3 - scanner: wl_* prefix removal: split it out into a namespace part so
4 we can call variables "surface" instead of "wl_surface"?
6 - We need rotation information in the output (multiples of 90
7 degress) and we'll need a way for a client to communicate that it
8 has rendered it's buffer according to the output rotation. The
9 goal is to be able to pageflip directly to the client buffer, and
10 for that we need the client to render accordingly and the
11 compositor needs to know that it did.
13 - Atomicity. Currently a lot of the atomicity in Wayland relies on
14 how we batch up all requests in a protocol buffer and only flushes
15 in the "blockhandler" in the client. Consensus was that we need
16 something more reliable and explicit. The suggestion is that we
17 make surface.attach a synchronization point such that everything
18 before that is batched and applied atomically when the
19 surface.attach request comes in. For cases where we need atomicity
20 beyond a surface.attach, we can add an atomic grouping mechanism,
21 that can group together multiple surface.attach requests into a
22 bigger atomic change. To be researched a bit.
24 - We should make pointer sprites regular surfaces. Something like
25 input_device.set_sprite(surface). This also make client side
26 animated cursors simple/possible, since we now get a frame event
27 that can drive the animation.
29 - Maybe try to make remote wayland actually happen, to see if there
30 is something in the protocol/architecute that makes it harder than
33 - Reconsider data types for coordinates in events. double, floats or
34 fixed point. Transformed and/or accelerated input generates
35 sub-pixel positions. 24.8 fixed point could work. Need to think
36 about the range of coordinates we need. Different from X problem,
37 since we don't have sub-windows, which is where X hits the 16 bit
38 limitations. Monitor and window sizes haven't yet out-grown the 16
41 - Key events need a bit more work/thinking/redesign. As it is we need
42 a mechanism to change and synchronize keymaps, repeat rates. But
43 as we started talking, we decided that we needed to go back and
44 research what other systems do instead of just essentially copying
45 the X model. Sending out unicode events in addition to keycode
46 events has a lot of benefits (OSK can send out unicode events
47 instead of fake keycode events, apps become much simpler...) Move
48 keymap handling and repeat to the server? Needs more research.
50 - Pointer axis events need modifiers (ctrl-scroll eg), but we either
51 need to send the modifier state with each axis/scroll event or send
52 keys down on pointer_focus and subsequent key events... or just key
53 events for modifier keys... or for the non-repeating subset?
55 - Input protocol restructuring: break up events into wl_pointer
56 (enter/leave/motion/button/axis events, set_pointer_surface
57 request), wl_keyboard (enter/leave/key events... what
58 else... unicode event, set_map request? pending kb work), and
59 wl_touch (down/up/motion/cancel events) interfaces. Rename
60 wl_input_device to wl_seat. wl_seat has zero or one of each, and
61 will announce this at bind time. Raw devices are also tied to a
62 wl_seat, but we may not do that for 1.0, we just need to make sure
63 wl_seat has a forward compatible way to announce them.
65 - Add timestamp to touch_cancel, add touch id to touch_cancel (?)
67 - The output protocol needs to send all the ugly timing details for the modes.
71 - clipboard manager interface? what's needed? just notification
72 that the selection has gone away. should the clipboard manager be
73 able to take over the selection "seamlessly", ie, with the same
74 timestamp etc? Doesn't seem like we need that, the clipboard will
75 have to set a new data_source anyway, with the subset of mimetypes
76 it offers (the clipboad manager may only offer a subset of the
77 types offered by the original data_source)
79 - mime-type guidelines for data_source (ie, both dnd and selection):
80 recommended types for text or images, types that a clipboard
81 manager must support, mime-types must be listed in preferred order
83 - we need a "no kb focus please" mechanism. Or should this be
84 implicit in a specific surface type?
88 - configure should provide dx_left, dx_right, dy_top, dy_bottom, or
89 dx, dy, width and height.
91 - move to workspace, keep on top, on all workspaces, minimize etc
92 requests for implementing client side window menu? or just make a
93 "show window menu" request to let the compositor display and manage
96 - window move and resize functionality for kb and touch.
98 - dnd loose ends: self-dnd: initiate dnd with a null data-source,
99 compositor will not offer to other clients, client has to know
100 internally what's offered and how to transfer data. no fd passing.
102 - Protocol for specifying title bar rectangle (for moving
103 unresponsive apps). Rectangle for close button, so we can popup
104 force-close dialog if application doesn't respond to ping event
105 when user clicks there. We could use the region mechanism here
108 - popup placement protocol logic.
