want live
150 thumb nails for. etc.
154 - pull in actions logic from xserver
156 - pull in keycode to keysym logic from libX11
158 - expose alloc functions in libxkbcommon, drop xserver funcs?
160 - pull the logic to write the xkb file from xkb_desc and names into
161 libxkbcommon and just build up the new xkb_desc instead of
162 dump+parse? (XkbWriteXKBKeymapForNames followed by
163 xkb_compile_keymap_from_string in XkbDDXLoadKeymapByNames)
165 - pull in keysym defs as XKB_KEY_BackSpace
167 - figure out what other X headers we can get rid of, make it not
168 need X at all (except when we gen the keysyms).
170 - Sort out namespace pollution (XkbFoo macros, atom funcs etc).
172 - Sort out 32 bit vmods and serialization
179 - draw window decorations in gtkwindow.c
181 - Details about pointer grabs. wayland doesn't have active grabs,
182 menus will behave subtly different. Under X, clicking a menu
183 open grabs the pointer and clicking outside the window pops down
184 the menu and swallows the click. without active grabs we can't
185 swallow the click. I'm sure there much more...
189 - Investigate DirectFB on Wayland (or is that Wayland on DirectFB?)
191 - SDL port, bnf has work in progress here:
192 http://cgit.freedesktop.org/~bnf/sdl-wayland/
194 - libva + eglimage + kms integration
199 - A wayland settings protocol to tell clients about themes (icons,
200 cursors, widget themes), fonts details (family, hinting
201 preferences) etc. Just send all settings at connect time, send
202 updates when a setting change. Getting a little close to gconf
203 here, but could be pretty simple:
205 interface "settings":
206 event int_value(string name, int value)
207 event string_value(string name, string value)
209 but maybe it's better to just require that clients get that from
210 somewhere else (gconf/dbus).
215 - AF_WAYLAND - A new socket type. Eliminate compositor context
216 switch by making kernel understand enough of wayland that it can
217 forward input events as wayland events and do page flipping in
218 response to surface_attach requests:
220 - ioctl(wayland_fd, "surface_attach to object 5 should do a kms page
223 - what about multiple crtcs? what about frame event for other
226 - forward these input devices to the client
228 - "scancode 124 pressed or released with scan codes 18,22 and 30
229 held down gives control back to userspace wayland.
231 - what about maintaining cursor position? what about pointer
232 acceleration? maybe this only works in "client cursor mode",
233 where wayland hides the cursor and only sends relative events?
234 Solves the composited cursor problem. How does X show its
237 - Probably not worth it.