110 - subsurface mechanism. we need this for cases where we would use an
111 X subwindow for gl or video other different visual type.
115 - Don't wl_display_iterate in eglSwapBuffer, send an eventfd fd?
117 - Land Robert Braggs EGL extensions: frame age, swap with damage
119 - Make it possible to share buffers from compositor to clients.
120 Tricky part here is how to indicate to EGL on the server side that
121 it should make an EGLImage available to a client. We'll need a
122 "create a wl_buffer for this EGLImage for this client" kind of
125 - Protocol for arbitrating access to scanout buffers (physically
126 contiguous memory). When a client goes fullscreen (or ideally as
127 the compositor starts the animation that will make it fullscreen)
128 we send a "give up your scanout buffer" to the current fullscreen
129 client (if any) and when the client acks that we send a "try to
130 allocate a scanout buffer now" event to the fullscreen-to-be
137 - Needs a mechanism to pass buffers to client.
139 buffer = drm.create_buffer(); /* buffer with stuff in it */
141 cache.upload(buffer, x, y, width, height, int hash)
143 drm.buffer: id, name, stride etc /* event to announce cache buffer */
145 cache.image: hash, buffer, x, y, stride /* event to announce
146 * location in cache */
148 cache.reject: hash /* no upload for you! */
150 cache.retire: buffer /* cache has stopped using buffer, please
151 * reupload whatever you had in that buffer */
153 - A "please suspend" event from the compositor, to indicate to an
154 application that it's no longer visible/active. Or maybe discard
155 buffer, as in "wayland discarded your buffer, it's no longer
156 visible, you can stop updating it now.", reattach, as in "oh hey,
157 I'm about to show your buffer that I threw away, what was it
158 again?". for wayland system compositor vt switcing, for example,
159 to be able to throw away the surfaces in the session we're
160 switching away from. for minimized windows that we don't want live
161 thumb nails for. etc.
165 - pull in actions logic from xserver
167 - pull in keycode to keysym logic from libX11
169 - expose alloc functions in libxkbcommon, drop xserver funcs?
171 - pull the logic to write the xkb file from xkb_desc and names into
172 libxkbcommon and just build up the new xkb_desc instead of
173 dump+parse? (XkbWriteXKBKeymapForNames followed by
174 xkb_compile_keymap_from_string in XkbDDXLoadKeymapByNames)
176 - pull in keysym defs as XKB_KEY_BackSpace
178 - figure out what other X headers we can get rid of, make it not
179 need X at all (except when we gen the keysyms).
181 - Sort out namespace pollution (XkbFoo macros, atom funcs etc).
183 - Sort out 32 bit vmods and serialization
190 - draw window decorations in gtkwindow.c
192 - Details about pointer grabs. wayland doesn't have active grabs,
193 menus will behave subtly different. Under X, clicking a menu
194 open grabs the pointer and clicking outside the window pops down
195 the menu and swallows the click. without active grabs we can't
196 swallow the click. I'm sure there much more...
200 - Investigate DirectFB on Wayland (or is that Wayland on DirectFB?)
202 - SDL port, bnf has work in progress here:
203 http://cgit.freedesktop.org/~bnf/sdl-wayland/
205 - libva + eglimage + kms integration
210 - A wayland settings protocol to tell clients about themes (icons,
211 cursors, widget themes), fonts details (family, hinting
212 preferences) etc. Just send all settings at connect time, send
213 updates when a setting change. Getting a little close to gconf
214 here, but could be pretty simple:
216 interface "settings":
217 event int_value(string name, int value)
218 event string_value(string name, string value)
220 but maybe it's better to just require that clients get that from
221 somewhere else (gconf/dbus).
226 - AF_WAYLAND - A new socket type. Eliminate compositor context
227 switch by making kernel understand enough of wayland that it can
228 forward input events as wayland events and do page flipping in
229 response to surface_attach requests:
231 - ioctl(wayland_fd, "surface_attach to object 5 should do a kms page
234 - what about multiple crtcs? what about frame event for other
237 - forward these input devices to the client
239 - "scancode 124 pressed or released with scan codes 18,22 and 30
240 held down gives control back to userspace wayland.
242 - what about maintaining cursor position? what about pointer
243 acceleration? maybe this only works in "client cursor mode",
244 where wayland hides the cursor and only sends relative events?
245 Solves the composited cursor problem. How does X show its
248 - Probably not worth it.